Luminous Landscape Forum
Raw & Post Processing, Printing => Colour Management => Topic started by: daicehawk on March 17, 2018, 03:04:04 pm
-
The DXOmark use a LED lightbox, the Imaging Resource HMI bulbs for the "skylight" component in the "sunlit" tests, which IMO renders the results invalid and defies the purpose of the color accuracy measurement. Any peaky spectrum is a NO-NO in my book no matter the CRI score.
What do you guys think?
-
Absolutely agree. But the only light source that produces a standard illuminate is 93 million miles away. Using a man made source, the quality and matching differ. CRI is kind of a hack too. CQS (15 very colorful patches) a bit better. That doesn't tell us about the spectrum which is even a better way to evaluate the illuminant.
-
I see no problem with any continuous spectrum be it the sun or a man-made filtered halogen (though very inefficient and hot). The sun\sky light balance varies a lot during the day and across the seasons with the spectrum continuity being the main attribute of the high CRI no matter the test sample set reflectance spectrum.
The Blue-Yellow LEDs lack the cyan range, the HMI are just a nightmare and both must be tweaked to get a high CRI. I think it is a shame the QDots +LEDs solution has not been implemented.
-
The DXOmark use a LED lightbox, the Imaging Resource HMI bulbs for the "skylight" component in the "sunlit" tests, which IMO renders the results invalid and defies the purpose of the color accuracy measurement. Any peaky spectrum is a NO-NO in my book no matter the CRI rate.
What do you guys think?
Tests run with a particular illuminant and a particular target patch set are useful in determining the camera's response to that patch set under that illuminant. That is the high road.
It is thought that tests run with illuminants and patch sets with broad smooth spectra generalize to other broad-spectrum illuminants and other broad-spectrum patch sets. That is usually the case, but spiky camera response can throw that for a loop.
jim
-
Tests run with a particular illuminant and a particular target patch set are useful in determining the camera's response to that patch set under that illuminant. That is the high road.
It is thought that tests run with illuminants and patch sets with broad smooth spectra generalize to other broad-spectrum illuminants and other broad-spectrum patch sets. That is usually the case, but spiky camera response can throw that for a loop.
jim
Those are obvious things, the point of my post being such test lights invalidate any judgments about the camera response and it degree of conformity to the Luther-Ives condition.
-
Those are obvious things, the point of my post being such test lights invalidate any judgments about the camera response and it degree of conformity to the Luther-Ives condition.
What camera conforms to the Luther-Ives condition?
-
What camera conforms to the Luther-Ives condition?
That depends on the acceptance threshold and selected preferences on what colors are more important to be reproduced accurately. Please read "its degree of conformity" and do not start an argument.
-
That depends on the acceptance threshold and selected preferences on what colors are more important to be reproduced accurately. Please read "its degree of conformity" and do not start an argument.
So you can’t answer my question.
-
That depends on the acceptance threshold and selected preferences on what colors are more important to be reproduced accurately. Please read "its degree of conformity" and do not start an argument.
Is 'degree of conformity' a scalar, and if so, how is it defined?
Jim
-
Those are obvious things, the point of my post being such test lights invalidate any judgments about the camera response and it degree of conformity to the Luther-Ives condition.
Even with a perfectly smooth, black body spectra, how well a camera can track a ColorChecker is only imperfectly correlated with how well it meets LI. A hypothetical camera and a perfect black body source, might be great with a ColorChecker and not so good with other colors, even daylight metamers of ColorChecker colors. To really characterize a camera system against LI, the gold standard is to measure its post CFA spectral responses.
But that's hard.
Of course spikey illuminants only add uncertainty. It would be interesting to compare the two sources of error.
-
Is 'degree of conformity' a scalar, and if so, how is it defined?
Jim
That’s an easy one Jim. It’s a scale from 1 to 10 with 10 being best. 😀
-
Is 'degree of conformity' a scalar, and if so, how is it defined?
Jim
I define this as "a wide color range scene including memory colors such as the sky, skin, sea water, foliage, fruits etc. shot under sunlight looks good and real on a wide gamut display merely developed with a matrix profile without having to resort to selective color correction in PS"
-
CQS (15 very colorful patches) a bit better.
The CRI replacement that is gaining the most interest lately is IES TM-30-15 (https://www.ies.org/store/technical-memoranda/ies-method-for-evaluating-light-source-color-rendition/).
-
Even with a perfectly smooth, black body spectra, how well a camera can track a ColorChecker is only imperfectly correlated with how well it meets LI. A hypothetical camera and a perfect black body source, might be great with a ColorChecker and not so good with other colors, even daylight metamers of ColorChecker colors. To really characterize a camera system against LI, the gold standard is to measure its post CFA spectral responses.
But that's hard.
Of course spikey illuminants only add uncertainty. It would be interesting to compare the two sources of error.
The colorchecker is good in representing memory colors in a small sample set. BTW you don`t have many "daylight metamers" for orange, yellow, red and aqua in the nature. Highly saturated reflected colors are less prone to metamerism.
100% LI conformity is not the goal, the "acceptable" color accuracy is. Perfect is the enemy of good.
-
Jim, was that an acceptable answer to your question?
My one question wasn't acceptably answered.
I think perhaps other than a very interesting post by GWGILL (thanks, I'll certainly look into that), this is a discussion better suited for the "Open Forums" over on DP Review where I will not enter. :o
-
What camera conforms to the Luther-Ives condition?
This one. (http://djjoofa.com/data/images/hyperspectral.jpg)
-
This one. (http://djjoofa.com/data/images/hyperspectral.jpg)
I still have want bumps for the RX10 :o
-
The CRI replacement that is gaining the most interest lately is IES TM-30-15 (https://www.ies.org/store/technical-memoranda/ies-method-for-evaluating-light-source-color-rendition/).
The posted video in that link is the best I've seen explaining these issues with various manufactured lighting. It made total sense on how it will be useful for me.
-
Jim, was that an acceptable answer to your question?
My one question wasn't acceptably answered.
I think perhaps other than a very interesting post by GWGILL (thanks, I'll certainly look into that), this is a discussion better suited for the "Open Forums" over on DP Review where I will not enter. :o
This one: "That’s an easy one Jim. It’s a scale from 1 to 10 with 10 being best."?
I was hoping for something objective.
Jim
-
I define this as "a wide color range scene including memory colors such as the sky, skin, sea water, foliage, fruits etc. shot under sunlight looks good and real on a wide gamut display merely developed with a matrix profile without having to resort to selective color correction in PS"
I was asking for an objective way to make that measurement. "Looks good and real" sounds subjective to me. I might think that camera A is better than camera B by that test, and you might think the reverse. Besides, you left a lot of wiggle room in the criteria for selecting the compromise matrix.
Jim
-
The colorchecker is good in representing memory colors in a small sample set. BTW you don`t have many "daylight metamers" for orange, yellow, red and aqua in the nature. Highly saturated reflected colors are less prone to metamerism.
100% LI conformity is not the goal, the "acceptable" color accuracy is. Perfect is the enemy of good.
No, reaching LI nirvana is unrealistic. However, I have a white LED that has quite good spectral characteristics outside of the hump at 460nm and dip around 500nm. It would have a hard time rendering a highly saturated cyan (in chromaticity). The luminosity of a saturated cyan is quite low. But it has no problem rendering the cyan in a colorchecker because it isn't very saturated and the peak/valley pretty much overlaps the cyan patches reflectance spectrum. But if I had a cyan that had a passband from 480 to 530 the LED would fail, creating a much darker cyan than D50 with the same Lux.
I also took a lot at a couple oranges that are D50 metamers. Turns out the reflectance spectrum of one differs quite a bit from the other with the main shift occurring 15nm different from each other. They would fail to be metamers if the illuminant had abrupt changes in that 15nm interval and there are fluorescents that do. And orange is one of those colors very sensitive to hue shift.
But it comes down to both the illuminant and the CFA filters. Both contribute to essential error. How would you quantify their relative contributions? I've looked at the color shifts induced by F8 fluorescents and a high CRI ( >95 ) LED. The LED wins out over the F8 but still produces dEs of 1-3 with my print inks. I believe Jim K. has done some work with camera imagers. IIRC, the dEs were somewhat larger.
In any case the contribution of both the illuminant and sensor/CFA to color error needs to be quantified. I don't put much credence in imatest's color accuracy measurements because they don't have a way to measure error from the illuminant even though I suspect it is smaller than the sensor/CFA. But who knows for sure?
-
No, reaching LI nirvana is unrealistic.
So what about post #16?
-
It occurs to me that if the camera were LI, the light source wouldn't make any difference. The camera would faithfully recreate the response of the Standard Observer to the patch under consideration with the illuminant chosen.
Jim
-
So what about post #16?
I can't find it at B&H.
-
It occurs to me that if the camera were LI, the light source wouldn't make any difference. The camera would faithfully recreate the response of the Standard Observer to the patch under consideration with the illuminant chosen.
Jim
I absolutely agree!
-
It occurs to me that if the camera were LI, the light source wouldn't make any difference. The camera would faithfully recreate the response of the Standard Observer to the patch under consideration with the illuminant chosen.
Jim
I absolutely agree!
Which is why the spectral response of the CFA/sensor is the gold standard. Illuminant drops out of the equation.
-
The CRI replacement that is gaining the most interest lately is IES TM-30-15 (https://www.ies.org/store/technical-memoranda/ies-method-for-evaluating-light-source-color-rendition/).
In the video the hue shift of the bell peppers caused by saturation and luminance changes (supposedly caused by the light source), does not happen in reality as far as I've seen. But how much of the color changes are due to the camera and software in what appears to me to be from a scene that doesn't look quite accurate with regards to contrast levels. I realize it's just a demo but it does show what I have to deal with editing images that has nothing to do with the light source used.
After editing 1000's of Raws over the years I am constantly custom shaping ACR's Point Curve in the mid range to deep shadows section in order to override this contrast inducing hue/sat shift that's part of the "filmic" look engineered into ACR's color rendering tools. Flat Linear curve looks too washed out making it not so accurate to the actual scene. Clarity and definition takes a hit.
Accuracy of the scene to the viewer is highly affected by this contrast weighted appearance which has to be viewed on a display having the same contrast and overall luminance. I don't see how anyone can separate what image rendering characteristic is being influenced by other capture/rendering mechanics and know for sure it's caused by the light source the scene is lit by compared to the display.
Even when image to scene accuracy examples are rendered by measured numbers it doesn't look so accurate as well. The camera will always get in the way and a camera profile only affects hue/sat and some contrast rendering but not enough to make the scene captured appear accurate.
-
Even when image to scene accuracy examples are rendered by measured numbers it doesn't look so accurate as well. The camera will always get in the way and a camera profile only affects hue/sat and some contrast rendering but not enough to make the scene captured appear accurate.
Because (again!) there is a massive difference between color perception and color appearance. We've been over this no*?
I'll refrain from asking you what you mean by 'accuracy' below.
*Colorimetry and the dE testing is about color perception. It is not about color appearance. The reason why viewing a print is more valid than measuring it is because measurement is about comparing solid colors. Color appearance is about evaluating images and color in context which measurement devices can't provide. Colorimetry is about color perception. It is not about color appearance. Colorimetry was never, designed as a color appearance model. It was never designed to even be used as an interchange space between device dependent color models. It's not designed for imagery at all. Colorimetry based on solid colors in very specific ambient and surround conditions.
-
I can't find it at B&H.
Doesn't matter. The point was having a camera. Not one at B&H.
-
Because (again!) there is a massive difference between color perception and color appearance. We've been over this no*?
I'll refrain from asking you what you mean by 'accuracy' below.
*Colorimetry and the dE testing is about color perception. It is not about color appearance. The reason why viewing a print is more valid than measuring it is because measurement is about comparing solid colors. Color appearance is about evaluating images and color in context which measurement devices can't provide. Colorimetry is about color perception. It is not about color appearance. Colorimetry was never, designed as a color appearance model. It was never designed to even be used as an interchange space between device dependent color models. It's not designed for imagery at all. Colorimetry based on solid colors in very specific ambient and surround conditions.
You don't offer solutions, Andrew. You only state absolutes based on data that doesn't help photographers or explain the issue I just described about contrast and what and why there are hue shifts with saturation and luminous changes. You don't provide any practical information that makes a connection between light source used to photograph or view prints under that cause this contrast/hue change.
And thanks for refraining from stating the obvious about color perception and appearance. It would help if you could provide data other than spectral graphs and dry color science articles.
-
You don't offer solutions, Andrew. You only state absolutes based on data that doesn't help photographers or explain the issue I just described about contrast and what and why there are hue shifts with saturation and luminous changes.
I'd advise photographers to ignore this entirely and render the images as they desire. Which has absolutely nothing to do with color accuracy nor must it!
-
You only state absolutes based on data that doesn't help photographers or explain the issue I just described about contrast and what and why there are hue shifts with saturation and luminous changes.
The reason that most -- but not all -- saturation and luminance changes performed in RGB color spaces result in hue shifts is not a mystery, nor is it without mitigation:
http://www.ingentaconnect.com/content/ist/cic/1994/00001994/00000001/art00023
Ps has blending modes that do similar things to the techniques that I developed in the paper. If you're interested in the math behind this and don't want to spring for the 12 bucks, I might be able to dredge up a copy of the paper and scan it.
Jim
-
The reason that most -- but not all -- saturation and luminance changes performed in RGB color spaces result in hue shifts is not a mystery, nor is it without mitigation:
http://www.ingentaconnect.com/content/ist/cic/1994/00001994/00000001/art00023
Ps has blending modes that do similar things to the techniques that I developed in the paper. If you're interested in the math behind this and don't want to spring for the 12 bucks, I might be able to dredge up a copy of the paper and scan it.
Jim
I don't see how that would explain how this has anything to do with a light source spectra' affect on color appearance accuracy in repro work where the subject lit by manufactured lighting is right next to the calibrated editing workstation. I also don't see how that can be integrated within a Raw workflow within the Raw converter. Not interested in working on pixels if that's where the math is being applied.
Yes, I know this RGB hue shift behavior is not a mystery. It's a known issue for some time but the corrections and the complexity involved to mitigate this I've seen implemented show a complete lack of knowing when something doesn't look right because the methods used are so complex and scientific based that to find what to do to make it look right isn't worth the time to dig into.
Nothing's perfect I guess even when math and science is used.
Scientists seem to have a different perception of color relationships.
Thanks for the offer to scan the papers you mentioned but I'm not going to be able to understand it anyway.
-
I don't see how that would explain how this has anything to do with a light source spectra' affect on color appearance accuracy in repro work where the subject lit by manufactured lighting is right next to the calibrated editing workstation..
Repro work is quite a different beast. There colorimetry rules. You don't even need a good monitor because you are trying to get the same colors (in a colorimetric sense) on the repro as are on the original. The appearances should match if the perceptual (colorimetry) matches. The largest sources of errors doing this are from the illuminant's divergence from D50 (if using D50 for profiles) and the camera (or scanner) divergence from LI.
Generally, for repro work one doesn't tweak colors (there are exceptions where the imager capture is off in critical areas due to LI) . One also uses Abs. Col. Intent. The most important thing is to match the white point or most luminous part of the original to the captured image. This all assumes the gamut of the printer is sufficient to encapsulate the colors in the original. If not your reproduction will be flawed at the start.
-
The largest sources of errors doing this are from the illuminant's divergence from D50 (if using D50 for profiles) and the camera (or scanner) divergence from LI.
Is what is degree such errors?
-
Repro work is quite a different beast. There colorimetry rules. You don't even need a good monitor because you are trying to get the same colors (in a colorimetric sense) on the repro as are on the original. The appearances should match if the perceptual (colorimetry) matches. The largest sources of errors doing this are from the illuminant's divergence from D50 (if using D50 for profiles) and the camera (or scanner) divergence from LI.
Generally, for repro work one doesn't tweak colors (there are exceptions where the imager capture is off in critical areas due to LI) . One also uses Abs. Col. Intent. The most important thing is to match the white point or most luminous part of the original to the captured image. This all assumes the gamut of the printer is sufficient to encapsulate the colors in the original. If not your reproduction will be flawed at the start.
Don't you think the video's demonstration of colored LED's shifting hue/sat in the bell pepper examples as a predictor of white light of various spectra output isn't being cause by the light but through maybe software manipulation? I've never seen any of the white lights I've used do that to a bell pepper.
That site's info and video is made by very learned color scientists and engineers, so I'm surprised they attribute that color behavior to lighting. Most white lighting I've worked with applies a more flat filter effect as an evenly distributed wash/stain of the color temp hue over the entire scene but never to the extent they shift hue/sat that severely.
I've never seen a camera profile attempt to characterize that behavior and correct it applying the profile in post. Editing is further required.
-
Don't you think the video's demonstration of colored LED's shifting hue/sat in the bell pepper examples as a predictor of white light of various spectra output isn't being cause by the light but through maybe software manipulation? I've never seen any of the white lights I've used do that to a bell pepper.
That site's info and video is made by very learned color scientists and engineers, so I'm surprised they attribute that color behavior to lighting. Most white lighting I've worked with applies a more flat filter effect as an evenly distributed wash/stain of the color temp hue over the entire scene but never to the extent they shift hue/sat that severely.
I've never seen a camera profile attempt to characterize that behavior and correct it applying the profile in post. Editing is further required.
Tim, what they are doing is rather unique. They are creating "white" light by selecting and changing the contributions of a large number of LEDs, each with a fairly narrow bandwidth. They have a diffuser so that you don't see any of the LED colors separately. And the "whites" are all computer controlled so that the chromaticity values are those of D60 above 6000K and black body below 4000K, avoiding the jump that is made at D50 to 4999K.
They then modify the spectrum but maintain the chromaticity of the white point.
Think of the specific example of your monitor. Let's assume you profiled it for D65 and filled it with white. That white does not have the same spectrum as D65. It just has the same chromaticity so it looks white and will appear to have the same white as D65 daylight. But spectrally it is VERY lumpy. One of the consequences of that is if you use your monitor's "white" output to illuminate say, a red apple. It will actually appear to be a more saturated red than if you illuminated it with D65 daylight at the same Lux.
What they are doing is demonstrating that by creating lumpy spectra that is still "white" one can increase the color saturation of many common objects. They believe this might be of interest commercially. For instance to enhance the appearance of produce or make product packaging stand out more.
And, there are ways to decrease saturation by moving the spectrum inwards from both ends a bit so long as they maintain the chromaticity white point. The whites still look the same but all the colors from the objects change.
Along the way they describe a couple metrics, one of which could be used as an improved CRI.
-
Is what is degree such errors?
Not sure exactly what you're asking. Generally, what few problems I run into where repro colors are noticeably different are in things like magenta and oranges that are more perceptually sensitive to hue errors. For instance a purplish blue dress might not get rendered by the imager accurately. The problem is almost always that my imaging system simply has the color too far off. LI strikes again. I'll take a spectro reading of the original if it's critical. Still, pretty rare and easily fixed in Photoshop.
-
Hi,
I think Jim Kasson did present a list of cameras fulfilling the Luther-Ives condition. It is reproduced here:
Best regards
Erik
What camera conforms to the Luther-Ives condition?
-
since there was nothing there. I assume that's what you mean.
-
Hi,
According to this article: https://www.imaging-resource.com/TIPS/TESTS/TESTS.HTM, Imaging Resource uses Solux lamps at 15.9V. That is continuous spectrum.
Best regards
Erik
The DXOmark use a LED lightbox, the Imaging Resource HMI bulbs for the "skylight" component in the "sunlit" tests, which IMO renders the results invalid and defies the purpose of the color accuracy measurement. Any peaky spectrum is a NO-NO in my book no matter the CRI score.
What do you guys think?
-
The DXOmark use a LED lightbox, the Imaging Resource HMI bulbs for the "skylight" component in the "sunlit" tests, which IMO renders the results invalid and defies the purpose of the color accuracy measurement. Any peaky spectrum is a NO-NO in my book no matter the CRI score.
What do you guys think?
I think that you need to approach this a bit more carefully - unlike Xenon sources for example, just plain reference to LED as a light source says absolutely nothing. Simply because they are all different, use different phosphors and pursue different goals. The quality and chromacity of the LED source is also very dependent on the current passing through it so a very good power supply is a must - i.e. what passes as good enough for kitchen light may not do well for colour applications.
As far as LEDs go - Yuji D50 LEDs are excellent light source for colour related applications (even their VTC 6500K LEDs are very good - slightly less flat but more powerful).
-
I think that you need to approach this a bit more carefully - unlike Xenon sources for example, just plain reference to LED as a light source says absolutely nothing. Simply because they are all different, use different phosphors and pursue different goals. The quality and chromacity of the LED source is also very dependent on the current passing through it so a very good power supply is a must - i.e. what passes as good enough for kitchen light may not do well for colour applications.
As far as LEDs go - Yuji D65 LEDs are excellent light source for colour related applications (even their VTC 6500K LEDs are very good - slightly less flat but more powerful).
There is reference to "Kiyoritsu color lightbox" on the DXOmark's site, no model whatsoever, I doubt they use Yuji (which still has a cyan dip and a green peaklet).
-
Hi,
According to this article: https://www.imaging-resource.com/TIPS/TESTS/TESTS.HTM, Imaging Resource uses Solux lamps at 15.9V. That is continuous spectrum.
Best regards
Erik
https://www.imaging-resource.com/ARTS/TESTS/SUNSOURCE2.HTM
-
since there was nothing there. I assume that's what you mean.
+1
Request for specifics in this thread (like those with respect to accuracy) fall into the same camp it appears. :o
-
There is reference to "Kiyoritsu color lightbox" on the DXOmark's site, no model whatsoever, I doubt they use Yuji (which still has a cyan dip and a green peaklet).
What's the point of speculating - just ask them for a reference? My point however was that simply dismissing all LEDs is not very wise.
-
I think Jim Kasson did present a list of cameras fulfilling the Luther-Ives condition. It is reproduced here:
There were some qualifiers and context. Here is the post where I said that:
https://blog.kasson.com/the-last-word/how-cameras-and-people-see-color/
Jim
-
Jim,
It occurs that the setup Graeme linked to Spectrally Variable illuminant (https://www.ies.org/store/technical-memoranda/ies-method-for-evaluating-light-source-color-rendition/) could be used to get much closer to L/I.
Since it uses a spectral range of computer controlled LEDs to shape the illuminant, this could be used to similar effect as the papers you have linked to. But a big advantage is the LEDs can be instantly changed enabling rapid, successive image shots. Beyond L/I, one could also make reasonably good estimates of the reflectance spectra at each location. This might make preservation of historical/museum artwork images cheaper, faster, and with very good quality.
-
Jim,
It occurs that the setup Graeme linked to Spectrally Variable illuminant (https://www.ies.org/store/technical-memoranda/ies-method-for-evaluating-light-source-color-rendition/) could be used to get much closer to L/I.
Since it uses a spectral range of computer controlled LEDs to shape the illuminant, this could be used to similar effect as the papers you have linked to. But a big advantage is the LEDs can be instantly changed enabling rapid, successive image shots. Beyond L/I, one could also make reasonably good estimates of the reflectance spectra at each location. This might make preservation of historical/museum artwork images cheaper, faster, and with very good quality.
The LI criterion ensures that the camera sees the scene, lit with an arbitrary illuminant, the same way as the Standard Observer.
Changing the light source isn't going to change that.
If you mean that the objective is that the camera, using a experimentally controlled illuminant, sees the scene the same way as the Standard Observer with some other illuminant, then the light source referenced could help. If you are allowed three exposures, a light source capable of artibrary spectrum could satisfy that objective. But that's not LI.
Or maybe I misunderstand your point.
Jim
-
I'll direct this to the attention of those who may not have read it:
http://dougkerr.net/Pumpkin/articles/Sensor_Colorimetry.pdf (http://dougkerr.net/Pumpkin/articles/Sensor_Colorimetry.pdf)
The colorimetric researchers (von)Luther and Ives10 showed mathematically that the outputs of a set of photodetectors will consistently describe the color of the light, regardless of the specific spectrum involved, if (and only if):
The response curves of the three types of photodetectors are linear combinations of the eye cone response functions l, m, and s (there being a couple of other requirements, some pesky stuff about orthogonality and so forth).
-
The LI criterion ensures that the camera sees the scene, lit with an arbitrary illuminant, the same way as the Standard Observer.
Changing the light source isn't going to change that.
If you mean that the objective is that the camera, using a experimentally controlled illuminant, sees the scene the same way as the Standard Observer with some other illuminant, then the light source referenced could help. If you are allowed three exposures, a light source capable of artibrary spectrum could satisfy that objective. But that's not LI.
Or maybe I misunderstand your point.
Jim
Of course. It's the ability to take multiple exposures with different spectral illuminants which allows one to process the exposures to produce an image much closer to L/I and with flexibility on the illuminant. That is, one could process the exposures and produce images closer to L/I for relatively arbitrary illuminants. That would require maintaining each of the sequential images or, alternately, processing them to extract subsets that have minimal information loss via SVD.
-
Of course. It's the ability to take multiple exposures with different spectral illuminants which allows one to process the exposures to produce an image much closer to L/I and with flexibility on the illuminant. That is, one could process the exposures and produce images closer to L/I for relatively arbitrary illuminants. That would require maintaining each of the sequential images or, alternately, processing them to extract subsets that have minimal information loss via SVD.
If I were really going to build a camera that worked this way, I'd skip the CFA and go for a monochromatic sensor, and make three exposures per setup. Alternatively, you could set the light source to as close to single wavelength as you could get and make one set of exposures incrementing lambda by 5 or 10 nm after each one. Then, knowing the camera's native spectral sensitivity, you could compute what the scene would look like with an arbitrary illuminant. This might be practical for digitizing artwork.
Jim
-
Of course. It's the ability to take multiple exposures with different spectral illuminants which allows one to process the exposures to produce an image much closer to L/I and with flexibility on the illuminant. That is, one could process the exposures and produce images closer to L/I for relatively arbitrary illuminants. That would require maintaining each of the sequential images or, alternately, processing them to extract subsets that have minimal information loss via SVD.
By "closer to L/I" above, I believe you mean "closer to colorimetric". Right?
Jim
-
By "closer to L/I" above, I believe you mean "closer to colorimetric". Right?
Jim
That would be accurate as well. If we add additional filters to the standard 3 in a CFA to N, any 3 of the filters will be a certain distance from L/I but using all N filters, in a linear combination, allows a substantial improvement and a much smaller distance away from L/I. That this worked quite well with only a small increase in filters was a principal result of the paper you linked in your blog.
Their multi-spectral approach allows not only better L/I approximation but could provide good approximations with different illuminants. A perfect camera that produces XYZ values, as opposed to spectral values, meeting L/I under one illuminant would only work for that one illuminant.
-
Alternatively, you could set the light source to as close to single wavelength as you could get and make one set of exposures incrementing lambda by 5 or 10 nm after each one. Then, knowing the camera's native spectral sensitivity, you could compute what the scene would look like with an arbitrary illuminant. This might be practical for digitizing artwork.
The camera that was mentioned earlier, and which is not available at B&H, did that. Though, the light source was not tempered with. The camera had special filters to scan through a narrow bandwidth (~1nm) in the visual range.
-
The camera had special filters to scan through a narrow bandwidth (~1nm) in the visual range.
ouch, those were that narrow-band pass-through filters ?
-
ouch, those were that narrow-band pass-through filters ?
Yes, specialized AOTF filters.
-
Hi,
The Bayer pattern could easily be modified for four colours, as it has two green pixels. So, using two different green would be possible. It was tried but never caught on.
Best regards
Erik
That would be accurate as well. If we add additional filters to the standard 3 in a CFA to N, any 3 of the filters will be a certain distance from L/I but using all N filters, in a linear combination, allows a substantial improvement and a much smaller distance away from L/I. That this worked quite well with only a small increase in filters was a principal result of the paper you linked in your blog.
Their multi-spectral approach allows not only better L/I approximation but could provide good approximations with different illuminants. A perfect camera that produces XYZ values, as opposed to spectral values, meeting L/I under one illuminant would only work for that one illuminant.
-
My point however was that simply dismissing all LEDs is not very wise.
FYI if interested, SPD for Yuji VTC 5600K LED (with CSV file as well) - in attachments.
Their D50 series is even better.
-
FYI if interested, SPD for Yuji VTC 5600K LED (with CSV file as well) - in attachments.
Their D50 series is even better.
Alexey, what do those bulbs cost and where do they sell them? Thanks for posting the info.
-
Alexey, what do those bulbs cost and where do they sell them? Thanks for posting the info.
Tim - this is their general site https://www.yujiintl.com/ which has some generic info (but spread across various sections somewhat illogically). They don't do bulbs, but they do LED strips or just LEDs themselves.
They have webstore with most of the LEDs data sheets. I got mine directly from them - promptly delivered. The relevant links for their LEDs:
- D50 - https://store.yujiintl.com/collections/vtc-d50-series (spec is at the bottom of one of the products page - here for example (https://cdn.shopify.com/s/files/1/0344/6401/files/YJWJ005-1.0.0_YJ-VTC-5730-D50-G01.pdf?244800397304522160)). I got an assorted LED pack - in the process of doing my own lighting panel to have a D50 lit box for target shooting
- VTC series - this is the COB I used (https://store.yujiintl.com/collections/vtc-series/products/vtc-series-high-cri-cob-led-135l-pack-5-pcs) in 5600K version for spectrum I posted. PDF spec of those is here (https://cdn.shopify.com/s/files/1/0344/6401/files/YJWJ040-1.0_YJ-VTC-135L-G01.pdf?13081682755939051710). I am evaluating using them as a light source for my camera spectral measurements device instead of Xenon.
For VTC series (CRI 98) LEDs they have big variety of chips, I just chose a largest COBs in those models to fit my application. I also designed my own driver for them since having a constant current supply for LEDs ensured the stability of the output and avoids colour shifts (but there are plenty readily available of course).
-
Tim - this is their general site https://www.yujiintl.com/ which has some generic info (but spread across various sections somewhat illogically). They don't do bulbs, but they do LED strips or just LEDs themselves.
They have webstore with most of the LEDs data sheets. I got mine directly from them - promptly delivered. The relevant links for their LEDs:
- D50 - https://store.yujiintl.com/collections/vtc-d50-series (spec is at the bottom of one of the products page - here for example (https://cdn.shopify.com/s/files/1/0344/6401/files/YJWJ005-1.0.0_YJ-VTC-5730-D50-G01.pdf?244800397304522160)). I got an assorted LED pack - in the process of doing my own lighting panel to have a D50 lit box for target shooting
- VTC series - this is the COB I used (https://store.yujiintl.com/collections/vtc-series/products/vtc-series-high-cri-cob-led-135l-pack-5-pcs) in 5600K version for spectrum I posted. PDF spec of those is here (https://cdn.shopify.com/s/files/1/0344/6401/files/YJWJ040-1.0_YJ-VTC-135L-G01.pdf?13081682755939051710). I am evaluating using them as a light source for my camera spectral measurements device instead of Xenon.
For VTC series (CRI 98) LEDs they have big variety of chips, I just chose a largest COBs in those models to fit my application. I also designed my own driver for them since having a constant current supply for LEDs ensured the stability of the output and avoids colour shifts (but there are plenty readily available of course).
OK, so you have to roll your own bulb with Yuji LED chips. And quite expensive chips.
You wouldn't happen to know of any company that's manufactured a finished bulb using these Yuji chips that can be purchased in the US?
Thanks for the links and for the reply.
-
You wouldn't happen to know of any company that's manufactured a finished bulb using these Yuji chips that can be purchased in the US?
https://dtdch.com/dt-photon-custom-cultural-heritage-lighting/
$6K a piece or so
PS: what you want is a "panel", not a "bulb"... if you are to illuminate a target you want less flat-fielding
-
https://dtdch.com/dt-photon-custom-cultural-heritage-lighting/
$6K a piece or so
PS: what you want is a "panel", not a "bulb"... if you are to illuminate a target you want less flat-fielding
$6K?! Ya' say! I guess that's the go-away-ya'-bothering-price.
Thanks for the link. Interesting info.
-
One thing that is nice (or not) is that these LEDs use violet as the base emission and have a considerable amount of uV. It may well provide full D50 levels. Not spectrally, of course, but enough near and below 400nm to fluoresce OBAs. Might work reasonably well for uV compliant D50ish light booths. If so one will need to use M1 or possibly M0 profiles for papers with OBAs.
-
One thing that is nice (or not) is that these LEDs use violet as the base emission and have a considerable amount of uV. It may well provide full D50 levels. Not spectrally, of course, but enough near and below 400nm to fluoresce OBAs. Might work reasonably well for uV compliant D50ish light booths. If so one will need to use M1 or possibly M0 profiles for papers with OBAs.
I believe I may be seeing the uV spike in the 5000K Hyperikon LED I bought on Amazon for $9. To get good Caucasian skin tone (not reddish or jaundice) the whites of the eyes turn pinkish blue compared to the Soraa LED 5000K Vivid with Gan technology where color is more balanced and realistic. Whites of the eyes are more of an ivory off white. Not as perfect using xenon or mercury based flash though.
But what I'm describing may have nothing to do with your point about these lights exciting OBA's in paper.
-
I believe I may be seeing the uV spike in the 5000K Hyperikon LED I bought on Amazon for $9. To get good Caucasian skin tone (not reddish or jaundice) the whites of the eyes turn pinkish blue compared to the Soraa LED 5000K Vivid with Gan technology where color is more balanced and realistic. Whites of the eyes are more of an ivory off white. Not as perfect using xenon or mercury based flash though.
Could be. It's not often discussed but there is a lot of variation amongst individuals, natural from small genetic differences and also age related. Even excluding the 8% of males that have color deficiencies. This is one of the main reasons the old CIE 1931 color matching functions are still used today in spite of significant technical improvements in lab equipment.
It's also in the shorter wavelengths (blues) that some of the larger individual variations occurs. LEDs generally, because they have such a large blue/violet spike, may not be as visually consistent among a group of people as flatter illuminants.
In other words, YMMV is a truism.
-
OK, so you have to roll your own bulb with Yuji LED chips. And quite expensive chips.
You wouldn't happen to know of any company that's manufactured a finished bulb using these Yuji chips that can be purchased in the US?
Thanks for the links and for the reply.
Expensive? 90$ buys 3000 lumens worth of chips (if you solder them yourself) which is not too expensive for D50 light source. Easiest route if you don't want minimal DIY is to get their D50 strips and put them on 2 panels with off the shelf power supply. You will get ready to use lighting setup for target shooting for 150$ and minimum DIY
-
Expensive? 90$ buys 3000 lumens worth of chips (if you solder them yourself) which is not too expensive for D50 light source. Easiest route if you don't want minimal DIY is to get their D50 strips and put them on 2 panels with off the shelf power supply. You will get ready to use lighting setup for target shooting for 150$ and minimum DIY
What is the cost of labor and buying the tools and acquiring the knowledge to soldier all those chips? You've got to be joking, right?
It's not worth it to me. I'm a hobbyist photographer who'ld rather make images with the lights affordable and available. I'm just looking for improvements in the industry that don't cost a fortune in time and money.
I appreciate your setting the boundaries on where and how that happens.
-
What is the cost of labor and buying the tools and acquiring the knowledge to soldier all those chips?
An evening watching Youtube tutorials
It's not worth it to me. I'm a hobbyist photographer who'ld rather make images with the lights affordable and available. I'm just looking for improvements in the industry that don't cost a fortune in time and money.
I appreciate your setting the boundaries on where and how that happens.
Have you really looked at links I sent? Soldering is only required is you buy naked chips - the panels (https://store.yujiintl.com/collections/vtc-d50-series/products/vtc-series-d50-5000k-high-cri-mcpcb-led-module-unit-5-pcs) as I said require no special skills - only connecting things together. At $180 + power supply this is still a lot cheaper than any other D50 light sources I have seen.
Their VTC series (LEDs with 98 CRI) also do come in COB (ceramic packages common in LED world) and as such they require no soldering (just mounting on heatsink and wiring a power supply).
BTW, I am a hobbyist as well not a professional and the reason I am looking at all this is that in photography most of the ready to use things do not come at cheap prices (there is of course a lot of cheap and not so good things) so I am finding and using alternatives. Printer profiling hardware and software for example cost a fortune especially if you use it occasionally - but researching a bit and getting used Spectroscan works quite well and on par with current solutions in quality.
-
An evening watching Youtube tutorials
Have you really looked at links I sent?
If you expect some to work at such understanding, it will likely fall on deaf ears. You offer solutions but like others, get more push back; the irony is seen in this comment:
http://forum.luminous-landscape.com/index.php?topic=123799.msg1033646#msg1033646 (http://forum.luminous-landscape.com/index.php?topic=123799.msg1033646#msg1033646)
-
If you expect some to work at such understanding, it will likely fall on deaf ears. You offer solutions but like others, get more push back; the irony is seen in this comment:
http://forum.luminous-landscape.com/index.php?topic=123799.msg1033646#msg1033646 (http://forum.luminous-landscape.com/index.php?topic=123799.msg1033646#msg1033646)
Thanks Andrew, I can only give the information in a hope that someone will find it useful.
-
Thanks Andrew, I can only give the information in a hope that someone will find it useful.
I did!
-
An evening watching Youtube tutorials
Have you really looked at links I sent? Soldering is only required is you buy naked chips - the panels (https://store.yujiintl.com/collections/vtc-d50-series/products/vtc-series-d50-5000k-high-cri-mcpcb-led-module-unit-5-pcs) as I said require no special skills - only connecting things together. At $180 + power supply this is still a lot cheaper than any other D50 light sources I have seen.
Their VTC series (LEDs with 98 CRI) also do come in COB (ceramic packages common in LED world) and as such they require no soldering (just mounting on heatsink and wiring a power supply).
Thanks for the additional info.
I'll have to pass on constructing electronic lighting with regard to heatsinks and wiring a power supply. I'm sure someone here will use this info and have enough confidence in what they're doing. I don't. Me and electricity don't get along and I now understand why there are several insignias on my Hyperikon LED's that include "UL" Underwriter's Laboratory standards and FCC with regard to frequency band regulations.
-
Thanks Andrew, I can only give the information in a hope that someone will find it useful.
Found it very interesting thanks. I am now wondering how they would perform as a relatively inexpensive print viewing booth light source compared to a GTI booth.
FWIW found this datasheet from another company with details on how to connect
https://led.cdiweb.com/Datasheets/citizen/Citizen-Instruction-Manual-COB-LED-Package.pdf
-
Thanks for the additional info.
I'll have to pass on constructing electronic lighting with regard to heatsinks and wiring a power supply. I'm sure someone here will use this info and have enough confidence in what they're doing
It would help if you actually read any of those links I posted (heatsinks mounting is only for COBs). The strips I referred to do not require any heatsinks - in fact they are pretty much the same as most of the led setups out there and those you frequently find in kitchens for example. The connectors are mechanical - plug and play style and come with the strips. Power supply if you want a simplest solution is also sold by Yuji and linked from the very same page (but here it is (https://store.yujiintl.com/collections/high-cri-led-strips-ribbon/products/flicker-free-dimmable-power-supply-48w-pack-1-pcs)). All it requires is plugging a few wires.
-
It would help if you actually read any of those links I posted (heatsinks mounting is only for COBs).
All it requires is plugging a few wires.
Too damn difficult! :P
For some....
-
Found it very interesting thanks. I am now wondering how they would perform as a relatively inexpensive print viewing booth light source compared to a GTI booth.
I am in the process of building something like that with D50 LEDs, placing them in two panels individually dimmable and using as a sides of the booth.
FWIW found this datasheet from another company with details on how to connect
https://led.cdiweb.com/Datasheets/citizen/Citizen-Instruction-Manual-COB-LED-Package.pdf
Thanks for those - they are very useful.
That Citizen COB on the pictures actually has mechanical holder from Molex (https://www.molex.com/molex/products/datasheet.jsp?part=active/1804140001_SOLID_STATE_LIGHTI.xml) that fits the VTC Yuji COB LEDs (13.5mm version). With that connector no soldering is required whatsover - just plugging in the wires.
-
I am in the process of building something like that with D50 LEDs, placing them in two panels individually dimmable and using as a sides of the booth.
Thanks for those - they are very useful.
That Citizen COB on the pictures actually has mechanical holder from Molex (https://www.molex.com/molex/products/datasheet.jsp?part=active/1804140001_SOLID_STATE_LIGHTI.xml) that fits the VTC Yuji COB LEDs (13.5mm version). With that connector no soldering is required whatsover - just plugging in the wires.
It would be great to see the finished booth and your thoughts on the build when finished if you have time :D
-
It would help if you actually read any of those links I posted (heatsinks mounting is only for COBs). The strips I referred to do not require any heatsinks - in fact they are pretty much the same as most of the led setups out there and those you frequently find in kitchens for example. The connectors are mechanical - plug and play style and come with the strips. Power supply if you want a simplest solution is also sold by Yuji and linked from the very same page (but here it is (https://store.yujiintl.com/collections/high-cri-led-strips-ribbon/products/flicker-free-dimmable-power-supply-48w-pack-1-pcs)). All it requires is plugging a few wires.
It would help me if you had read what I wrote several times that I wasn't interested in building my own Yuji light and that I would like to know if there are any companies that have the lights already built that use the Yuji chip.
Since you didn't answer the last part I'll take it that you and anyone else here don't know of any companies that's made this light at an affordable price like say under $50.
I also now remember having this same conversation in this forum several years ago and still not getting an answer. My memory of these discussions starts to fritter away once I see there's no solution.
And probably several years later this conversation will come up again and will now have to commit to memory the name Yuji as an LED light that hasn't or isn't going to built at an affordable price.
And BTW I did read those links. It would've saved me some time if I'ld known I would have to build the light myself with the components bought off that site. Again, not interested.
-
It would help me if you had read what I wrote several times that I wasn't interested in building my own Yuji light....
Naturally he never said that. Let alone several times....
-
we are going in circles now...
-
Since you didn't answer the last part I'll take it that you and anyone else here don't know of any companies that's made this light at an affordable price like say under $50.
The solution if exists out there is not going to cost $50. You do need to research LEDs because the panels, strips and COBs is by far the most popular packages they come in - not as a prepackaged panels. They are also the only cost effective solution if you want high CRI and full spectrum coverage.
Btw, Yuji does sell ready made lights (not with D50 to my understanding) here (https://store.yujiintl.com/collections/high-cri-led-lights) but their prices of course are nowhere near $50.
As with anything - all those manufacturers are happy to answer questions, so why not to write to them and ask instead of complaining here that no one answers what you want to know? As with any quest: where there's a will - there's a way
-
It would be great to see the finished booth and your thoughts on the build when finished if you have time :D
Sure, as for most of my projects full details will go on github so they can be reproduced.
The panels now taking a slight backseat though as I am in a process of replacing Xenon light source in my Spectron device (camera sensors spectral responsivity measurements) with the Yuji LEDs (4x VTC COBs in a single compact and controllable package).
I will try to make panels ready though before summer and will post the details if interested. My primary goal with them is a well lit booth for target shooting but of course that can be used for more than this. If I were to construct viewing booth from scratch, I would probably go for the prebuilt panels and Yuji power supply. This will require just mounting the panels at the top with a few screws and connecting power supply with a few wires and mechanical clips. The panels they sell will be enough for 2 booth so it could be joint purchase to spread the cost (well just a thought anyway).
-
So now asking questions on the availability of a bulb is considered complaining? WTF?! :o
-
So now asking questions on the availability of a bulb is considered complaining? WTF?! :o
Complaining for not getting an answer of course. An with that tone I doubt anyone will be willing to answer.
-
Complaining for not getting an answer of course. An with that tone I doubt anyone will be willing to answer.
+1; plus, no reason to feed the trolls!
-
Tim, don't give up... go trolls, go !!!
-
Tim, don't give up... go trolls, go !!!
It's in their nature; a lack of experience too. As seen in this series of posts:
http://forum.luminous-landscape.com/index.php?topic=124034.msg1037298#msg1037298
-
The CRI replacement that is gaining the most interest lately is IES TM-30-15 (https://www.ies.org/store/technical-memoranda/ies-method-for-evaluating-light-source-color-rendition/).
There are now LED light sources that have high CRI, TLCI and R9 values. I'm wondering if the manufacturers are gaming the system and that the bulbs/panels don't really render color as well as the CRI (et. al) specs indicate.
Will ColorMeter support this?
[Edit. It has been a while since I posted on Luminous Landscape. This was a response to a post by GWGill.]
[2nd Edit. Sorry again. I just realized that GWGill addressed this in another thread. (http://forum.luminous-landscape.com/index.php?topic=94691.new;topicseen#new)
-
Soldering is only required is you buy naked chips - the panels (https://store.yujiintl.com/collections/vtc-d50-series/products/vtc-series-d50-5000k-high-cri-mcpcb-led-module-unit-5-pcs) as I said require no special skills - only connecting things together. At $180 + power supply this is still a lot cheaper than any other D50 light sources I have seen.
Alexey, I'm jumping in here because I'm interested. Just to make sure I understand, I'd need to order one
VTC Series D50 5000K High CRI MCPCB LED Module - Pack: 10 pcs $209 (https://store.yujiintl.com/collections/vtc-d50-series/products/vtc-series-d50-5000k-high-cri-mcpcb-led-module-unit-5-pcs)
This page says
One pack (10pcs MCPCB) needs 2pcs power supply.
When choosing a power supply, please be sure to select 24V to ensure compatibility with this product.
So I'd need to order two
Yuji Flicker Free Dimmable Power Supply, 48W - Pack: 1 pcs $45 (https://store.yujiintl.com/collections/high-cri-led-strips-ribbon/products/flicker-free-dimmable-power-supply-48w-pack-1-pcs) and choose the 24V versions.
OK, but the modules come like this
(https://cdn.shopify.com/s/files/1/0344/6401/files/P7152142_large.jpg)
This shows five strips bonded together. "Pack: 10 pcs" means two of these modules, correct?
The connectors look like
(https://cdn.shopify.com/s/files/1/0344/6401/products/DSC_0967_grande.jpg)
which means a lot of pins sticking out of each module (of five strips) Don't we need some kind of jumper wire harness to connect to the power supply? (I can handle soldering, but it would be easier to start out with jumper wires that I can cut apart in the middle to solder to.)
They don't give the voltage requirements. Are the strips connected in serial or in parallel?
I'm trying to get it down to knowing what to order so I can plug everything together to mount in enclosures that I'd fabricate.
An evening watching Youtube tutorials
Suggestions on Youtube search terms?
-
I ordered bare naked LEDs not the board and do all soldering myself. The boards with presoldered LEDs you looking at seems to be an easy option in term of connecting them all. I would suggest you email them with all the questions - they so far have been answering mine promptly. I can only base my replies on what I can read on their site regarding the strips.
So I'd need to order two
Yuji Flicker Free Dimmable Power Supply, 48W - Pack: 1 pcs $45 (https://store.yujiintl.com/collections/high-cri-led-strips-ribbon/products/flicker-free-dimmable-power-supply-48w-pack-1-pcs) and choose the 24V versions.
Yes - one for each panel of 5 strips it seems.
OK, but the modules come like this
(https://cdn.shopify.com/s/files/1/0344/6401/files/P7152142_large.jpg)
This shows five strips bonded together. "Pack: 10 pcs" means two of these modules, correct?
No idea - but I would guess the picture may be for illustration and it may come as a single block or two blocks. They can be broken down to individual strips and connected into one long strip with supplied Wago straight connectors.
The connectors look like
(https://cdn.shopify.com/s/files/1/0344/6401/products/DSC_0967_grande.jpg)
which means a lot of pins sticking out of each module (of five strips) Don't we need some kind of jumper wire harness to connect to the power supply? (I can handle soldering, but it would be easier to start out with jumper wires that I can cut apart in the middle to solder to.)
The connectors on those boards are Wago 2060 series and from description the package you looking at comes with 10 straight Wago 2060-902 connectors. Those will allow to build one long strip of lights mechanically. If you want to connect them as they are in the panel, I think all you would need is a 1mm thick wire (pair) since connector is locking and wire them at the sides like you would if you connect those strips into one long one. You will notice that the panel are already done like that so connecting part of each strip are adjacent to each other.
They don't give the voltage requirements. Are the strips connected in serial or in parallel?
I'm trying to get it down to knowing what to order so I can plug everything together to mount in enclosures that I'd fabricate.
No idea but judging by power supply specification they all connected in parallel (power supply does produce voltage enough for one LED with current for 30-40 of them).
For my board/panel I am targeting 4 strips of LEDs in parallel, 12 LEDs in each strip in series.
Suggestions on Youtube search terms?
"Soldering SMD", "wiring LED COB" etc
-
"Soldering SMD", "wiring LED COB" etc
Thanks for your informative response! But I see that I phrased my question backwards. I should have limited it to asking about this. I Googled for both these phrases and the results show that the subject isn't trivial.
Starting with "wiring LED COB" It would take a lot more than an evening to be comfortable with powering up the LED modules described above. There are a lot of questions that need to be answered about connecting in parallel (namely about avoiding thermal runaway.) Parsing component data sheets into finished project schematics (let alone pictorial diagrams) isn't something that can be learned in an evening. I have a sort of head start because I was a reasonably hard core electronics experimenter in the 1970s and 80s (metric: I have an electronics workbench and test equipment. But dating from the 1980s.)
"Soldering SMD" is an even longer detour. It looks like it may be cheap, but only when you know what you are doing. I also have a sort of head start because I know people that can do this. But not well enough to get personal lessons. I looked through some of the written tutorials. I get the impression that the people who write the tutorials are very optimistic about what is "easy."
I've done enough electronics project assembly that I'm not afraid of wiring from AC to finished project. I know the safe guards. But I wouldn't recommend that a non technical person do this without guidance. In person.
I'm not afraid of being electrocuted with the LED modules. I'm afraid of hearing a loud pop and then seeing a cloud of magic smoke.
For those of us that are queasy about wiring up LED panels and power supplies, what about the finished bulbs?
BC Series A60 High CRI Remote Phosphor LED Bulb - Pack: 4 pcs $80 (https://store.yujiintl.com/collections/high-cri-led-lights/products/bc-series-a60-high-cri-remote-phosphor-led-bulb-unit-2-pcs)
I note that they don't give any spectral plots for these. They only give a bland
CRI Value (Ra typ.) Above 95
which leaves a lot of wiggle room. These are pricey but the cost would be livable. If they gave good light.
I recently found some Cree bulbs locally (https://www.homedepot.com/p/Cree-100W-Equivalent-Daylight-5000K-A21-Dimmable-Exceptional-Light-Quality-LED-Light-Bulb-TA21-16050MDFH25-12DE26-1-11/304006413) that cost less than $10 ea. for 100w (equiv.) and measure a lot better than older hardware store bulbs.
(http://www.fishcreekstudio.com/timages/01/03-Cree-100w-90-CR.png)
The CRI, TLCI, and R9 are nice. (Duv is iffy, though.) They are real easy to use. But the spectral plot is telling me that I need better. $160 worth of the above Yuji bulbs would give me enough light to do what I'm doing now with the Cree bulbs. But I don't know if the Yuji bulbs' spectrum is closer to my Cree bulbs, or is closer to the COB LEDs that you are using.
-
Thanks for your informative response! But I see that I phrased my question backwards. I should have limited it to asking about this. I Googled for both these phrases and the results show that the subject isn't trivial.
Starting with "wiring LED COB" It would take a lot more than an evening to be comfortable with powering up the LED modules described above.
I gave those as an example (since those were what I referred to in my original post) - the modules you have been looking at are not COBs and already come presoldered on boards so require only wiring to their power supply.
Regarding COBs, in reality - there is nothing complicated about LEDs and COBs specifically if all you want to do is to have a lighting system. Plenty of COB mechanical fixing packages exists that allow them to be fastened to a radiator with a few screws and won't require any soldering (just connecting wires). For COBs all you generally need to do is pick radiator(s), drill some holes, fasten the COBs to radiators and attach the wires to power supplies.
Reference to soldering SMD was given in for the "going all the way" case - soldering LED chips yourself. True, it is a bit of a learning curve but this will give the ultimate control over LED configuration package you can create.
I'm not afraid of being electrocuted with the LED modules. I'm afraid of hearing a loud pop and then seeing a cloud of magic smoke.
As with anything else - it is a matter of reading specifications for the chosen LED module and matching it to power supply connected.
But I don't know if the Yuji bulbs' spectrum is closer to my Cree bulbs, or is closer to the COB LEDs that you are using.
The page you referenced has a spec with the spectral plot. Their BTC series have lowest CRI from all their offerings and pretty much match those CREE ones with lack in blue part of the spectrum (400-420nm) and typical for these LEDs hump just after that. All that is on Yuji site - if you look at "SHOP BY CRI LEVEL" you can see all those categories in decreasing CRI order.
-
Regarding COBs, in reality - there is nothing complicated about LEDs and COBs specifically if all you want to do is to have a lighting system. Plenty of COB mechanical fixing packages exists that allow them to be fastened to a radiator with a few screws and won't require any soldering (just connecting wires). For COBs all you generally need to do is pick radiator(s), drill some holes, fasten the COBs to radiators and attach the wires to power supplies.
COBs + mechanical fixing packages may be more attractive. Drilling holes and plugging in wires is reasonable. (If I can get past that thermal runaway issue.) SMD soldering is a long detour. If fabrication is truly the only option to get usable lights.
The page you referenced has a spec with the spectral plot. Their BTC series have lowest CRI from all their offerings and pretty much match those CREE ones with lack in blue part of the spectrum (400-420nm) and typical for these LEDs hump just after that. All that is on Yuji site - if you look at "SHOP BY CRI LEVEL" you can see all those categories in decreasing CRI order.
Hmm, the spec sheet I found for the BTC series doesn't have a spectral plot (https://cdn.shopify.com/s/files/1/0344/6401/files/YJ-BC-RP-10-G01.pdf?15136077673316622282). It just shows luminosity distribution.
"SHOP BY CRI LEVEL" not informative. For instance, here is the spectral plot of an Aputure Amaran AL-H198 (https://www.bhphotovideo.com/c/product/1114779-REG/aputure_amaran_al_198_198_on_camra_daylight.html) I have that has a CRI of almost 97.
(http://www.fishcreekstudio.com/timages/01/Aputure-Amaran-AL-H198-700.png)
We can see that judging by CRI, TLCI, R9, and Duv alone is problematic. The only way to get a spectral plot was for me to buy the AL-H198 and measure it myself. The cost of these panels was low enough that I don't mind having a pair. But I'm trying to avoid getting into a "buying, measuring, and then returning" cycle, which is what I'd need to do if I wanted to find out if more expensive LED panels were really worth the additional expense. (I've gravitated back to prefabricated panels here.)
There is a gap in the luminaire industry. Photographers really shouldn't have to be reading spec sheets and moonlighting in DIY electrical engineering in order to have usable lighting. Or be independently wealthy (to afford museum quality lighting.)
However, right now photographers should be more conversant with products like ColorMeter. Getting going with ColorMeter isn't such a long detour. So we can hold big luminaire's feet to the fire.
-
It's less of a problem than it "appears" :)
The spikey characteristic of any illuminant isn't what matters directly. It's the degree of matching after the illuminant's spectral curves are combined with the print's spectral reflectance curves then the color matching functions.
The various measures of color rendering are based on a much larger variety of reflective surfaces than what exists with an inkjet print.
Printers, fortunately for us, have a much more limited spectral distribution. So the important measure of an illuminant used for prints is how closely they match D50, D55, D65, or whatever your preferred illuminant. The spectral distribution of an illuminant, especially LED types, will not correlate well with actual sensed color rendition from a print.
You can get a better sense of how different illuminants impact a print by making two custom profiles. The first made with D50 illuminant and the second with the illuminant you wish to compare the print to. Then convert to printer RGB space using the standard D50 profile. Next, select "assign profile" but don't hit OK, and toggle the "view" box to see the difference in how the print will look under one illuminant compared to D50.
There are ways to do this quantitatively as well but the point here is that it's really hard to tell how well an illuminant will work for a print looking at the illuminant spectrum.
-
It's less of a problem than it "appears" :)
The spikey characteristic of any illuminant isn't what matters directly. It's the degree of matching after the illuminant's spectral curves are combined with the print's spectral reflectance curves then the color matching functions.
The various measures of color rendering are based on a much larger variety of reflective surfaces than what exists with an inkjet print.
I'm not illuminating prints. I'm shooting paintings. (And eBay pictures.) I don't think that painting pigments are as predictable as broad spectrum inkjet print colors, are they?
This is my second go round shooting family paintings. For my first attempt I used "full spectrum" daylight balanced CFLs. I never could get accurate color, no matter how many Adobe DNG Profile Editor camera profiles I made. So I'm a bit leery of illuminant spikes.
-
I'm not illuminating prints. I'm shooting paintings. (And eBay pictures.) I don't think that painting pigments are as predictable as broad spectrum inkjet print colors, are they?
This is my second go round shooting family paintings. For my first attempt I used "full spectrum" daylight balanced CFLs. I never could get accurate color, no matter how many Adobe DNG Profile Editor camera profiles I made. So I'm a bit leery of illuminant spikes.
Yep, colorimetric rendering of paintings is harder. Much harder. Paint pigments are much more diverse spectrally than inkjet prints. So they are impacted by both the illuminant and the camera. And so-called "full spectrum" CFLs are pretty bad to start with. And so is the camera. You can always take pictures of the paintings in daylight but that won't fix the camera's divergence from Luther/Ives.
As for the spike at 450nm and dip at 480, these both affect mostly the Z in XYZ colorspace and only where there are significant reflectance differences in 450-500nm transitioning in the painting pigments over the range of 440nm to 520nm. One way to quantify it would be spectrally sampling colors in the painting and comparing calculated Lab values for the led v D50. It's a major hassle to do.
-
I'm not illuminating prints. I'm shooting paintings. (And eBay pictures.) I don't think that painting pigments are as predictable as broad spectrum inkjet print colors, are they?
This is my second go round shooting family paintings. For my first attempt I used "full spectrum" daylight balanced CFLs. I never could get accurate color, no matter how many Adobe DNG Profile Editor camera profiles I made. So I'm a bit leery of illuminant spikes.
Thanks for the info and links on the Yuji and Cree bulbs, Wayne. That was very helpful.
My local Home Depot indicates they have 15 of the Crees. I'm going to check them out and see if they render Cadmium Yellow as Lemon Yellow (bad CRI) shooting and AWB'ing a WhiBaL card and whether the Xrite CCchart cyan renders as just another sky blue. I'm also going to test how they render the whites of my eyes shooting a selfy. The Soraa Vivid LED I have is the best at this but finding it hard to buy a second one online of the exact model. The Hyperikons make the whites of my eyes overly blue. Most of the other colors look pretty close but they employ a trick manipulating the color of white by making it slightly greenish desaturated yellow which of course will add magenta to other cooler colors AWB a neutral target.
It's never a perfect rendering. I always have to manipulate WB against HSL panel in ACR. It's never a one off and I'm done.
-
COBs + mechanical fixing packages may be more attractive. Drilling holes and plugging in wires is reasonable. (If I can get past that thermal runaway issue.) SMD soldering is a long detour. If fabrication is truly the only option to get usable lights.
Soldering is only if you want to go down that road. The packages you were looking for (panels) require no soldering. I am struggling to understand the logic really. Yuji site has comprehensive selection of various packages for pretty much different installation requirements - bare LED chips (the ones you need to solder), COBs (where available), prearranged panels with LED chips already soldered on PCB and flexible strips (I am not considering various lamps at this moment). The latter two are by far the easiest to connect - a few wires, connectors to be connected to power supply and you are done.
Hmm, the spec sheet I found for the BTC series doesn't have a spectral plot (https://cdn.shopify.com/s/files/1/0344/6401/files/YJ-BC-RP-10-G01.pdf?15136077673316622282). It just shows luminosity distribution.
They do but not for that lamp. Yuji have 3 categories of LEDs - BTC, VTC and D50 (in order of increasing CRI). Select any chip in that category (COB for example) and spec there will contain technical details including spectral curves. Here for example https://cdn.shopify.com/s/files/1/0344/6401/files/YJWJ007-1.2_YJ-BC-135L-G02.pdf?1044822003027063112
Same for VTC and D50 series.
"SHOP BY CRI LEVEL" not informative. For instance, here is the spectral plot of an Aputure Amaran AL-H198 (https://www.bhphotovideo.com/c/product/1114779-REG/aputure_amaran_al_198_198_on_camra_daylight.html) I have that has a CRI of almost 97.
I was referring to Yuji site not anybody elses and how informative it will be in all other cases. Yuji tend to structure their LEDs in categories by CRI and within those by CCT. I had a look through pretty much all of those in search of the spectral response that I need vs light output and its a good starting point.
As I also said before - do ask them. In my view it is a waste of time to hypothesize what may or may not be the case - much simpler just to ask Yuji or any other manufacturer for spectral plots if that is what you need.
-
To somewhat clarify the questions going around what is a module and how to connect it:
- Modules are a string of individual LEDs connected in series, with a resistor connected in series with each string.
- The resistor does two things - firstly, limits current to the string, and secondly, prevents thermal runaway if you connect modules in parallel.
- 12V modules usually have three LEDs in series, 24V modules 6 or 7, depending on LED color, type, etc. On the photos in previous posts, the LEDs are the yellow blocks, the resistors the black blocks
- The amount of current drawn depends on the module power - you have to look that up on the data sheet.
- You can wire as many modules in parallel as you want, subject to the power supply being able to provide enough current.
- Usually, the modules are internally wired such that if you just connect them end-to-end, that gives you an parallel connection.
- If you want constant intensity, any old DC power supply of the right voltage will do. If want to dim LED modules, things get a bit more complicated.You need a power supply that can vary current smoothly, not voltage, such as the one linked to in a previous post.
I don't recommend trying to deal with individual LEDs without specialist equipment and knowledge. But anyone that's reasonably competent with wiring should be ok with modules.
-
12V modules usually have three LEDs in series, 24V modules 6 or 7, depending on LED color, type, etc. On the photos in previous posts, the LEDs are the yellow blocks, the resistors the black blocks
Yuji D50 LED strips referenced above have individual LEDs running at 3.3-3.4V with LED nominal current 120mA (at these specifications their output and CRI corresponds to the specified one). The specification of the individual D50 LED is here https://cdn.shopify.com/s/files/1/0344/6401/files/YJWJ011-1.1_YJ-VTC-5730-D50-G01.pdf?5186286384022133237
If you want constant intensity, any old DC power supply of the right voltage will do. If want to dim LED modules, things get a bit more complicated.You need a power supply that can vary current smoothly, not voltage, such as the one linked to in a previous post
It is not that simple for colour critical applications since for the LEDs to stay tightly within their specs the power supply needs to be constant current power supply. That is hardly any old DC one :). But generally this is true for a strip solution yes.
-
It is not that simple for colour critical applications since for the LEDs to stay tightly within their specs the power supply needs to be constant current power supply. That is hardly any old DC one :). But generally this is true for a strip solution yes.
<non-techies tune out now>
Yes, that's true - a constant current supply would be marginally better. But practically, the bin variation (aka the variation you'd get from module to module) is more than anything that you'd get from temperature variation. (Temperature variation is the only reason why current would vary given a constant voltage supply - temperature variation causes variation in the LED forward voltage, which in turn causes current variation)
<non-techies can tune in again now> ;)
-
Thank you sandymc and Alexey!
It is not that simple for colour critical applications since for the LEDs to stay tightly within their specs the power supply needs to be constant current power supply. That is hardly any old DC one :). But generally this is true for a strip solution yes.
Does the Yuji Flicker Free Dimmable Power Supply, 48W (https://store.yujiintl.com/collections/high-cri-led-strips-ribbon/products/flicker-free-dimmable-power-supply-48w-pack-1-pcs) qualify?
I'm looking at making lights for family memories reproduction work. Assuming that the Yuji modules and power supply have reduced the problem to plug-and-play, the next step is figuring out how to fabricate some semblance of
https://dtdch.com/dt-photon-custom-cultural-heritage-lighting/ (https://dtdch.com/dt-photon-custom-cultural-heritage-lighting/)
Looking that site over, I need just about everything. (But not at those prices. This means going to Home Depot, walking past the Cree bulbs and looking in the lighting fixture, electrical, and building materials departments.)
However...earlier Doug discussed sidestepping the camera's Luther/Ives divergence (http://forum.luminous-landscape.com/index.php?topic=123799.msg1033821#msg1033821) by using the variable illuminants used in the IES TM-30-15 color rendition method (https://www.ies.org/store/technical-memoranda/ies-method-for-evaluating-light-source-color-rendition/)
Any ideas on fabricating these kinds of illuminants? Or would this take me down a very deep rabbit hole? (Should I forget about Luther/Ives divergence and stick to emulating D50 as best I can?)
-
In my view, the Yuji power supply looks fine.
Also in my view, Luther/Ives is in of itself a massive rabbit hole. But I know that others on this forum would violently disagree.
Thank you sandymc and Alexey!
Does the Yuji Flicker Free Dimmable Power Supply, 48W (https://store.yujiintl.com/collections/high-cri-led-strips-ribbon/products/flicker-free-dimmable-power-supply-48w-pack-1-pcs) qualify?
I'm looking at making lights for family memories reproduction work. Assuming that the Yuji modules and power supply have reduced the problem to plug-and-play, the next step is figuring out how to fabricate some semblance of
https://dtdch.com/dt-photon-custom-cultural-heritage-lighting/ (https://dtdch.com/dt-photon-custom-cultural-heritage-lighting/)
Looking that site over, I need just about everything. (But not at those prices. This means going to Home Depot, walking past the Cree bulbs and looking in the lighting fixture, electrical, and building materials departments.)
However...earlier Doug discussed sidestepping the camera's Luther/Ives divergence (http://forum.luminous-landscape.com/index.php?topic=123799.msg1033821#msg1033821) by using the variable illuminants used in the IES TM-30-15 color rendition method (https://www.ies.org/store/technical-memoranda/ies-method-for-evaluating-light-source-color-rendition/)
Any ideas on fabricating these kinds of illuminants? Or would this take me down a very deep rabbit hole? (Should I forget about Luther/Ives divergence and stick to emulating D50 as best I can?)
-
Yes, that's true - a constant current supply would be marginally better. But practically, the bin variation (aka the variation you'd get from module to module) is more than anything that you'd get from temperature variation. (Temperature variation is the only reason why current would vary given a constant voltage supply - temperature variation causes variation in the LED forward voltage, which in turn causes current variation)
The D50 LEDs that I ordered came from a singe bin. From my experience so far - the variations in the current do get chromacity shifts. It is also a reason why I don't use power supplies with analogue dimming that just vary current. The power supplies that sense and control the current in LED chain in my view are far more stable approach than chain of LEDs with a resistors and constant voltage DC power supply.
-
Does the Yuji Flicker Free Dimmable Power Supply, 48W (https://store.yujiintl.com/collections/high-cri-led-strips-ribbon/products/flicker-free-dimmable-power-supply-48w-pack-1-pcs) qualify?
No idea - since it is powering LED strips (with resistors) I would guess it is not the constant current power supply. I build my own but this approach of course is not for everyone.
Their supply though is perfectly fine to drive their LED strips and panels - it will still give you a much better light source than other off the shelf solutions in the same price range. If you look at utmost quality and stability then I am afraid no off the shelf components do exist in the price range Yuji offers - unless you do build them yourself (I have chosen that road).
-
No idea - since it is powering LED strips (with resistors) I would guess it is not the constant current power supply. I build my own but this approach of course is not for everyone
Resistors in series with a constant voltage supply is commonly used to drive LEDs. Selecting the right resistor and supply voltage will automatically compensate for LED output with temp drift since the forward voltage decreases with temp.
-
Resistors in series with a constant voltage supply is commonly used to drive LEDs. Selecting the right resistor and supply voltage will automatically compensate for LED output with temp drift since the forward voltage decreases with temp.
I know they do as well the drawbacks of this approach (as well as the fact that its the cheapest way to drive strings of LEDs). Variations of resistance, variations caused by temperature makes the current through the LEDs not that well controlled as in constant current approach.
-
I know they do as well the drawbacks of this approach (as well as the fact that its the cheapest way to drive strings of LEDs). Variations of resistance, variations caused by temperature makes the current through the LEDs not that well controlled as in constant current approach.
Voltage driven LEDs can cause thermal runaway at higher levels because the forward voltage drops as the LEDs heat up. Constant current approaches completely avoid thermal runaway but the right resistor, in series with the LEDs is very good. As the LEDs heat up and the forward voltage drops, the voltage across the resistor increases and that increase is also seen as increased current to the LEDs. Goal is to increase the current just enough to compensate for the lower efficiency as the LEDs heat up. By balancing the resistor and voltage, one can compensate quite well for the drop in light that would otherwise occur with a constant current source.
Sometimes the cheapest way is the best way.
-
Sometimes the cheapest way is the best way.
Its a good thing manufacturers of the LED monitors do not agree with you and use constant current for LED backlighting ;)
Driving LEDs with resistor as current limiter and constant voltage does nothing to control chromacity variations as current varies in that case. It also does not help that current and voltage through LEDs have non-linear relationship - tiny variations in voltage can cause substantial shift in current and changes in chromacity. For kitchen lighting it may be ok, for target shooting and profiling - well, I would not use anything but constant current power supply. Less variables to control is always better.
There are quite a few articles on a subject out there for those interested in technical aspects - for example this https://www.maximintegrated.com/en/app-notes/index.mvp/id/3256
For my application driving four Yuji VTC 5600K LED COBs with constant current supply on full output at rated current (well below maximum incidentally) - there is almost no variation in spectral output (there is of course some variation as LEDs get hotter to the working temperature which stabilizes after first 10 mins of running). I did periodic spectral measurements over 35 mins period starting from turning the LEDs on from cold.
-
Interesting read up on the thermal dynamics within LED lighting constructs. I'm sure some of the members of Underwriter's Laboratory (UL) would probably find this interesting as well.
But as a photographer, and this is a photography forum, I'm sure we are now all in agreement that build your own Yuji LED full spectrum lighting is not as easy as Alexey points out.
And I'ld like to add for the photographers interested that the Cree 100w equivalent, 1600 Lumens, 90+ CRI, 5000K bulbs I bought at Home Depot yesterday are about the same in color rendering as the Walmart Great Value branded 5000K, 60w's at 800 Lumens. The Crees get VERY HOT after a while.
Cadmium Yellow turns Lemon Yellow. WhiBal gray is a slight greenish yellow.
-
But as a photographer, and this is a photography forum, I'm sure we are now all in agreement that build your own Yuji LED full spectrum lighting is not as easy as Alexey points out.
Nothing is easy until you actually try it :)
The solution I am building with constant current is comparable and aiming more towards high end solutions. In this aspect it is far easier than paying small fortune for the equivalent.
There is very little point in comparing it to LED bulbs as those offer neither predictable and repeatable lighting nor the control over brightness and chromacity. If comparing those then yes Yuji LEDs solutions are very comparable and easy to build (and due to spectrum will be better).
There is of course always a compromise - for example attempting to find a constant current supply (readily made) that matches the existing panels/LEDs. Practically no soldering (perhaps to de-solder resistors from panels). I considered going with this approach when building LED light from COBs but in the end doing my own solution gave me better control over the whole circuit.
An example of this ready to use approach is this constant current power supply (https://uk.rs-online.com/web/p/led-drivers/9091364/) (one of many) matching Yuji VTC 5600K COB LEDs (https://store.yujiintl.com/collections/vtc-series/products/vtc-series-high-cri-cob-led-135l-pack-5-pcs) (this power supply should be enough to support 5 LEDs in series - that's 3000lm). All that remains is attaching them onto radiator with these Molex connectors (https://www.mouser.co.uk/ProductDetail/Molex/180414-0001?qs=sGAEpiMZZMvuxEGc54HwjfX8RoxiEYdwojo%2fi0fbVqo%3d). No soldering whatsoever and gives you quality LED driver.
-
And I'ld like to add for the photographers interested that the Cree 100w equivalent, 1600 Lumens, 90+ CRI, 5000K bulbs I bought at Home Depot yesterday are about the same in color rendering as the Walmart Great Value branded 5000K, 60w's at 800 Lumens.
When I lasted tested hardware store quality LED bulbs several years ago, CRI was in the low 80s, including, Walmart bulbs. But I hadn't tried Walmart bulbs lately. I'll pick some up and test them. Recent Home Depot house brand LEDs were in the mid 80s (but rendered skin tone in paintings very well.)
Did you measure CRI? Or are you going by how well the colors you are interested in were captured?
The Crees get VERY HOT after a while.
Cadmium Yellow turns Lemon Yellow. WhiBal gray is a slight greenish yellow.
Duv measured on the high side for the Crees. Duv is the amount of tint away from the Planckian (white) locus.
What I'm interested in is if the increase in measured CRI for cheap bulbs (was low 80s several years ago, now (Cree anyway) is in the 90s) correlates with colors being captured more accurately.
Otherwise, what is the point of measuring CRI (etc.)?
-
An example of this ready to use approach is this constant current power supply (https://uk.rs-online.com/web/p/led-drivers/9091364/) (one of many) matching Yuji VTC 5600K COB LEDs (https://store.yujiintl.com/collections/vtc-series/products/vtc-series-high-cri-cob-led-135l-pack-5-pcs) (this power supply should be enough to support 5 LEDs in series - that's 3000lm). All that remains is attaching them onto radiator with these Molex connectors (https://www.mouser.co.uk/ProductDetail/Molex/180414-0001?qs=sGAEpiMZZMvuxEGc54HwjfX8RoxiEYdwojo%2fi0fbVqo%3d). No soldering whatsoever and gives you quality LED driver.
Thank you again. I know you are frustrated but, until this post, this information was illusive. For example, your earlier posts referenced constant current power supplies but I didn't know if you were, for example, making a CC supply for each LED (rather than connecting them in series. Or if connecting in series caused problems.) Or that prepackaged CC supplies and COB connectors were readily available. (Why didn't Yuji have CC supplies instead of CV supplies? Why didn't they offer COB connectors?)
I know that this component sourcing information must seem elementary for you but it isn't for those of us whose electronics skills are rusty.
An open question: Have other forum participants built lighting as Alexey has described? If so, could you share any construction tips?
-
Thank you again. I know you are frustrated but, until this post, this information was illusive.
I am frustrated by the people spending time posting here about how difficult that is and complaining that no cheap solution exists (not referring to anyone specific here but rather generally) where a fraction of the time spent on writing those posts could be spent to research the subject. These days information is at everybody's fingertips - it took me less than 2 mins in Google yesterday to find the power supply and connectors for COBs and I was not searching for some special terminology (LED, COBs, constant current LED power supplies etc).
Yuji as well as any other LED manufacturers provide enough packages to cater for various levels of application. All you need to do is look at what's there, research a bit: pros and cons of various solutions, whether the cons outweigh for you the pros and then select the target LED package and the rest follows from there (power supplies, necessary skills if needed, youtube searches etc).
LEDs by far the easiest in this kind of research (try to build a lighting solution with pulsed point Xenon arc lamps for example to see what really difficult lighting looks like) since the information is out there in spades.
For example, your earlier posts referenced constant current power supplies but I didn't know if you were, for example, making a CC supply for each LED (rather than connecting them in series. Or if connecting in series caused problems.)
You didn't ask. The constant current supplies are done with the goal of LEDs connected in series simply because the current regulation will be met this way for each LED regardless of their small variation in forward voltages. Connecting them in parallel means that unless they are precisely aligned there will be always one LED that passes more current than the other and you still need to balance them with resistors. Basically connecting them in parallel should only be done if the power supply has multiple channels. For example, VTC COB LEDs that I am using have 450mA per LED at 19-21V, my power supply can provide up to 50V at 450mA and has 2 channels individually controlled. So my LED wiring is two strings of two LEDs in series - one string on each power supply channel.
If you go down the COB route - the most important thing to remember is that COB packages do need good heat dissipation (they can produce substantial amount depending on a COB). So unlike panels from lost of small LED chips where board itself may provide enough heat dissipation, COBs do need a heatsink (be mounted one one).
Or that prepackaged CC supplies and COB connectors were readily available. (Why didn't Yuji have CC supplies instead of CV supplies? Why didn't they offer COB connectors?)
None of the LED manufacturers offer connectors simply because the cannot cater for every single application. A typical search for the LED COB connector reveals that Molex (a most common connector manufacturer) has lots of variations. Short research from there using the size of the COBs (from spec) gives you a couple of variants to try. And they are really cheap to try and see what fits.
I know that this component sourcing information must seem elementary for you but it isn't for those of us whose electronics skills are rusty.
There are no electronic skills involved in searching for mechanical connection. A little research into how LED work is needed to make necessary connections. I had no electronic skills (apart from soldering an occasional wire to the battery connector) at all before a few years ago when I decided to build my spectral measurement device. Now I do have some rudimentary skills (and most of that is just research and practice) but none of those were needed in search for LED COBs attachments.
-
When I lasted tested hardware store quality LED bulbs several years ago, CRI was in the low 80s, including, Walmart bulbs. But I hadn't tried Walmart bulbs lately. I'll pick some up and test them. Recent Home Depot house brand LEDs were in the mid 80s (but rendered skin tone in paintings very well.)
Did you measure CRI? Or are you going by how well the colors you are interested in were captured?
Duv measured on the high side for the Crees. Duv is the amount of tint away from the Planckian (white) locus.
What I'm interested in is if the increase in measured CRI for cheap bulbs (was low 80s several years ago, now (Cree anyway) is in the 90s) correlates with colors being captured more accurately.
Otherwise, what is the point of measuring CRI (etc.)?
I have no instruments to measure CRI. Like you I'm interested in very good reproduction of color appearance in a controlled lighting environment (studio work) where I can shoot a scene containing a wide variety of artificially and naturally colored objects lit under these "Daylight" balanced LED bulbs and do the least amount of editing in order to develop a turnkey process. So far it's not possible but some of these LED lights get pretty close excluding metameric failure of the object being photographed.
I haven't tested this on artist's paints which might have the potential for metameric failure no amount of spectrally accurate light will fix. Most of the subjects lit by these daylight light bulbs photograph well enough but do require HSL/White Balance and some SplitTone edits (including some Adjustment Brush work) in ACR to get it to look right but then that's not a turnkey process and not very efficient.
I really don't need instruments to tell me when something doesn't look right. It really doesn't make any sense since I have to edit every image I shoot anyway even lit by the sun. But what I've seen, your 80 CRI measurements seem in the ballpark but I'ld favor maybe a bit under 90. But I don't have any visual to relate the scaling of the numbers to color appearance matching so I don't find any use in measuring instruments.
I have enough trouble visualizing Delta E differences. Different colors are more pronounced than others when they look off but the number doesn't indicate how much so I don't find that system very useful.
-
I am frustrated by the people spending time posting here about how difficult that is and complaining that no cheap solution exists (not referring to anyone specific here but rather generally) where a fraction of the time spent on writing those posts could be spent to research the subject. These days information is at everybody's fingertips - it took me less than 2 mins in Google yesterday to find the power supply and connectors for COBs and I was not searching for some special terminology (LED, COBs, constant current LED power supplies etc).
You're referencing me and you know it. I don't take that personally though because I respect those who have that level of technical expertise.
I just think you're wasting your time here. Like you, I see something I say something. You don't seem to grasp the more you post explaining how to build your own LED lights the more complicated it appears. And as a musician I do understand some electronics since I installed my own sound system in my car and wired my two 10in. subwoofers in series to my 250w Alpine amp in order to raise its crossover max peak on the low end because the subs were rated to reproduce signals down to 10Hz meaning peak (where most of kick bass resides around 50Hz) would barely be heard if I wired them in parallel (at 2 ohm impedance instead of 4), but that's DC current so I won't add any more complexity to this subject. See, I know some technical electronic stuff, too, but I know when I've gotten too deep to make it worth while. Wiring for AC to drive DC current is above my pay grade.
This is a photography forum, not a structural engineering forum. Every post you make explaining the intricacies of electronic regulatory issues with resistors, power supplies and wiring in series verses parallel makes your comments useless to those that spend most of their time making images. I'm one of those photographers and I've got over 3000 edited Raw images under my belt and my car's audio system sounds incredible especially digitally recorded wide dynamic range EDM (Electronic Digital Music).
I'm frustrated that you aren't making any headway in convincing a whole bunch of photographers to build their own full spectrum LED lighting on the cheap. But maybe if you keep explaining, it will come across a lot more simple outside of entertaining the retired engineers and technical folks who like to collect and talk about gear instead of making images.
-
You're referencing me and you know it. I don't take that personally though because I respect those who have that level of technical expertise.
Not just to you
This is a photography forum, not a structural engineering forum.
And I am not one either - I write software.
Every post you make explaining the intricacies of electronic regulatory issues with resistors, power supplies and wiring in series verses parallel makes your comments useless to those that spend most of their time making images.
Every? I have not started about resistors even - someone else did. When discussion goes that way I can support it. What I did start with were simple things (perhaps you just fail to read them).
The thing is - I need nothing of this. I already built my lighting systems (not one)and it works. I was merely responding here to the statement that LED lighting is rubbish - it is not and provided the reference. The technical bits came out after people were asking questions. It is completely fine with me not to answer any of those so I will stop from now on.
-
But I did start with simple things - you just fail to read them.
I did read them and my understanding of electronics tells me there's nothing simple about the subject. Wayne pretty much backed that up with his questions which was just as much over my head.
I have gotten shocked working on simple AC electronic components in my youth. That cured me of ever attempting a DIY project involving AC wiring no matter how simple it may appear. It just makes me nervous.
-
The thing is - I need nothing of this. I already built my lighting systems (not one)and it works. I was merely responding here to the statement that LED lighting is rubbish - it is not and provided the reference. The technical bits came out after people were asking questions. It is completely fine with me not to answer any of those so I will stop from now on.
Can you post an unedited photo shot of a scene having a wide variety of colored objects lit by your DIY Yuji lights? And if you have to edit, just mention them. Please include a Cadmium Yellow colored object. If you don't know what that looks like, Google it. That color is dominant in a lot of artist's paints and paintings especially sunsets.
Lemon Yellow ruins the effect and in digital photos of sunsets indicates clipping in one or more channels.
-
Can you post an unedited photo shot of a scene having a wide variety of colored objects lit by your DIY Yuji lights?
Nop sorry. My current lighting system is inside spectral measurements device. Eventually (when I build next one) I can do that for something like Color Checker DC.
-
Nop sorry. My current lighting system is inside spectral measurements device. Eventually (when I build next one) I can do that for something like Color Checker DC.
I take it then you didn't even bother to read the original OP which I'll quote...
The DXOmark use a LED lightbox, the Imaging Resource HMI bulbs for the "skylight" component in the "sunlit" tests, which IMO renders the results invalid and defies the purpose of the color accuracy measurement. Any peaky spectrum is a NO-NO in my book no matter the CRI score.
What do you guys think?
So basically you've been off topic from the get go. I thought your Yuji lights were for lighting scenes similar to the LED lightbox mentioned in the OP.
-
I take it then you didn't even bother to read the original OP which I'll quote...
So basically you've been off topic from the get go. I thought your Yuji lights were for lighting scenes similar to the LED lightbox mentioned in the OP.
And you came to this conclusion how exactly?
-
And you came to this conclusion how exactly?
Your Yuji lights are inside measuring instruments. So you took DXOmark and Imaging Resource as using measuring instruments that use spikey illuminant to light what to test and judge DSLR reproduction of color?
What are these two online sites testing using LED/HMI lights? Maybe you don't even read the title of the OP.
-
Your Yuji lights are inside measuring instruments. So you took DXOmark and Imaging Resource as using measuring instruments that use spikey illuminant to light what to test and judge DSLR reproduction of color?
Have you even read my replies to OP? Or you just inventing things as you go along?
Anyway - my last attempt to put some reason into all that.
OP has asked essentially this (question in bold):
The DXOmark use a LED lightbox, the Imaging Resource HMI bulbs for the "skylight" component in the "sunlit" tests, which IMO renders the results invalid and defies the purpose of the color accuracy measurement. Any peaky spectrum is a NO-NO in my book no matter the CRI score.
What do you guys think?
To which my answer was that LEDs (referenced here as one of the light sources using high CRI I presume as one of their characteristics) should not be dismissed because there are so many varieties now. Then later I posted a spectral reference for Yuji LED lights that I have used in my spectral measurements apparatus. I have made no claims whatsoever regarding the quality of equipment DXO or IR are using or relation of Yuji to them.
Now that was in March when my LED light was still being developed so I was able to test it out of the box. I don't just spend time handing on these forums so since March I have built and tested that so these are now installed inside my apparatus (it was referenced in dcamprof thread here if you don't know what I am referring to) - to take them out means I will have to disassemble all again. And frankly I have no idea what me shooting random scene and providing raw file will actually give you (you may or may not have the camera I have, have no spectral responses for my sensor and no profiles built for it with Yuji lights) so to me taking what I have done so far apart to satisfy just your curiosity seems an exercise in futility. As I said, if anyone will need a reference I will do so later when I build enhanced version of my board with the lights and will provide the CC DC shot under them.
-
I have no instruments to measure CRI. Like you I'm interested in very good reproduction of color appearance in a controlled lighting environment (studio work) where I can shoot a scene containing a wide variety of artificially and naturally colored objects lit under these "Daylight" balanced LED bulbs and do the least amount of editing in order to develop a turnkey process. So far it's not possible but some of these LED lights get pretty close excluding metameric failure of the object being photographed.
I haven't tested this on artist's paints which might have the potential for metameric failure no amount of spectrally accurate light will fix. Most of the subjects lit by these daylight light bulbs photograph well enough but do require HSL/White Balance and some SplitTone edits (including some Adjustment Brush work) in ACR to get it to look right but then that's not a turnkey process and not very efficient.
I really don't need instruments to tell me when something doesn't look right. It really doesn't make any sense since I have to edit every image I shoot anyway even lit by the sun. But what I've seen, your 80 CRI measurements seem in the ballpark but I'ld favor maybe a bit under 90. But I don't have any visual to relate the scaling of the numbers to color appearance matching so I don't find any use in measuring instruments.
I'm also trying to find the value of measuring instruments. I just bought and measured a Walmart daylight LED
(http://www.fishcreekstudio.com/timages/01/03_Walmart%20Great%20Value%20daylight%20LED-700.png)
75w (equiv.) Walmart Great Value daylight LED 1100 lumens (less than $4 from Walmart)
And to recap
(http://www.fishcreekstudio.com/timages/01/03-Cree-100w-90-CR-700.png)
Cree 100w (equiv.) daylight bulb 1600 lumens (about $9.60 from Home Depot)
(http://www.fishcreekstudio.com/timages/01/Aputure-Amaran-AL-H198-700.png)
Aputure Amaran AL-H198 approx. 2000 lumens (About $58 from B&H)
We've already seen prices and lumen output from the DIY Yuji LEDs.
I displayed the above plots in order of (supposedly) increasing color rendering accuracy. But how does the measured rating correlate with real world color rendering? Tim, you said that you found the Cree LEDs to be about the same as the Walmart LEDs. But the Cree LEDs measure a lot better. The H198s measure even better (and are available pre-assembled from B&H.)
The whole reason I bought ColorMeter and my ColorMunki was to test lights to see which affordable ones had the best color rendering. I was excited when I measured the Cree LEDs several weeks ago because I thought that finally bulbs from Home Depot were close enough to "good" (TLCI of almost 97) so they could be used to light art with minimal compromise. But you are reporting otherwise.
So I remain bewildered.
-
I have gotten shocked working on simple AC electronic components in my youth. That cured me of ever attempting a DIY project involving AC wiring no matter how simple it may appear. It just makes me nervous.
The power supplies that Alexey has recommended are preassembled. Any AC is hidden inside the power supply. The only voltage you'd be exposed to is 24 volts DC.
I attempt to sidestep any problems by always having the power supply plugged into an outlet strip with a pilot light and switch and keeping the outlet strip turned off while I'm wiring. I only turn the outlet strip on when I'm done wiring.
As I said earlier, I'm not worried about being electrocuted (because the AC is hidden inside the power supplies), I'm worried about thermal runaway (meaning that the LEDs self destruct, which is a concern when the LEDs in question cost several hundred dollars.)
Fabrication is also a concern. After thinking about it some, fabrication using the led modules (panels) and Yuji power supplies would be easiest. Assuming that the LED modules don't run too hot, they could be mounted on white foam core board and then suspend the foam core the same way that scrims are suspended. And have additional real scrims in front of them for diffusion. If the panels run too hot to mount on foam core, then use Masonite that is painted white and use more clamps and light stands. The power supplies could be mounted on the light stand and use some kind of plug and socket to connect the power supplies to the LED panels.
Fabricating a workable enclosure for COB LEDs mounted on heavy heat sinks sounds more difficult.
-
Fabrication is also a concern. After thinking about it some, fabrication using the led modules (panels) would be easiest. Assuming that they don't run too hot, they could be mounted on white foam core board and then suspend them like scrims. And have additional real scrims in front of them for diffusion. If they run too hot to mount on foam core, then use Masonite that is painted white and use more clamps and light stands. The power supplies could be mounted on the light stand and use some kind of plug and socket to connect the power supplies to the LED panels.
What type of lighting are you constructing? Perhaps that is starting point for the selection? The light size, power, uniformity?
The COBs as I said are the easiest to assemble but they are powerful point (9mm diameter of light emitting surface) light source so if light panel is needed then something like this (https://www.youtube.com/watch?v=N-gNAJZ-GMM) or this (https://www.youtube.com/watch?v=elBzBTtE78I) for mounting would keep them cool. And the diffuser in front.
Panels in most cases just need aluminium backing but should be ok as is (they usually contain a set of low powered LEDs).
I needed a point light source for my application so mounted 4 COBs adjacent to each other on CPU heatsink with reflector (see attachment).
-
I'm also trying to find the value of measuring instruments. I just bought and measured a Walmart daylight LED
75w (equiv.) Walmart Great Value daylight LED 1100 lumens (less than $4 from Walmart)
And to recap
We've already seen prices and lumen output from the DIY Yuji LEDs.
I displayed the above plots in order of (supposedly) increasing color rendering accuracy. But how does the measured rating correlate with real world color rendering? Tim, you said that you found the Cree LEDs to be about the same as the Walmart LEDs. But the Cree LEDs measure a lot better. The H198s measure even better (and are available pre-assembled from B&H.)
The whole reason I bought ColorMeter and my ColorMunki was to test lights to see which affordable ones had the best color rendering. I was excited when I measured the Cree LEDs several weeks ago because I thought that finally bulbs from Home Depot were close enough to "good" (TLCI of almost 97) so they could be used to light art with minimal compromise. But you are reporting otherwise.
So I remain bewildered.
Would you be less bewildered with the idea that the accuracy of light measuring instruments seems to be related to how much they cost?
I mean my LG LED display was calibrated at the factory by $10,000 Minolta Color Analyzer. What makes it more expensive than the equipment you used to measure those bulbs which I'm assuming costs far less? Could it be it's more accurate at defining D50 which as I've said about shooting quite a few daylight scenes there is a prominent fuchsia/magenta component in actual daylight that flash tries to emulate with mercury content.
Last night I gave the Crees a second look and they're worse than the Walmart 60w. And now I'll post how it renders red, blue and cadmium yellow against 5PM direct sunlight outside my apartment.
-
And this is how the Cree renders color by clicking for R=G=B on the WhiBaL Gray area...
-
Oops! I've never seen this happen with these types of bulbs but the Cree is shifting in color temp hue to less green and more toward slight orange the longer I leave it on. My camera's AWB recorded the first color temp in ACR as 4350/tint +22.
From the time I posted the first shot to now my camera's AWB records it as 4500/tint +11. See below how far it drifted away from green.
-
What type of lighting are you constructing? Perhaps that is starting point for the selection? The light size, power, uniformity?
Ideally, a pair of fixtures like the DT Photons for copy stand work (large documents, camera scanning) and a pair of DT Photon XLs for lighting artwork. Both from Digital Transitions (https://dtdch.com/dt-photon-custom-cultural-heritage-lighting/). (In my dreams.)
But in the real world I've been making do with hardware store clamp fixtures and hardware store bulbs (first "full spectrum" CFLs and now "high CRI" LED bulbs). When shooting art I use cross polarization which sucks up a lot of light. I need about 600 watts (equiv.) to keep shutter speeds in the one or two second range when shooting base ISO. Six clamp fixtures (5 inch reflectors) give me enough uniformity for my standards.
The Aputure Amaran AL-H198 (https://www.bhphotovideo.com/c/product/1114779-REG/aputure_amaran_al_198_198_on_camra_daylight.html)s is my venture into high end lighting (at $58.00 ea.) I gave the ColorMeter test results for the H198s in my response to Tim, above.
The COBs as I said are the easiest to assemble but they are powerful point (9mm diameter of light emitting surface) light source so if light panel is needed then something like this (https://www.youtube.com/watch?v=N-gNAJZ-GMM) or this (https://www.youtube.com/watch?v=elBzBTtE78I) for mounting would keep them cool. And the diffuser in front.
I looked at those videos and also related ones that Youtube offered. That involves a lot of heavy duty metal work with the heat sinks. (Also I figured out what "grow lights" are usually used for. Apparently not for growing roses.)
Panels in most cases just need aluminium backing but should be ok as is (they usually contain a set of low powered LEDs).
Aluminum backing? Do you mean heat sinks?
I needed a point light source for my application so mounted 4 COBs adjacent to each other on CPU heatsink with reflector (see attachment).
For Spectron (https://github.com/Alexey-Danilchenko/Spectron)?
Which is really cool but is a different use case than mine.
-
I mean my LG LED display was calibrated at the factory by $10,000 Minolta Color Analyzer. What makes it more expensive than the equipment you used to measure those bulbs which I'm assuming costs far less?
What color analyzer, measuring device are you using?
-
Would you be less bewildered with the idea that the accuracy of light measuring instruments seems to be related to how much they cost?
I mean my LG LED display was calibrated at the factory by $10,000 Minolta Color Analyzer. What makes it more expensive than the equipment you used to measure those bulbs which I'm assuming costs far less?
I'm using a ColorMunki spectrophotometer that is similar to this one (https://www.bhphotovideo.com/c/product/550833-REG/X_Rite_CMUNPH_ColorMunki_Photo_Color_Management.html) with ArgyllPro ColorMeter (http://www.argyllcms.com/pro/) The absolute accuracy is less than with more expensive instruments, but I'm confident that the relative measurements are reasonably accurate. i.e., when I measure a bulb as having a CRI of (say) 83, that its CRI is close to 83. When I measure a TLCI of 97, the real TLCI is close to 97. The spectral plots I display may vary from what more expensive instruments give, but the shapes are approximately accurate.
If I am incorrect with my assumptions, will somebody please correct me!
I am not bewildered at the results of my measurements. I'm bewildered at the lack of correlation between the measurements and the color rendering that you are reporting.
The Cree bulbs repeatedly measure a TLCI close to 97. If a TLCI of 97 means extremely bad color rendering, then something is wrong.
-
If I am incorrect with my assumptions, will somebody please correct me!
Not incorrect. I suspect differing software products may report CRI a bit differently but far better than someone posting images from a camera and reporting what the camera states for CCT Kelvin. Nearly useless.
-
I'm using a ColorMunki spectrophotometer that is similar to this one (https://www.bhphotovideo.com/c/product/550833-REG/X_Rite_CMUNPH_ColorMunki_Photo_Color_Management.html) with ArgyllPro ColorMeter (http://www.argyllcms.com/pro/) The absolute accuracy is less than with more expensive instruments, but I'm confident that the relative measurements are reasonably accurate. i.e., when I measure a bulb as having a CRI of (say) 83, that its CRI is close to 83. When I measure a TLCI of 97, the real TLCI is close to 97. The spectral plots I display may vary from what more expensive instruments give, but the shapes are approximately accurate.
If I am incorrect with my assumptions, will somebody please correct me!
I am not bewildered at the results of my measurements. I'm bewildered at the lack of correlation between the measurements and the color rendering that you are reporting.
The Cree bulbs repeatedly measure a TLCI close to 97. If a TLCI of 97 means extremely bad color rendering, then something is wrong.
I'm not knocking your instrument's accuracy. I have the same question you do about what the numbers mean to what I'm suppose to see. To cut to the chase do you see the Crees color of white (greenish to orangish yellow) to be the same as I've posted in my images?
If your Crees are showing you perfectly neutral colorless white light viewing a spectrally flat gray target then either your eyes or my eyes are playing tricks on us or the Crees have quality control issues in maintaining consistency bulb to bulb in rendering neutral white light or the measuring instruments are not that sensitive.
I have two Hyperikon LED floods that are rated at 5000K. One has the greenish yellow tint and the other a more neutral blueviolet tint which my camera's AWB shows an ACR tint slider in the minus (greenish) region meaning it sees the magenta pink portion of violet. The camera is seeing what my eyes are seeing. My eyes know when it is seeing something that looks neutral and when it is not.
What and how does a measuring instrument define neutrality if they're going to assess the same same number to two different tints of white lights?
What accuracy measurement number does one apply to tell us which bulb is the most accurate to daylight?
-
Ideally, a pair of fixtures like the DT Photons for copy stand work (large documents, camera scanning) and a pair of DT Photon XLs for lighting artwork. Both from Digital Transitions (https://dtdch.com/dt-photon-custom-cultural-heritage-lighting/). (In my dreams.)
Ok - so rectangular flat panels. The easiest then would be to start with the panelled LEDs (the ones you have been looking at) and perhaps DC (constant voltage) power supply. That will be straightforward to connect. Then after mounting it on some backing panel in the case with diffuser, you can run them in for some time, monitoring temperature and doing periodic measurements to monitor for thermal/current related colour shifts and whether they are within your tolerance levels. If they are then you may be perfectly happy with just DC power supply. The constant current approach is not going to work with the panels (due to the ways the soldered) or strips.
In case of Yuji - their panels are MCPCB (metal core boards) so they do have better heat dissipation that normal PCB and you may not need any heatsink installed.
Those Yuji panels (https://store.yujiintl.com/collections/vtc-d50-series/products/vtc-series-d50-5000k-high-cri-mcpcb-led-module-unit-5-pcs) by my calculations contain 21 LED in each strip: 3 parallel strings of 7 LEDs in series with resistor(s) each. Incidentally the panel's single strip specification (https://cdn.shopify.com/s/files/1/0344/6401/files/YJWJ013-1.1_YJ-VTC-MOD-5730-24V-G01-D50.pdf?3774671522798093348) is shared on that very page (dimensions current etc).
But in the real world I've been making do with hardware store clamp fixtures and hardware store bulbs (first "full spectrum" CFLs and now "high CRI" LED bulbs). When shooting art I use cross polarization which sucks up a lot of light. I need about 600 watts (equiv.) to keep shutter speeds in the one or two second range when shooting base ISO. Six clamp fixtures (5 inch reflectors) give me enough uniformity for my standards.
That powerful lighting may need to consider COBs - only they will give you that light density in a relatively compact package. But as you know that means fiddling with heatsinks and backing and perhaps even adding active cooling (fans).
I looked at those videos and also related ones that Youtube offered. That involves a lot of heavy duty metal work with the heat sinks. (Also I figured out what "grow lights" are usually used for. Apparently not for growing roses.)
Orchids from what I looked at. Some of those videos go through more complex setups (multicoloured LEDs etc) but in essence it is simple. Finding the metal backing, drilling a pair of wholes for each COB fixture (if you attach them mechanically with connectors), add thermal paste and fasten the COBs with connectors. Then on the back of the metal panel all extra heatsinks or fans if needed. There is alternative approach of course and that is why I posted a photo of my setup - no COBs are screwed to radiator there at all and no holes are drilled.
Aluminum backing? Do you mean heat sinks?
I mean aluminium extruded strip designed to hold LEDs (those are frequently used in kitchen lightings). But those are for long and narrow strip mountings usually.
For Spectron (https://github.com/Alexey-Danilchenko/Spectron)?
Which is really cool but is a different use case than mine.
Again I posted those not as something you should do but as example of fairly easy and compact package giving me 2400lm light source with fairly good heat dissipation and not a lot of mechanical work. I put 4 COBs on thermal glue (flexible one that expands and contracts) on a heatsink, soldered their inner bits with small wires (so they connected in series). Then cut out the board to the radiator size (used the single sided PCB board material) with the square inside to go around LEDs and attached metal pins to the board (soldered but could be glued or screwed) that aligned with the LED contacts. Those go off the board as wires to the power supply. The board simply sits on top of the LEDs screwed to the heatsink. Took me about couple of hours to do the prototype (mainly to cut square hole). What you see on a photo is the same board fabricated on one of the PCB services (mainly because I like things neat and tidy) but it worked fine with just what I described.
-
I'm not knocking your instrument's accuracy.
At least he has an instrument that's actually designed to measure an illuminant, seems you don't. You were asked FWIW.
At least you can't knock someone using the right tool for the right job. That fellow is under no obligation not to knock someone for discussing CCT and CRI values with a fellow who has no such equipment or understanding on how to produce such measurements or values. ::)
"I often say that when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind".
-Lord Kelvin
-
As I said earlier, I'm not worried about being electrocuted (because the AC is hidden inside the power supplies), I'm worried about thermal runaway (meaning that the LEDs self destruct, which is a concern when the LEDs in question cost several hundred dollars.)
There is no (as in zero) chance of thermal runaway if you're using the modules and you stick to wiring them in parallel. That's what the resistors built into them are for. If you want to work with the LEDs directly, then, as mentioned above, you need specialist knowledge and equipment.
-
There is no (as in zero) chance of thermal runaway if you're using the modules and you stick to wiring them in parallel. That's what the resistors built into them are for. If you want to work with the LEDs directly, then, as mentioned above, you need specialist knowledge and equipment.
Small correction - the strips are wired to each other sequentially, this way the actual LEDs end up wired in parallel strings (7 LEDs in each string)
-
I'm not knocking your instrument's accuracy. I have the same question you do about what the numbers mean to what I'm suppose to see. To cut to the chase do you see the Crees color of white (greenish to orangish yellow) to be the same as I've posted in my images?
If your Crees are showing you perfectly neutral colorless white light viewing a spectrally flat gray target then either your eyes or my eyes are playing tricks on us or the Crees have quality control issues in maintaining consistency bulb to bulb in rendering neutral white light or the measuring instruments are not that sensitive.
I have two Hyperikon LED floods that are rated at 5000K. One has the greenish yellow tint and the other a more neutral blueviolet tint which my camera's AWB shows an ACR tint slider in the minus (greenish) region meaning it sees the magenta pink portion of violet. The camera is seeing what my eyes are seeing. My eyes know when it is seeing something that looks neutral and when it is not.
What and how does a measuring instrument define neutrality if they're going to assess the same same number to two different tints of white lights?
I'm not really sure. I just did my own visual test. I set up a small test arrangement of several objects which include a French's mustard container, a ColorChecker chart, and a WhiBal card. The other objects were chosen to be commonly available (at least in the US) to get red, (sort of) green, and blue into the scene. (Maybe we need a scavenger hunt to find the best commonly available color test obects (that aren't a ColorChecker card). Go to Walmart with an empty shopping cart and get...)
I don't see a greenish tint because I white balance on a WhiBal card as the first step in ACR. I have no doubt that the hardware store lights have a consistency issue. There is only so much that can be done in the power supply of a bulb that sells for less than $10. This is why Alexey is paying so much attention to the power supply with his Yuji lights.
You mentioned cadmium yellow. I don't paint and don't have the foggiest idea of what pigments were used in the paintings I'm shooting.
I shot raw and white balanced on the lower right portion of the WhiBal card. (This is the way I always shoot paintings.) I adjusted exposure (in ACR) so that this portion of the card was approx. 186. They are
1. Walmart 75w daylight LED
2. Cree 100w daylight LED
3. Aputure Amaran AL-H198 LED panel
4. Outside under a cloudy sky, about 1:00 PM EDT. ColorMeter measured the CCT as being about 6000K
I'm attaching the four images. The three done under artificial light look pretty similar. The one done outside looks different, 1., because the CCT is higher (shows on the background) and 2., because the angle of the light is different. (And 3, because the one done outdoors is slightly out of focus because I quickly shot handheld because I felt like a doofus crouching down in my driveway shooting a small white thing balanced on a storage tub.)
When I expand the images so the ColorChecker chart fills the field of view I can't see much difference between any of them.
What accuracy measurement number does one apply to tell us which bulb is the most accurate to daylight?
This is what I want to know. As I showed in my earlier post, the Walmart, Cree, and Aputure LEDs have different measurements. The TLCI (which is how cameras render color) are close to 97 for both the Cree and Aputure lamps. This should mean something.
-
Thanks for posting the images, Wayne. I'm glad I chose the French's Mustard bottle as a target because toggling through all four shows the gradual increase of green with Walmart being the worst, then progressing toward less green to the intended slightly reddish yellow that is Cadmium Yellow in the 6000K shot.
The CCchart yellow patch which is too small to detect color change bares me out with the standard known yellow Lab numbers of L*80,a*4,b*80 which is spot on with 6000K but the Caucasian skin patch has a pink bias of L*70,a*18,b*15 which indicates that overcast clouds have a slight green bias which is why I shot my outdoor target under direct sunlight.
I'ld say the Aputure is the most accurate with spot on Lab numbers for Caucasian skin and yellow patch (just off by 2 points in the (a*) channel for yellow which is undetectable). Lab numbers can be off by as much as 5 points in either channel without a visual difference which is why I chose the mustard bottle target which does show differences even though the yellow patch Lab numbers are so close between all four. Walmart's bulb is in the minus region in the a* channel. That indicates quite a bit of green.
So visually you don't see a greenish yellow hue in the white light of the Crees or Walmart bulbs when you first turn them on?
-
To clear up something about my questioning the accuracy of your measuring instruments and software I was under the impression you were measuring the lights as they are with their greenish yellow white hue but I don't know what surface you pointed the Colormunki spectro to arrive at the numbers.
Your measurements are based on reflected light off something but you didn't make it clear what it was. What doesn't make sense and might be cleared up at least for me regarding the greenish yellow tint of white I see from the Cree and Walmart bulbs is maybe you were measuring off a surface these lights were lighting that has optical brighteners.
What were you measuring off of to arrive at the CRI/CCT numbers you posted?
-
Lab numbers can be off by as much as 5 points in either channel without a visual difference which is why I chose the mustard bottle target which does show differences even though the yellow patch Lab numbers are so close between all four.
Utter rubbish! Easy to dismiss as shown below where one Lab channel differs by 5 units and we see (numerically via deltaE of 3.88 and visually) that statement is totally untrue.
Readers of Mr. Lookingbill, beware of such kinds of statements made that are colorimetrically untrue and by someone without the hardware or software to make such incorrect generalities about color! He may ignore my corrections to his text, but you can ignore the misinformation from his posts after I've corrected them colorimetrically as we see here:
-
What were you measuring off of to arrive at the CRI/CCT numbers you posted?
The illumination! If you owned or ever used a spectrophotometer that measured an light source, you would know this and maybe HOW to do so correctly.
You used the wrong tool (a camera) for this task and the question you asked Wayne applied to your incorrect use of a camera to “capture” not measure anything really. What software did YOU not use to get CRI values you didn’t provide?
-
To clear up something about my questioning the accuracy of your measuring instruments and software I was under the impression you were measuring the lights as they are with their greenish yellow white hue but I don't know what surface you pointed the Colormunki spectro to arrive at the numbers.
Your measurements are based on reflected light off something but you didn't make it clear what it was. What doesn't make sense and might be cleared up at least for me regarding the greenish yellow tint of white I see from the Cree and Walmart bulbs is maybe you were measuring off a surface these lights were lighting that has optical brighteners.
What were you measuring off of to arrive at the CRI/CCT numbers you posted?
I need to reread my books on color science. It has been several years since I studied them. Now I'm working through Color Management for Photographers, then on to more advanced ones. Right now, I don't know how to do the L*a*b analysis you are doing. (But I do generally understand L*a*b.) I've been concentrating on the raw CRI, TLCI, Duv, and R9 measurements. And on the overall shape of the spectrums.
When I turn the lights on for photographing I don't pay attention to tints I see--cameras renders color differently than eyes do. (Especially with the Luther/Ives situation.) This is why there is CRI for human eyes and TLCI for cameras. (I think.) Clicking on a WhiBal card in the frame takes care of tints (I think).
The ColorMunki Design spectrophotometer (https://www.xrite.com/categories/calibration-profiling/colormunki-design) I use has two ports for reading. The one at the top is for reading ambient light through a diffuser. When it is set up for reading ambient light, I point it right at the light, usually six to eight inches away from the light to swamp out room reflections. I always do several measurements from slightly different angles to ensure that reflections aren't contaminating the reading. (Think of putting a microphone right next to a guitar speaker--any room reflections are too small to be meaningful.)
The bottom port on the ColorMunki is undiffused. It is for either reading reflected color (say, from printer profile charts, or a paint/fabric sample). Or for reading undiffused light (spot emission.)
The ColorMunki Design is out of production. I think that is replaced by the i1Studio Spectrophotometer (https://www.bhphotovideo.com/c/product/1369004-REG/x_rite_eostudio_i1studio.html), which is a better name. Before, XRite had ColorMunki spectrophotometers and ColorMunki colorimeters (can only be used to calibrate monitors) which confused people.
Photographers use spectrophotometer for profiling printers and for measuring light. Unless you go used, the ColorMunki (or i1Studio) is the cheapest spectrophotometer you can get new. If you want to profile a printer, it comes bundled with Xrite software. Or you can use ArgyllCMS. (If you really want to profile printers the more expensive XRite options are better. But cost about three times as much.)
Even though spectrophotometers can be used to calibrate monitors, colorimeters are better. I use an i1 Display Pro to calibrate my monitors.
I put the images up at 2400x1600. If you look at the original images I attached the ColorChecker patches are larger than they are in real life so you should be able to measure them.
-
Right now, I don't know how to do the L*a*b analysis you are doing. (But I do generally understand L*a*b.)
Then you're way above Tim's understanding here. He is again not using proper tools or software to comment let alone analyze Lab values or measured the color of anything like you've done with the right equipment. Just examine the silly notion that's been easily dismissed below after Tim incorrectly wrote: "Lab numbers can be off by as much as 5 points in either channel without a visual difference". Pretty clear to someone without color vision deficiencies that the two yellow Lab values provided by him have a visual difference and that's just ONE set of Lab triplets I've analyzed (correctly) in CT&A that Tim didn't/cannot. He doesn't even understand how you correctly used your Spectrophotometer to measure the illuminant!
Go back to reading about color management from others :o don't get stuck into another of Tims rabbit hole of color management absurdities.
-
Ok - so rectangular flat panels. The easiest then would be to start with the panelled LEDs (the ones you have been looking at) and perhaps DC (constant voltage) power supply. That will be straightforward to connect. Then after mounting it on some backing panel in the case with diffuser, you can run them in for some time, monitoring temperature and doing periodic measurements to monitor for thermal/current related colour shifts and whether they are within your tolerance levels. If they are then you may be perfectly happy with just DC power supply. The constant current approach is not going to work with the panels (due to the ways the soldered) or strips.
In case of Yuji - their panels are MCPCB (metal core boards) so they do have better heat dissipation that normal PCB and you may not need any heatsink installed.
Those Yuji panels (https://store.yujiintl.com/collections/vtc-d50-series/products/vtc-series-d50-5000k-high-cri-mcpcb-led-module-unit-5-pcs) by my calculations contain 21 LED in each strip: 3 parallel strings of 7 LEDs in series with resistor(s) each. Incidentally the panel's single strip specification (https://cdn.shopify.com/s/files/1/0344/6401/files/YJWJ013-1.1_YJ-VTC-MOD-5730-24V-G01-D50.pdf?3774671522798093348) is shared on that very page (dimensions current etc).
It also gives the output in lumens. 480 lumens per strip, 4800 lumens for the package of ten strips.
Transposing, I currently use six 100w (equiv.) lights to get enough light to shoot paintings cross polarized. A 100w bulb is about 1500 lumens. Six bulbs is 9000 lumens. So I would need two sets of Yuji strips at $209 for each set and four Yuji power supplies at $45 ea. Which adds up to around six hundred dollars, before I get into fabrication.
At this point I don't think I can justify this for shooting family paintings. I'll spend additional time shifting hues around in Photoshop to make my images look pleasing (if not accurate.)
Orchids from what I looked at.
Look at the videos again. They aren't growing orchids.
(https://www.youtube.com/watch?v=MNKL9onYB_8)
Thanks for all your help. Maybe in six months or a year I might want to tackle this if I build up a fresh head of steam. But right now, I'll stick with the lights I have.
-
Go back to reading about color management from others
Your book, along with Real World Color Management, are solid introductions to the parts of color science that matter most to photographers. Before I tackle Wyszecki and Stiles again.
Have you considered writing a second edition? Even though almost everything in your book is relevant today, there have been advances in technology. P3 and HDR are coming real soon (and already exist to a certain extent in recent smartphones.) Cinema standards are driving things. Even for those of us that aren't videographers, the technology is coming. And is all about color science.
-
Utter rubbish! Easy to dismiss as shown below where one Lab channel differs by 5 units and we see (numerically via deltaE of 3.88 and visually) that statement is totally untrue.
Readers of Mr. Lookingbill, beware of such kinds of statements made that are colorimetrically untrue and by someone without the hardware or software to make such incorrect generalities about color! He may ignore my corrections to his text, but you can ignore the misinformation from his posts after I've corrected them colorimetrically as we see here:
Andrew, your deltaE of 3.88 is a deltaE2000. Perhaps Tim is referring to deltaE1976, since he didn't specify the version and the latter is usually meant when not specified. In your example the deltaE1976 is 5. I'm pretty sure there are colors in the yellows with a deltaE1976 of 5 but have a generally more visually accurate metric, deltaE 2000, that is less than 1. Perhaps he stumbled across some of those visually close colors and was thinking that they look awfully close even though the had a large deltaE1976. And he did say "as much as."
Looking at a set of L*a*b* values one can guess pretty closely what the deltaE1976 is. If 2 of the 3 components are the same the difference of the third is the deltaE1976. Actually, if you just think of a box with the length, height, and width having lengths that are the differences between L*'s, a*s, and b*s the distance between opposite corners of the box is the deltaE1976. Fairly easy to make a rough guess if you can visualize a box.
No such way to guess for deltaE2000 w/o digging out the somewhat messy formula.
-
When I turn the lights on for photographing I don't pay attention to tints I see--cameras renders color differently than eyes do. (Especially with the Luther/Ives situation.) This is why there is CRI for human eyes and TLCI for cameras. (I think.) Clicking on a WhiBal card in the frame takes care of tints (I think).
The ColorMunki Design spectrophotometer I use has two ports for reading. The one at the top is for reading ambient light through a diffuser. When it is set up for reading ambient light, I point it right at the light, usually six to eight inches away from the light to swamp out room reflections. I always do several measurements from slightly different angles to ensure that reflections aren't contaminating the reading. (Think of putting a microphone right next to a guitar speaker--any room reflections are too small to be meaningful.)
The bottom port on the ColorMunki is undiffused. It is for either reading reflected color (say, from printer profile charts, or a paint/fabric sample). Or for reading undiffused light (spot emission.)
I believe you may be using the spectro not the way it was intended for measuring ambient light which by that very word is a human perception function for measuring the surrounding light of one's display during calibration. I don't think your ambient light is you looking straight into the light bulb 8 inches from your eyes while you edit images on the display.
The Colormunki spectro's "Ambient" feature is designed to measure the surrounding light which is not as bright around the display which the software takes into account during calibration/profiling when building the display profile. What you're doing is akin to measuring the city's water pressure with your face 8 inches from the front of a fire hose set to full blast without taking the firetruck's pump into account and any errant backwash pressure and over spray. It's not practical and its accuracy is suspect even if your face and eyes are still intact. A measuring instrument won't be accurate either without accounting for all known and unknown variables.
This is all unnecessary complexity anyway that keeps you from solving the issue you described about your family paintings having some colors not reproducing accurately on your display and thus in the print and questioning the lighting being used during the photographic process. I think you solved your lighting issue especially using the Aputure LED's.
I'm just not going to trust your spectro in the way you're using it to measure off the shelf and inexpensive daylight bulb's spectral quality. As your photos show the CCT/CRI numbers don't tell me anything that's practical in color reproduction work using the photographic process.
The Lab readings I took from Photoshop's Info Palette by downloading your images of the CCchart. It's all you can go by to tell you no matter how greenish yellow a 5000K light source makes a WhiBal neutral target appear and thus photographed by the camera (CCT/CRI numbers having no absolute correlation to visual perception in this case), when the WhiBal card is clicked for R=G=B in ACR/LR, the question remains how many colors will be distorted compared to the original photographed subject.
From your posted photo of the Aputure lit CCchart I don't think you're going to have a problem or require a lot of editing.
-
Wayne, can't you post a photo of a cropped section of one of your family paintings showing colors you're having trouble reproducing? That would simplify a lot of this.
-
I need to reread my books on color science. It has been several years since I studied them. Now I'm working through Color Management for Photographers, then on to more advanced ones. Right now, I don't know how to do the L*a*b analysis you are doing. (But I do generally understand L*a*b.) I've been concentrating on the raw CRI, TLCI, Duv, and R9 measurements. And on the overall shape of the spectrums.
The ultimate book about measuring colours in general is Hunt's "Measuring Colour" (https://www.amazon.co.uk/Measuring-Colour-Imaging-Science-Technology/dp/1119975379/ref=sr_1_1?ie=UTF8&qid=1528447843&sr=8-1&keywords=Measuring+colour). It does go into details about colour spaces, measurements, comparisons and light sources and is your definitive source (this has become my most used book in the past couple of years as I keep going and re-reading bits ;)).
Regarding more precise light source measurements - a good idea is to do it in integrating sphere (does not have to be commercial - it's an easy enough DIY project for weekend) to get the correct figures.
-
I believe you may be using the spectro not the way it was intended for measuring ambient light which by that very word is a human perception function for measuring the surrounding light of one's display during calibration. I don't think your ambient light is you looking straight into the light bulb 8 inches from your eyes while you edit images on the display.
The Colormunki spectro's "Ambient" feature is designed to measure the surrounding light which is not as bright around the display which the software takes into account during calibration/profiling when building the display profile. What you're doing is akin to measuring the city's water pressure with your face 8 inches from the front of a fire hose set to full blast without taking the firetruck's pump into account and any errant backwash pressure and over spray.
I am sorry but that is a load of nonsense. The ambient measurements on spectrometers achieved (with the lack of integrating sphere to aid such measuremengts) by measuring through diffuser. Pointing it towards the sole light source is the correct way to measure in this case. The power of illumination has nothing much to do with it - diffuser will attenuate the illumination to start with and the spectrophotometers sensors allow varying integration times to adjust for less or more light.
-
The ultimate book about measuring colours in general is Hunt's "Measuring Colour" (https://www.amazon.co.uk/Measuring-Colour-Imaging-Science-Technology/dp/1119975379/ref=sr_1_1?ie=UTF8&qid=1528447843&sr=8-1&keywords=Measuring+colour). It does go into details about colour spaces, measurements, comparisons and light sources and is your definitive source (this has become my most used book in the past couple of years as I keep going and re-reading bits ;)).
Regarding more precise light source measurements - a good idea is to do it in integrating sphere (does not have to be commercial - it's an easy enough DIY project for weekend) to get the correct figures.
And after all that reading can you provide any real world example of the practical application from the info that solved a problem concerning color reproduction using the photographic process?
I'm into getting things done and solving problems. If all this reading still hasn't solved a problem concerning the subject of color reproduction then this is nothing but info-tainment.
I get it that people like to read technical stuff but there isn't anyone here that has a degree or is certified as a digital imaging scientist or color scientist, not even the one consultant here who amounts to being a technical writer using color science nomenclature to come across as an authority on the subject and shilling for X-rite, Epson and other photographer services and products businesses.
-
I am sorry but that is a load of nonsense. The ambient measurements on spectrometers achieved (with the lack of integrating sphere to aid such measuremengts) by measuring through diffuser. Pointing it towards the sole light source is the correct way to measure in this case. The power of illumination has nothing much to do with it - diffuser will attenuate the illumination to start with and the spectrophotometers sensors allow varying integration times to adjust for less or more light.
You are not a color scientist, Alexey. You're a programmer as you've pointed out.
If you're so sure about this process then provide practical application that solves the problem of color reproduction using the photographic process.
As it is you don't know diddly squat about what you are doing. Get a degree in color science and come back and report that you solved a problem with that education that actually helped people, not just informed them. Talk is cheap and you're no expert. No one here is an expert, not even me.
I'm a pragmatist. I get things done! I don't talk about getting it done!
-
As it is you don't know diddly squat about what you are doing.
Speak for yourself please because this is impression I have after reading all the crap you are posting
-
Speak for yourself please because this is impression I have after reading all the crap you are posting
He really has no clue about this topic and his agenda is not learning or aiding others; ignore his severe misinformed posts and maybe he will go away.
-
Andrew, your deltaE of 3.88 is a deltaE2000. Perhaps Tim is referring to deltaE1976, since he didn't specify the version and the latter is usually meant when not specified. I
Doug, you of all people here can't be taking this misstatement, (one of so many just in this thread), seriously are you, no matter the dE formula?
Posted by: Tim Lookingbill: Lab numbers can be off by as much as 5 points in either channel without a visual difference...
Tim hasn't the foggiest idea of what he's suggesting, please don't give his color science fiction any degree of seriousness.
-
He really has no clue about this topic and his agenda is not learning or aiding others; ignore his severe misinformed posts and maybe he will go away.
Duly noted - thanks Andrew.
-
No one here is an expert, not even me.
Wrong then absolutely correct; a rare occurrence from your keyboard. Not are you not an expert on color, nor play one on TV, you're generally wrong and without sufficient equipment or knowledge to be posing on this subject. Which is why you're called out and corrected so often from so many....
Now what is it you actually do for a living? Aside from posting a lot of rubbish on the subject of color, proving you're not an expert or sufficient in that topic, do give us an idea of what you believe you are an expert in, other than fiction on the subject of color. A lack of a reply indicates you still live with your parents. ;D
-
Before I tackle Wyszecki and Stiles again.
Have you considered writing a second edition?
Yeah, W&S isn't easy, evening reading but a good reference!
No desire to update the book; too much work, too little gain. It's enough work these days dismissing the nonsense from amateurs who think they should write about color in various forums but clearly should not. ;)
-
... where one Lab channel differs by 5 units and we see (numerically via deltaE of 3.88 and visually) ... after I've corrected them colorimetrically as we see here:
What program are you using? All I've got is ColorMeter and the various programs in ArgyllCMS. Which are very good but I don't think I can do the kind of comparative analysis that you did. (Or maybe I need to add the documentation for ColorMeter and ArgyllCMS to my stack of remedial color science reading.)
-
What program are you using?
CT&A
http://www.babelcolor.com (http://www.babelcolor.com)
Love how it shows the various deltaE differences visually (with the text and such).
ColorThink Pro can do the same. But what's great about CT&A unlike ColorThink Pro, is I can enter Lab or RGB values directly into the input fields (after receiving common numeric assumptions from Tim over the years), and actually show visually and numerically, the dE using many differing formulas.
-
Oh, for Doug and the fellow ignoring the colorimetric facts because he has no such tools, here's another set of Lab triplets that only differ 3 units in one channel and have a clear visual difference and a large dE using a differing formula. As I said, that statement about no visual differences when one Lab channel is off by 5 units is utter rubbish and the figment of someone's imagination:
-
This is a do-over for my earlier post (http://forum.luminous-landscape.com/index.php?topic=123799.msg1048418#msg1048418) because "4. Outside..." because 4. was out of focus. This time I took my tripod out in the driveway and shot it properly. (Somebody I probably know honked when they drove by. This isn't helping my "crazy old man" reputation.)
I shot raw and white balanced on the lower right portion of the WhiBal card. (This is the way I always shoot paintings.) I adjusted exposure (in ACR) so that this portion of the card was approx. 186. They are
1. Walmart 75w daylight LED
2. Cree 100w daylight LED
3. Aputure Amaran AL-H198 LED panel
4. Outside under a cloudy sky, about 1:00 PM EDT. ColorMeter measured the CCT as being about 6000K
I'm attaching the four images. The three done under artificial light look pretty similar. The one done outside looks different, 1., because the CCT is higher (shows on the background) and 2., because the angle of the light is different. (And 3, because the one done outdoors is slightly out of focus because I quickly shot handheld because I felt like a doofus crouching down in my driveway shooting a small white thing balanced on a storage tub.)
This time I have
1. Outside under a cloudy sky, about 12:30 PM EDT. ColorMeter measured the CCT as being about 6000K
2. Same, only with my Google Pixel 2. Taken about four minutes after 1.
3. What the clouds looked like. They were moving so light was changing.
4. ColorMeter done as close to 1. as possible (because of moving clouds.)
Because the Pixel 2 image was an OTC JPEG I didn't attempt to adjust white balance. I did adjust exposure (pulled up in Curves) so the lower right portion of the WhiBal card was about 186, like the others.
Um, why is the Duv of real sunlight 0.0029? Shouldn't it be closer to 0? Is this an indication of measurement error with my ColorMunki (or with ColorMeter)? Or is there something going on with the sun that the government isn't telling us? Maybe this is a hint. (https://xkcd.com/2004/)
-
CT&A
http://www.babelcolor.com (http://www.babelcolor.com)
Love how it shows the various deltaE differences visually (with the text and such).
ColorThink Pro can do the same. But what's great about CT&A unlike ColorThink Pro, is I can enter Lab or RGB values directly into the input fields (after receiving common numeric assumptions from Tim over the years), and actually show visually and numerically, the dE using many differing formulas.
Cool. I downloaded CT&A and am running it in trial mode. I can't seem to enter Lab values, only RGB. I used PhotoShop's info tool to get RGB values from the CC yellow patch on 1., the Cree LEDs I shot yesterday and 2., the real daylight shot I did today (that was in focus.) I attached a screenshot from CC&T.
Is the reason I can't enter Lab values because it is still in trial mode?
It doesn't support my ColorMunki Design. ArgyllCMS spotread (https://argyllcms.com/doc/spotread.html) can make CGATS files from ColorMunki spectro readings (I just verified.) I also don't see any file import in CT&A. Is this also because it is in trial mode?
Otherwise it looks like a good learning tool. I still have a lot of color science remedial work to do.
-
Wayne, will you post the photo of the family painting colors that aren't reproducing as intended and see if your Aputure lights correct it? This was the impetus of this entire discussion that started with you mentioning these paintings.
And disregard what I said about your using the "Ambient" port for measuring the bulbs. I don't know enough about what's going on with your measurement process nor do I have a spectro. Can't afford one anyway.
It wouldn't help me figure out how to predict what colors will distort with these daylight LED bulbs. The CRI numbers only tell me a progression of yellow and skin tone colors that appear to improve according to the CRI number but it doesn't answer why CCT 5000K daylight LED bulbs show this greenish yellow tint of white when there's nothing on planet Earth that shows this hue of white in any natural lighting situation including flash and direct sunlight.
And Andrew can kiss my ass with his unhelpful comments.
-
Oh, for Doug and the fellow ignoring the colorimetric facts because he has no such tools
Try L*a*b* values 80,0,90 and 80,0,95 and look at the delta E2000 which is currently the best estimate of whether two patches appear different or not. ;D
The vast majority of Lab triplets that vary by 5 in either a* or b* are quite visibly different. But not all.
-
And disregard what I said about your using the "Ambient" port for measuring the bulbs. I don't know enough about what's going on with your measurement process nor do I have a spectro. Can't afford one anyway.
That's progress. Wayne's comment about how to read an illuminant's spectrum was exactly correct and showed a good understanding of the issues involved.
Wayne:
The ColorMunki Design spectrophotometer I use has two ports for reading. The one at the top is for reading ambient light through a diffuser. When it is set up for reading ambient light, I point it right at the light, usually six to eight inches away from the light to swamp out room reflections.
-
Is the reason I can't enter Lab values because it is still in trial mode?
Unless he's made it inoperative in demo, you have to make sure the Lab check boxes on either side is clicked on. Then as you enter values, a little red "go" icon appears which you click on to enter the values. Move from field to field on each side. Then you should see the deltaE mode take place and you can alter the formula as desired.
-
The vast majority of Lab triplets that vary by 5 in either a* or b* are quite visibly different. But not all.
Agreed, quite visibly different and I don't believe I suggested none were. Simply that Tim's statement is utterly and consistently bogus; he's speaking of ALL Lab triplets if we make the massive mistake of taking his text seriously.
-
Wayne, will you post the photo of the family painting colors that aren't reproducing as intended and see if your Aputure lights correct it? This was the impetus of this entire discussion that started with you mentioning these paintings.
The paintings I struggled with were done a few years ago when I was using CFLs. The 2nd batch I was shooting recently weren't as out of bounds. I don't know if it is because the pigments in the paintings fell in a different portion of the spectrum (that the LEDs lit properly), or if the newer LEDs are indeed better. Several years ago I went through several rounds of having the images of the paintings printed by AdoramaPix (because I don't have a printer that prints 16x20 nor do I plan on buying one) before I went with the "least worst" prints.
I haven't reshot those paintings yet. I was trying to get a better handle on measuring color rendering before I started up the cycles with AdoramaPix.
And because I have a lot of 100 year old B/W photos to digitize that were also discovered. Color science isn't needed as much for those. I'm working on camera scanning techniques so I don't have to use a slow flat bed scanner again. This is mostly a tripod/copy stand and light stand thing, which is off topic for the color management forum.
And disregard what I said about your using the "Ambient" port for measuring the bulbs. I don't know enough about what's going on with your measurement process nor do I have a spectro. Can't afford one anyway.
[Edit: A ColorMunki grade spectro is less than the cost of a good lens. The ColorMeter Android app is about the cost of a nifty fifty. {Flame}Any photographer should already have an Android Pixel 2 because it is the only smartphone that takes images that withstand 100% scrutiny. P2 image quality ~= a M4/3 sensor.{/flame} A spectro and ColorMeter would be well worth it for any photographer if they can be used to take meaningful measurements of light sources. Without needing to take a long detour into reading a stack of expensive color science textbooks.]
It wouldn't help me figure out how to predict what colors will distort with these daylight LED bulbs. The CRI numbers only tell me a progression of yellow and skin tone colors that appear to improve according to the CRI number but it doesn't answer why CCT 5000K daylight LED bulbs show this greenish yellow tint of white when there's nothing on planet Earth that shows this hue of white in any natural lighting situation including flash and direct sunlight.
As I said, I don't pay too much attention to what my eyes can see (tints) because cameras render differently than eyes do. I pay attention to the image that is displayed on my NEC PA241W (and ultimately on AdoramaPix prints.) So I don't have any good painting examples now. I don't think it would be of much value to show the paintings I shot with CFLs several years ago.
I am also trying to understand how the numerical measurements (CRI, TLCI, Duv, and R9, etc.) correlate. As I have explained, I am refreshing what I used to know about color science, starting with Andrew's book. I am at a loss how the new Cree bulbs have a TLCI of almost 97 and aren't almost perfect for rendering color shot with a camera (other than that pesky Luther/Ives thing.)
ColorMeter and a ColorMunki grade spectro are useful tools, once I figure out how to use them properly.
-
And disregard what I said about your using the "Ambient" port for measuring the bulbs. I don't know enough about what's going on with your measurement process nor do I have a spectro. Can't afford one anyway.
Anyone who's read more than a few of your posts on this subject have already disregarded what you wrote about a piece of hardware you know very little of anything about. Hence your silly rant/question to Wayne after which again, I had to correct you then bring you into our world of facts and science. Indeed, the most accurate text you've written in a very long time here is your admission you don't know enough about the measuring process many of the rest of us do clearly understand. And have actually experienced. Now if you'd consider how much you know, you don't know before posting, we'd all get more done in these parts. But you don't know what you don't know even when you're told correctly about how the stuff you have no experience with, actually works by experts who do know how they work. None being color scientists if you can actually believe that fact!
It wouldn't help me figure out how to predict what colors will distort with these daylight LED bulbs.
Again, you know so little about this topic and how to measure color, your comment, like so many posted here is best ignored. Which is my goal in pointing out again, your massive misunderstandings of this topic along with your own admission of a lack of any experience measuring color.
The CRI numbers only tell me a progression of yellow and skin tone colors that appear to improve according to the CRI number but it doesn't answer why CCT 5000K daylight LED bulbs show this greenish yellow tint of white when there's nothing on planet Earth that shows this hue of white in any natural lighting situation including flash and direct sunlight.
When you ask stupid questions expect stupid answers. No, that isn't what CRI tells those who understand what CRI is and how rather useless it is.
And Andrew can kiss my ass with his unhelpful comments.
The comments were never meant to be useful to you; that's not possible as you're a fact denier. The comments were made to correct your misinformed posts for others and the comments at least today, seem to have been successful in getting a newer member who hasn't witnessed your inability to post facts on color science a tip he too should ignore your rants. The kissing the ass part shows you've got to lower yourself down further from an amateur who has so little experience with this topic of color, to someone morphing into a troll here which may deserve moderation due to that comment. I personally don’t care you wrote it, there's no way I could kiss your ass when your own lips seem to be stuck onto that area of anatomy. Perhaps that position is why you are unable to measure an illuminant like some many other's here. Amateurs like you should read more, type less, assuming your newer agenda is to learn this topic. Until that occurs, what is your agenda for regularly posting misinformation that has been regularly dismissed and corrected with facts, by experts and professionals in the field of color and imaging?
Do we need to talk to your mom and dad about taking away your keyboard son? Shorten your curfew?
-
Unless he's made it inoperative in demo, you have to make sure the Lab check boxes on either side is clicked on. Then as you enter values, a little red "go" icon appears which you click on to enter the values. Move from field to field on each side. Then you should see the deltaE mode take place and you can alter the formula as desired.
That worked. I'm pausing now to watch your sRGB Myths videos. It covers a lot of what I am asking questions about.
Thanks.
-
I'm also trying to find the value of measuring instruments. I just bought and measured a Walmart daylight LED
To add to your measurements in this post (http://forum.luminous-landscape.com/index.php?topic=123799.msg1048266#msg1048266) here are some measurements for Yuji VTC COB LEDs and daylight for comparison. Measurements below show importance of not overheating the LED and include measurement for 4xLEDs run for 10 mins heated up pretty much to max temperature on full power and similarly the same with reduced power (PWM still at full rated current) to 75% which show quite a bit of difference in CCT, CRI and TLCI. Then also included measurement for single LED on large heatsink at full power (it does not go hot) and daylight measurement.
-
Um, why is the Duv of real sunlight 0.0029? Shouldn't it be closer to 0? Is this an indication of measurement error with my ColorMunki (or with ColorMeter)? Or is there something going on with the sun that the government isn't telling us? Maybe this is a hint. (https://xkcd.com/2004/)
The Duv value goes along with the CCT. It's a measure of how far off, on CIEuv chromaticity, the measurement is from a black body at the CCT temperature. It's normal for sunlight, and D50, to have an offset. It certainly shouldn't be 0.
Note that the LED is closer to 0. Duv, together with the CCT, is only a measure of chromaticity and is more intuitive than just xy or uv. All are isomorphic, but they say nothing about how spikey the spectrum is. The various rendering indexes do that and to varying degrees of goodness.
-
The Duv value goes along with the CCT. It's a measure of how far off, on CIEuv chromaticity, the measurement is from a black body at the CCT temperature. It's normal for sunlight, and D50, to have an offset. It certainly shouldn't be 0.
Black body == Planckian locus, doesn't it?
Wikipedia:
The relative spectral power distribution (SPD) of a D series illuminant can be derived from its chromaticity coordinates in the CIE 1931 color space.
(https://upload.wikimedia.org/wikipedia/commons/thumb/3/39/Daylight-locus-in-CIE-1960-UCS.png/500px-Daylight-locus-in-CIE-1960-UCS.png)
Daylight locus in the CIE 1960 UCS. The isotherms are perpendicular to the Planckian locus. The two sections of the daylight locus, from 4000–7000 K and 7000–25000 K, are color-coded.
https://en.wikipedia.org/wiki/Standard_illuminant (https://en.wikipedia.org/wiki/Standard_illuminant)
Does "perpendicular" account for the offset?
I hadn't thought about this. I was extrapolating that because the CRI of the sun (and of a black body radiator) is a perfect 100, that the Duv should be 0. But they aren't the same thing.
Note that the LED is closer to 0. Duv, together with the CCT, is only a measure of chromaticity and is more intuitive than just xy or uv. All are isomorphic, but they say nothing about how spikey the spectrum is. The various rendering indexes do that and to varying degrees of goodness.
The "varying degrees of goodness" is at the root of my posts in this thread.
-
Black body == Planckian locus, doesn't it?
Does "perpendicular" account for the offset?
I hadn't thought about this. I was extrapolating that because the CRI of the sun (and of a black body radiator) is a perfect 100, that the Duv should be 0. But they aren't the same thing.
Nope, they sure aren't the same. People often think of D50 as the same as blackbody 5,000K but they aren't. While the CCT of D50 is very close to 5,000K, the major difference is that it's offset along the CCT locus. That offset is surprisingly large, on the order of dE 4 or so, so the two "whites," side by side are noticeably different. Apart, you really can't see a difference.
However, the L/I issue relative to camera CFAs is a much bigger factor for most colored objects than these differences in the illuminant. And how good that is depends on both the CFA spectral response and the particular spectrum of whatever objects one is photographing. Those are bigger factors than whether the illuminant is D50, D55, D65, or an incandescent at 5,000K.
-
Black body == Planckian locus, doesn't it?
Indeed:
https://www.xrite.com/service-support/please_explain_the_term_black_body_curve (https://www.xrite.com/service-support/please_explain_the_term_black_body_curve)
-
Nope, they sure aren't the same. People often think of D50 as the same as blackbody 5,000K but they aren't.
Yup, I wish I had a dollar for every post I've seen where people associate the the two as being the same. Which is why I strive to always place CCT somewhere around a numeric value that also has Kelvin associated with said value.
-
Yup, I wish I had a dollar for every post I've seen where people associate the the two as being the same. Which is why I strive to always place CCT somewhere around a numeric value that also has Kelvin associated with said value.
So true. Seems to be a widespread belief.
-
Yup, I wish I had a dollar for every post I've seen where people associate the the two as being the same. Which is why I strive to always place CCT somewhere around a numeric value that also has Kelvin associated with said value.
You go, Andrew! D50 doesn't even have a CCT of 5000 Kelvin; it's 5003, thanks to changes in the way that blackbody spectra are calculated since the D-illuminants were defined. It's a little difficult to credit in the computer age how the birth of the D-illuminants were affected by computational ease. And why should daylight act like a blackbody anyway?
Jim
-
You go, Andrew! D50 doesn't even have a CCT of 5000 Kelvin; it's 5003, thanks to changes in the way that blackbody spectra are calculated since the D-illuminants were defined. It's a little difficult to credit in the computer age how the birth of the D-illuminants were affected by computational ease. And why should daylight act like a blackbody anyway?
Jim
Being mathematically challenged, can I assume there is more than one calculation and thus, that 5003 value could change?
Further when comparing our measurements to standard illuminants or such a conversion to CCT, useful to consider that D-illuminants were produced from many (622) measurements made around the planet, using differing spectroradiometers, and averaged to produce that definition used today. So the YMMV warning may apply?
-
Being mathematically challenged, can I assume there is more than one calculation and thus, that 5003 value could change?
It could change if the values of Planck's constants are revised again.
Take a look at the last section of this before the references:
https://en.wikipedia.org/wiki/Planckian_locus
Further when comparing our measurements to standard illuminants or such a conversion to CCT, useful to consider that D-illuminants were produced from many (622) measurements made around the planet, using differing spectroradiometers, and averaged to produce that definition used today. So the YMMV warning may apply?
Absolutely! If D50 is useful, it is useful as a normative standard for comparison, not as a measure of daylight at a particular time of day at a particular spot with particular weather conditions, or even as an average of some of the above.
Jim
-
You go, Andrew! D50 doesn't even have a CCT of 5000 Kelvin; it's 5003, thanks to changes in the way that blackbody spectra are calculated since the D-illuminants were defined. It's a little difficult to credit in the computer age how the birth of the D-illuminants were affected by computational ease. And why should daylight act like a blackbody anyway?
Jim
Worth noting that the dE difference between 5000K and 5003K is less than 0.1 Actually, D50 is pretty far away from a 5000K Planckian white. the visual differences between 5000K and 4750K/5250K are similar to the differences between D50 and 5003K (or 5000K). However, the D50 difference is along the (green) magenta axis while the 5000K->5250 is along the blue (yellow) axis.
-
Worth noting that the dE difference between 5000K and 5003K is less than 0.1 Actually, D50 is pretty far away from a 5000K Planckian white. the visual differences between 5000K and 4750K/5250K are similar to the differences between D50 and 5003K (or 5000K). However, the D50 difference is along the (green) magenta axis while the 5000K->5250 is along the blue (yellow) axis.
Yes, that's been brought up before in this thread, but not, AFAIK, the CCT difference. Sure, it's academic, but it just points out one more way that D50 is not what some folks think it is.
Jim
-
That worked. I'm pausing now to watch your sRGB Myths videos. It covers a lot of what I am asking questions about.
I watched your video and then did a random walk looking at other spectrum analyzing programs. During my random walk I hit on a (39 minute!) video done by a photographer that had been upset that various photo printing labs were apparently using canned profiles instead of having custom profiles done by a professional. She had downloaded a bunch of profiles from the labs and ran them through ColorThink Pro's 3D grapher. She also compared the gamut of one of her own images that had a lot of saturated colors and showed how the gnarly printer profiles wouldn't reproduce as many of the colors from her image as a printer that used a better (custom) profile would. (Message: look for custom profiles. That were done recently.)
I think I can use various ArgyllCMS utility programs to do 3D plots that are similar to Colorthink Pro 3D plots. I had already done a bunch of 3D plots of profile gamuts (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/allgamuts.html).
While wondering how to better quantify color rendering under different illuminants, I decided to run the images I had previously took of a set under various LEDs (http://forum.luminous-landscape.com/index.php?topic=123799.msg1048418#msg1048418) (and under the real sun) that I previously posted through the ArgyllCMS programs. This set
(http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/rimages/700/01-Walmart-75w-daylight_IMG_0072.jpg)
OK, here is where I'm guessing. I loaded the set images into Photoshop from the raw files but this time I reconfigured ACR to load them as Lab. I saved them as 8 bit TIFFs (as Lab) and made ArgyllCMS gamut files from them and made various 3D plots.
Here is a 2D image of the gamut of the set lit by Walmart bulbs against sRGB.
(http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/rimages/700/01-Walmart-sRGB-2d.jpg)
Here is the 3D plot (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/xd3-2018-small/Set-01-Walmart-Lab_IMG_0072_sRGB_wrl.x3d.html). I think this shows that some colors from this image poke out of sRGB a bit, correct?
Here is the gamut of the set lit with the Walmart bulb against the gamut of the set lit with real daylight
(http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/rimages/700/01-Walmart-real-daylight-2d.jpg)
3D plot (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/xd3-2018-small/Set-01-Walmart-Lab_IMG_0072_Set-06_Real-cloudy-daylight-Lab_IMG_0107_wrl.x3d.html) There does seem to be some areas that real daylight covers that the Walmart bulb doesn't. And some areas where the Walmart bulb apparently has more saturation than real daylight does.
I didn't make a 2D screen shot, but here is the set illuminated with the Walmart bulbs against the set illuminated with the Apurture Amaran AL-H198 LED panel (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/xd3-2018-small/Set-01-Walmart-Lab_IMG_0072_Set-04-Aputure-Lab_IMG_0081_wrl.x3d.html). These are closer but the Walmart bulb's gamut is still smaller than the Apurture panel's gamut is.
Here are all the 3D plots (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/CC-Tiffs.html)I did comparing the gamuts of the set lit with different illuminants. Here are the 2D images of the set, ColorMeter spectrum plots, etc. (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/real-images.html)
Now, am I testing anything meaningful here? Am I getting any closer to quantifying color rendering ability?
-
I did gamut plots again but with just the ColorChecker portion. Also, I added in the CC chart in Lab format from Dry Creek Photo (https://www.drycreekphoto.com/Learn/monitor_calibration.htm#charts). Here is the Dry Creek version (with labels of Lab values as read from Photoshop, when you open the Lab version from DCP, not the sRGB version I'm showing here.)
(http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/rimages/600/MacbethCC-LAB-v-L.jpg)
The gamut looks like this in 2D (done from the Lab version of the DCP image with no labels.)
(http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/rimages/600/01-DRP.jpg)
3D plot (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/xd3-2018-small/CC-01_00-DCP-01_MacbethCC-LAB-L_wrl.x3d.html). Smoother (DCP's is an average. Or maybe was CGIed?)
Here is the CC portion of my set image shot in real daylight after I used Photoshop's Perspective Crop to crop out just the CC chart.
(http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/rimages/600/CC-01-04-CC-Daylight-01_IMG_0072.jpg)
Differences in processing: I opened the images up in Lab mode in ACR but this time I switched the profile from "Adobe Color" to "Adobe Standard". And white balanced on the 2nd gray chip (instead of on the WhiBal card.) The Dry Creek Photo chip was 81,-1,0. I adjusted exposure in ACR so this chip was 81,0,0 (The ColorCheck Passport card's bottom row isn't really neutral (http://www.chromaxion.com/information/ColorChecker_Passport_Technical_Report.pdf) according to Robin Myers, which is why I originally used my WhiBal card.)
Here is the 2D plot of the gamut of my CC card shot in daylight.
(http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/rimages/600/02-Real-Daylight.jpg)
3D plot. (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/xd3-2018-small/CC-01-04-CC-Daylight-01_IMG_0072_wrl.x3d.html)
And here is a 2D plot of the gamut of my CC card shot in daylight against the gamut of the Dry Creek Photo CC card.
(http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/rimages/600/03-Real-Daylight-DCP.jpg)
3D plot. (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/xd3-2018-small/CC-01-04-CC-Daylight-01_IMG_0072_CC-01_00-DCP-01_MacbethCC-LAB-L_wrl.x3d.html)
The gamut of the DCP card doesn't go as close to black as the gamuts of my CC card, presumably because the surround of the DCP card is gray while the surround of my ColorChecker Passport card is closer to black.
I don't know why the gamut of the DCP card is larger than my own CC card shots. Possibly because of different raw converters? Or because DCP used a different camera profile than Adobe Standard? All of my CC card shots were processed the same (opened and saved in Lab, the using Adobe Standard profile.) Or (after rereading the DCP page), maybe the DCP chart is synthetic and was created entirely in Photoshop; and wasn't shot from a camera at all.
Here is an updated page of my 3D gamut plots (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/CC-Tiffs.html) showing with all the illuminants I used before (Walmart, Cree, Aperture, real daylight) in different permutations. With the DCP card mixed in.
When looking at the table sorted by Cubic Colorspace Units at the bottom of the above page I see that the gamut plots I did earlier of the entire sets are a lot larger than the gamut of just the CC card. Maybe because the real world objects in my set are more saturated than the CC card? Or maybe they are just different hues, so there are more samples?
Is any of this showing us anything about color rendering under different illuminants? Or am I just making colorful Rorschach tests?
-
Is any of this showing us anything about color rendering under different illuminants? Or am I just making colorful Rorschach tests?
Well, they are quite pretty!
When trying to analyze the quality of lighting and camera profiles it's best to use "scene referred" techniques with the exception of adjusting the white point to a common value. Most all image converters do "output referred" conversions which typically increase saturation and introduce an "S" tone curve to accommodate the large dynamic range most photographed scenes have. This makes them look better when printed.
Also, with scene referred processing one can compare Lab values and associated statistics.
However, there isn't a lot of good info on this as it's more a specialty of people that replicate artwork but you can find some here and by googling.
-
Well, they are quite pretty!
I was trying to demonstrate a model for a better color rendering metric. Refer to my 3D plot (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/xd3-2018-small/Set-01-Walmart-Lab_IMG_0072_Set-06_Real-cloudy-daylight-Lab_IMG_0107_wrl.x3d.html) of the gamut of my set (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/rimages/01-Walmart-75w-daylight_IMG_0072.jpg) illuminated by Walmart bulbs against the gamut of the set illuminated by real daylight. The 3D plot shows both decreased and increased saturation of all colors. (Increased saturation can be just as bad as decreased saturation.) Plus or minus any deficiencies of my methodology (such as differing CCTs.)
This is an example of the kind of color rendering test results I'd like to see. I'd like to see a 3D plot showing exactly how much the artificial illuminant diverges from real daylight (or real tungsten). In addition to the very small list of single digit measurements (CRI, TLCI, etc.) that we now have. A 2D spectral plot isn't much better.
My demonstration only attempted to show saturation. It didn't cover hue shifts (that would require arrows.)
I don't know if my test is realizable from just a spectral measurement of the illuminant.
When trying to analyze the quality of lighting and camera profiles it's best to use "scene referred" techniques with the exception of adjusting the white point to a common value. ...
This covers the entire pipeline. This is important! Thanks for the suggestions. I'm already Googling.
But I was restricting my demonstration to color rendering. Photographers need better metrics so we can decide how high up the illuminant chain we need to go. To cut down on lighting manufacturers gaming the system with color rendering tests that do not adequately describe actual color rendering. Heck, just getting SPDs is difficult enough.
Before we get into Luther/Ives and the Sensitivity Metemerism Index (SMI) tarpit.
Wayne
-
This is an example of the kind of color rendering test results I'd like to see. I'd like to see a 3D plot showing exactly how much the artificial illuminant diverges from real daylight
Could you please define real daylight? This is quite a misnomer because daylight tend to differ and substantially depending on time of day, weather conditions etc.
A 2D spectral plot isn't much better.
SPD of a light source technically is all you need from a light source to model, evaluate and profile the rest (assuming you also have spectral responses of the sensor or sensor+lens combination).
-
Could you please define real daylight? This is quite a misnomer because daylight tend to differ and substantially depending on time of day, weather conditions etc.
(https://upload.wikimedia.org/wikipedia/commons/2/21/CIE_illuminants_D_and_blackbody_small.gif)
Relative spectral power distribution of illuminant D and a black body of the same correlated color temperature (in red), normalized about 560 nm.
https://en.wikipedia.org/wiki/Standard_illuminant (https://en.wikipedia.org/wiki/Standard_illuminant)
Real daylight with the same CCT as the illuminant that is being tested.
Or the D illuminant doppelgängers. For my above demonstrations, substitute a D illuminant for the set (or whatever) being shot in real daylight (or real tungsten). If that is more realizable then modeling lumpy real daylight.
Color rendering tests are for choosing artificial light sources. As a photographer, can I get away with LED bulbs from a hardware store? Or a LED panel from B&H? Or do I need to get into fabricating Yuji components? We need better color rendering tests.
SPD of a light source technically is all you need from a light source to model, evaluate and profile the rest (assuming you also have spectral responses of the sensor or sensor+lens combination).
So how do we do it? The ArgyllCMS programs are powerful (and are portable--Win 10 tablets or netbooks are cheap as chips.) I'm mathematically challenged but can slap together wrapper Perl scripts to call ArgyllCMS programs. If I had guidance.
-
... We need better color rendering tests.
So how do we do it? The ArgyllCMS programs are powerful (and are portable--Win 10 tablets or netbooks are cheap as chips.) I'm mathematically challenged but can slap together wrapper Perl scripts to call ArgyllCMS programs. If I had guidance.
In the early 90s, I wrote some programs to do this. More recently, Jack Hogan and I wrote Matlab code to optimize compromise matrices for specific cameras.
The process is not well defined. Some key decisions: What light spectrum or spectra? What set of reflectance spectra for the training set? What is the error function (avg? rms? Lab? Luv? Lab 2000?) Are some patches in the training set weighted more heavily than others? Is the test patch set different than the training set (it better be unless the training set is very large)? What do you do with a solution space that is polymodal (I have found the genetic algorithm inefficient and simulated annealing better)? Do you allow LUT-based conversions instead of or in addition to compromise matrices? If so, how do you constrain the LUTs? And finally, and practically, can you code all this up to run in a reasonable length of time, especially with training sets that are at least in the mid-hundreds?
I've decided that, for now, life's too short for me to pursue this any more. Jack took it further than I did, and maybe he's got some advice, but I don't think Jack extended the training set beyond the Macbeth CC, which is very small.
Jim
-
Real daylight with the same CCT as the illuminant that is being tested.
Or the D illuminant doppelgängers. For my above demonstrations, substitute a D illuminant for the set (or whatever) being shot in real daylight (or real tungsten). If that is more realizable then modeling lumpy real daylight.
So you were in fact referring to D type illuminants - anyone in particular that you are trying to match?
Color rendering tests are for choosing artificial light sources. As a photographer, can I get away with LED bulbs from a hardware store? Or a LED panel from B&H? Or do I need to get into fabricating Yuji components? We need better color rendering tests.
So how do we do it? The ArgyllCMS programs are powerful (and are portable--Win 10 tablets or netbooks are cheap as chips.) I'm mathematically challenged but can slap together wrapper Perl scripts to call ArgyllCMS programs. If I had guidance.
Any good book which explains CIE colour model and calculations will do to start with to see how to obtain colour from spectral data (light, object, observer essentially). With cameras/sensors it is the same - observer is the camera in this case.
-
So you were in fact referring to D type illuminants - anyone in particular that you are trying to match?
Whichever one that has a CCT that is closest to the light being tested.
Any good book which explains CIE colour model and calculations will do to start with to see how to obtain colour from spectral data (light, object, observer essentially).
I just reread chapter three of a good book (https://www.amazon.com/Color-Science-Concepts-Quantitative-Formulae/dp/0471399183/ref=sr_1_1?ie=UTF8&qid=1528830656&sr=8-1&keywords=color+science+wyszecki) and it tells how to do CRI. I was looking for something more meaningful. I'm not sure if IES TM-30-15 (https://www.ies.org/store/technical-memoranda/ies-method-for-evaluating-light-source-color-rendition/) will be much of an improvement (after Graeme adds it to ColorMeter.)
I'll repeat that I am mathematically challenged so I can't even do CRI by myself from spectral data. (I'm moving from ColorMeter to ArgyllCMS spotread so I can collect spectral data when I'm testing lights. I need to configure my Win 10 tablet with ArgyllCMS first.)
With cameras/sensors it is the same - observer is the camera in this case.
I believe that my above demonstrations show the difference in color rendering between the three illuminants I tested and real daylight at the time I made an image of the same set in real daylight. If I am using the same scene, camera, camera settings, and raw conversion for all four images, doesn't that null out the camera and post processing? (For this particular test.)
And somewhat less ambitiously, plotting an image's gamut against standard color spaces shows if that particular scene will fit in a given color space (sRGB, etc.) I think I got a correlation with the 2nd demonstration (just the ColorChecker card) where cyan pokes a bit out of sRGB (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/xd3-2018-small/CC-01-01-CC-Walmart-01_IMG_0072_sRGB_wrl.x3d.html). Andrew's "sRGB Myths" video also shows cyan poking out of sRGB (he used ColorThink Pro to plot an image's gamut against a color space, compared to me using ArgyllCMS programs.)
-
I just reread chapter three of a good book (https://www.amazon.com/Color-Science-Concepts-Quantitative-Formulae/dp/0471399183/ref=sr_1_1?ie=UTF8&qid=1528830656&sr=8-1&keywords=color+science+wyszecki) and it tells how to do CRI. I was looking for something more meaningful. I'm not sure if IES TM-30-15 (https://www.ies.org/store/technical-memoranda/ies-method-for-evaluating-light-source-color-rendition/) will be much of an improvement (after Graeme adds it to ColorMeter.)
If you are looking to a single quantitative measure to describe the quality of light for photographic application then in my view you are wasting your time. A "usefulness" of a single measure will be as "good" as DxO camera marks which is total rubbish. To many variables to pack into one measure, quite difficult to place weights on those variables simply because those could be application specific. For instance for someone shooting paintings/art matching of specific paints will be more important that any natural colours (skintones, sky etc).
Jim above fairly thoroughly described what is involved. I personally would build profiles with that light (lighting something like CC SG target or custom target if needed) and evaluate their accuracy against variety of shots and (if that is your goal) against the same profile built with daylight (though this is rather tricky).
In the light source I personally would look at SPD to see if the visible range is well covered, then for the smoothness. Most of the LEDs fail in this respect since they have no coverage below 420nm and the large spike to start with (which means for extended range you need to add more different LEDs to the mixture). Yuji so far was the only one I know that gave fairly broad coverage including <420nm (it starts with 380nm). The shape of their SPD loosely follows the D illuminant (no exact fit of course) which is what interested me personally so I replaced my Xenon lighting with them. No regrets so far. Any alternative that I know of is to assemble panel of LEDs covering various wavelength and in total covering the whole visible range - then controlling their intensities to vary the shape of SPD. But that gets progressively harder with the number of LEDs involved (different current requirements, different brightness ranges etc).
I do plan to do the target and profile testing with Yuji D50 LEDs (properly lit target etc) but first I need to build those panels (not the point light source I have already done).
-
If you are looking to a single quantitative measure to describe the quality of light
My experiments were designed to find a better CRI. This is how a human standard observer perceives color matches. I really liked the 3D plots from my demonstration that showed what colors in the artificial light had less saturation and what colors had more saturation. I thought that this is more meaningful than the standard palette of 8-15 colors that various flavors of CRI (and variants) use.
for photographic application
I wasn't attempting that in my test. As I pointed out the camera and processing were nulled out.
I assumed that everybody here would have realized what I was attempting to do (and what I was not attempting to do) so I didn't explicitly explain this. I expected that people would point out deficiencies in my methodology and would suggest improvements. I expected that people would ask for the parameters I passed to iccgamut, tiffgamut and viewgam.
And then, ideally, a way to skip shooting in real daylight. I understand that this is a lot more difficult than passing parameters to the existing ArgyllCMS programs. But skipping shooting in daylight isn't mandatory for this test. Possibly the accuracy of shooting the object in real daylight could be improved by making a table of different CCTs of daylight at different times of day, in shade and in direct sunlight. I haven't done this yet.
Baby steps first. First an improved standard observer metric. Before introducing the complexities of Luther/Ives and SMI. (I'm not sure how TLCI handles this.)
Jim above fairly thoroughly described what is involved.
I read Jim's paper (http://www.ingentaconnect.com/content/ist/cic/1994/00001994/00000001/art00023) last night. Even though I didn't understand it, I was impressed with the complexity that goes on behind the scenes in image processing programs.
I'm not underestimating complexity. I'm trying to harness complexity that has already been solved. Such as with the ArgyllCMS programs which are really powerful. I was hoping that others would suggest other ways they can be used.
Preferably augmented with additional algorithms to reprocess spectral data. I can do algorithms. It is math that I have a problem with.
-
Whichever one that has a CCT that is closest to the light being tested.
The problem is that getting the closest to some CCT ignores the green/magenta error, which is unlimited.
My demonstration only attempted to show saturation. It didn't cover hue shifts (that would require arrows.)
And those (hue shifts) are what human vision is most sensitive to as saturation increases. In fact, generally human vision becomes less sensitive to changes in saturation as the saturation increases while hue shifts remain more perceivable. This is a principal reason I find 3D gamut images not very useful. It's more useful for showing saturation than hue shifts.
-
I read Jim's paper (http://www.ingentaconnect.com/content/ist/cic/1994/00001994/00000001/art00023) last night. Even though I didn't understand it, I was impressed with the complexity that goes on behind the scenes in image processing programs.
Here's a better place to start:
https://blog.kasson.com/the-last-word/the-color-reproduction-problem/
Jim
-
My experiments were designed to find a better CRI.
If we're ignoring capture metameric error, then I think a reasonable way to improve the CRI is to extend the patch set. This is not too hard to do (but can take a while to run if you've got thousands of patches) with simulations but can be difficult IRL. There is a NASA set of natural spectra that you can download, and I think that Burns at RIT published a set of paint spectra. Sorry, I don't have links, but I might be able to find them if you want to go that route.
Jim
-
Good stuff, Jim!
-
If we're ignoring capture metameric error, then I think a reasonable way to improve the CRI is to extend the patch set. This is not too hard to do (but can take a while to run if you've got thousands of patches) with simulations but can be difficult IRL. There is a NASA set of natural spectra that you can download, and I think that Burns at RIT published a set of paint spectra. Sorry, I don't have links, but I might be able to find them if you want to go that route.
I would be very interested in those Jim if you can find them.
Thanks
-
I would be very interested in those Jim if you can find them.
Here's one:
https://speclib.jpl.nasa.gov/
Unfortunately, the lambda range greater than we'd prefer. I've seen more appropriate data sets from NASA. I'll keep looking.
Here's another set:
https://crustal.usgs.gov/speclab/QueryAll07a.php
Here is Berns' artist paint database:
https://www.rit.edu/cos/colorscience/mellon/pubs.php
By the way, there was a thread a while back that is tangentially related to this thread. It was concerned with CFA filters, but the same techniques could be used to evaluate lighting and camera interactions.
https://www.dpreview.com/forums/post/60278621
Jim
-
Here's one:
https://speclib.jpl.nasa.gov/
Unfortunately, the lambda range greater than we'd prefer. I've seen more appropriate data sets from NASA. I'll keep looking.
Here's another set:
https://crustal.usgs.gov/speclab/QueryAll07a.php
Here is Berns' artist paint database:
https://www.rit.edu/cos/colorscience/mellon/pubs.php
By the way, there was a thread a while back that is tangentially related to this thread. It was concerned with CFA filters, but the same techniques could be used to evaluate lighting and camera interactions.
https://www.dpreview.com/forums/post/60278621
Brilliant - thanks Jim.
-
Here's a better place to start:
https://blog.kasson.com/the-last-word/the-color-reproduction-problem/ (https://blog.kasson.com/the-last-word/the-color-reproduction-problem/)
Thanks. This is a more approachable explanation. The DPR CFA II modeling thread (https://www.dpreview.com/forums/post/60278621) explains the difficulties you alluded to in your earlier post. All of this is, of course, several orders of magnitude above my head. But reading your blog posts and looking at the DPR threads makes me appreciate the complexities.
I really appreciate your using prose instead of equations in the fifth post in your series (https://blog.kasson.com/the-last-word/compromise-matrix-construction-details/). If Amazon would make an equation-to-prose translator I'd understand Wyszecki and Stiles better. :)
In blog post 14 (https://blog.kasson.com/the-last-word/where-to-go-with-the-color-reproduction-series/) you noted how tedious it was to transcribe Lab numbers from Photoshop. This is exactly why I tried my 3D plot approach of visualizing spectra. I had followed Andrew's recommendation for BabelColor CC&T (http://www.babelcolor.com/cta_features.htm) and transcribing Photoshop Lab values from different ColorChecker chart readings got old real fast. I was looking at different spectral analysis programs hoping to find one where you could just refer to image files of a CC chart and a reference CC chart (like we can with the Adobe DNG Profile editor) when I got to the video of the photographer using Chromix ColorThink Pro (https://www.chromix.com/colorthink/) to plot the gamut of an image against a color profile and I remembered that I could do the same thing with ArgyllCMS programs. It was only a short leap to plot a CC gamut against a different CC gamut.
I'm looking for a middle ground that shows something meaningful about an illuminant's color rendering accuracy that doesn't require a very long detour into complicated math. I can handle scripting ArgyllCMS (or other) command line programs.
Doug said in an earlier post that hue shifts are more important than saturation changes. If so, then doesn't this also apply to printer profiling? Aren't hue shifts meaningful when profiling printers? I don't think we can plot hue shifts with ArgyllCMS 3D plotting.
I like 3D plots because, as Doug said, they are pretty. I'd also like them to be meaningful.
-
You are not a color scientist, Alexey. You're a programmer as you've pointed out.
If you're so sure about this process then provide practical application that solves the problem of color reproduction using the photographic process.
As it is you don't know diddly squat about what you are doing. Get a degree in color science and come back and report that you solved a problem with that education that actually helped people, not just informed them. Talk is cheap and you're no expert. No one here is an expert, not even me.
I'm a pragmatist. I get things done! I don't talk about getting it done!
From the colour science perspective, you are utterly failing to refute anything Alexey is saying. Maybe it is you who are in need of an education to even begin understanding what is discussed here.
-
The problem is that getting the closest to some CCT ignores the green/magenta error, which is unlimited.
Could you explain the unlimited green/magenta error? Is this something about Duv?
And those (hue shifts) are what human vision is most sensitive to as saturation increases. In fact, generally human vision becomes less sensitive to changes in saturation as the saturation increases while hue shifts remain more perceivable. This is a principal reason I find 3D gamut images not very useful. It's more useful for showing saturation than hue shifts.
OK, 3D gamut plots aren't useful for comparing different measured lights against daylight, but what about the simpler case of comparing the gamut of a real world image against the gamut of a color space (i.e., sRGB) as a 3D plot? Illustrated by this plot of a synthetic Lab CC chart against sRGB (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/xd3-2018-small/CC-01_00-DCP-01_MacbethCC-LAB-L_sRGB_wrl.x3d.html) Compared to my real world shot of a CC chart shot under Walmart lights (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/xd3-2018-small/CC-01-01-CC-Walmart-01_IMG_0072_sRGB_wrl.x3d.html). In both cases cyan pokes out. (The point of the two comparisons is to show that the gamut of my real world shot of the CC chart is comparable to the gamut of the synthetic CC chart, in addition to showing the general validity of plotting image gamuts against color space gamuts.)
Under the assumption that you shouldn't use a color space that is larger than is needed. We all know approximately what the gamut of a CC chart is but we don't know the gamut of any arbitrary real world image. Isn't plotting the gamut of an image against a color space useful?
-
Could you explain the unlimited green/magenta error? Is this something about Duv?
Yes, For any given CCT, there are an infinite number of xy (or uv) coordinates with the same CCT. Duv is one way to specify how far away from the black body locus having that CCT a specific color not on the black body locus is.
That said, how small the Duv is together with the CCT only determines what the "white illuminant" is and says almost nothing about how well rendered actual objects are. Here's a specific example. There exists multiple pairs of two saturated single wavelengths, each at the boundaries of the CIExy human vision gamut, that can produce a "white" at exactly D50*. One might look like a strong cyan, the other a strong orange. Mixed together in the right proportions the result is "white." But if you then illuminate a colorchecker with those two wavelengths all the colors except for the 6 neutrals will look highly distorted. So focusing on how close a "white" is to D50 (or any other white point) tells only a small part of the story. You can have an extremely spectrally spikey illuminant that is very close to D50 yet is awful rendering colors and likely has a very poor CRI. As inadequate as CRI is, it's more useful than Duv.
http://www.brucelindbloom.com/Eqn_XYZ_to_T.html
*This assumes a person with color perception that matches the "Standard Observer" and for cases of only two wavelengths there are significant variations between individuals.
OK, 3D gamut plots aren't useful for comparing different measured lights against daylight, but what about the simpler case of comparing the gamut of a real world image against the gamut of a color space (i.e., sRGB) as a 3D plot? Illustrated by this plot of a synthetic Lab CC chart against sRGB (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/xd3-2018-small/CC-01_00-DCP-01_MacbethCC-LAB-L_sRGB_wrl.x3d.html) Compared to my real world shot of a CC chart shot under Walmart lights (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/xd3-2018-small/CC-01-01-CC-Walmart-01_IMG_0072_sRGB_wrl.x3d.html). In both cases cyan pokes out. (The point of the two comparisons is to show that the gamut of my real world shot of the CC chart is comparable to the gamut of the synthetic CC chart, in addition to showing the general validity of plotting image gamuts against color space gamuts.)
Under the assumption that you shouldn't use a color space that is larger than is needed. We all know approximately what the gamut of a CC chart is but we don't know the gamut of any arbitrary real world image. Isn't plotting the gamut of an image against a color space useful?
Usefulness depends. People understand comparisons in different ways. I'm somewhat numbers oriented and would rather see the 24 Lab numbers resulting from the Walmart light compared with the D50 standard. Especially with the dE00 difference on each patch. It's a short text listing. For me it's more informative. I do like comparing to Colorchecker patches since I'm more familiar with them.
-
Under the assumption that you shouldn't use a color space that is larger than is needed.
I think that is old-fashioned thinking from the bad old days of 8-bit precision. I am perfectly happy to use PPRGB at Ps's 15-bit-plus-one-value precision.
https://blog.kasson.com/?s=color+space+conversion
Jim
-
I think that is old-fashioned thinking from the bad old days of 8-bit precision. I am perfectly happy to use PPRGB at Ps's 15-bit-plus-one-value precision.
https://blog.kasson.com/?s=color+space+conversion (https://blog.kasson.com/?s=color+space+conversion)
Jim
Yes, that concept, outside the bit depth considerations is an urban legend, perhaps created on the DPR forums that are colorimetrically false. First, those who propose it can't tell us how we should determine how the image gamut would fit. Sure, you can render from raw into various color spaces and plot them. Image by image? More importantly, unnecessary. When I heard this silly idea that you should use a color gamut working space that isn't any larger than the image, I colorimetrically proved it wrong. I took an image that from raw can easily fit into sRGB, did so, then rendered into ProPhoto RGB RGB. There's no difference using either color space!
If an image from raw can't fit into the resulting color space gamut, you clip colors; not good.
If an image easily fits, using something significantly larger makes NO difference.
These are empty containers of differing sizes and without an encoded pixel, they contain nothing.
Anyway, here's a video I did, dedicated to a fellow who states this nonsense over and over again in another forum and can't explain nor prove what he states has any validity, because it doesn't.
sRGB urban legend Part 1
In this 30 minute video I'll cover:
Is there benefit or harm using a wider color gamut working space than the image data?
Should sRGB image data always be encoded into sRGB?
What are RGB working spaces and how do they differ?
What is Delta-E and how we use it to evaluate color differences.
Color Accuracy: what it really means, how we measure it!
Using Photoshop to numerically and visually see color differences when using differing working spaces.
Using ColorThink Pro and BableColor CT&A to show the effects of differing working space on our data and analyzing if using a smaller gamut working space is necessary.
Appendix: testing methodology, how differing raw converters encode into working spaces, capturing in-camera JPEG data and color accuracy.
Low resolution (YouTube): https://www.youtube.com/watch?v=1w0zUIl-dzY (https://www.youtube.com/watch?v=1w0zUIl-dzY)
High resoution: http://digitaldog.net/files/sRGBMyths.mp4 (http://digitaldog.net/files/sRGBMyths.mp4)
-
Anyway, here's a video I did, dedicated to a fellow who states this nonsense over and over again in another forum and can't explain nor prove what he states has any validity, because it doesn't.
sRGB urban legend Part 1
Low resolution (YouTube): https://www.youtube.com/watch?v=1w0zUIl-dzY (https://www.youtube.com/watch?v=1w0zUIl-dzY)
High resoution: http://digitaldog.net/files/sRGBMyths.mp4 (http://digitaldog.net/files/sRGBMyths.mp4)
As usual, another excellent video. I've watched many of yours but this is the first time I saw this one. Quite good and should help folks get more of a handle on colorspaces and discard misconceptions many seem to have acquired.
One note. You pointed out Photoshop truncates Lab values in the info panel. They finally fixed that! Photoshop now allows high resolution of Lab values in the info panel but only if you select 32bits and Lab. Then you can get things like (71.23, 19.84, 18.64). It works even with 8 bit RGB images and shows the actual Lab values fractionally.
-
I think that is old-fashioned thinking from the bad old days of 8-bit precision. I am perfectly happy to use PPRGB at Ps's 15-bit-plus-one-value precision.
https://blog.kasson.com/?s=color+space+conversion
Jim
I also doubt "need" is something easily predictable. I know of 4 versions of Monet's San Giorgio Maggiore by Twilight; all in different colour, and from measurements one version doesn't fit sRGB.
-
You pointed out Photoshop truncates Lab values in the info panel. They finally fixed that! Photoshop now allows high resolution of Lab values in the info panel but only if you select 32bits and Lab. Then you can get things like (71.23, 19.84, 18.64). It works even with 8 bit RGB images and shows the actual Lab values fractionally.
Yes but the newer values don't make a lot of sense to me when set to 16-bit, why?. Sampler point on some pixels with the readout set to 8bit first, then 16-bit next.
Indeed, 32 bit is the way to go....
-
Just discovered a cool trick: Make multiple sampler points on an image. Hold down Option key (Mac), toggle to Lab then toggle to 32 bit; all sampler points update to reflect that setting.
Too bad we can't set this in Preferences so it always reflects the settings on all documents. :-[ Everything defaults back to RGB/8-bits expect the 'main' readout. At least it's sticky.
-
Yes but the newer values don't make a lot of sense to me when set to 16-bit, why?. Sampler point on some pixels with the readout set to 8bit first, then 16-bit next.
Indeed, 32 bit is the way to go....
2 taken to the power of 15, 2^15 = 32768
44 / 100 * 2^15 = 14417.92
Given that 44 is probably rounded, seems close enough.
-
2 taken to the power of 15, 2^15 = 32768
44 / 100 * 2^15 = 14417.92
Given that 44 is probably rounded, seems close enough.
Thanks, good to know. I wonder however, why Adobe provided that option. Is there any reason to use it? It's just a readout that's quite foreign to me.
-
Yes, that concept, outside the bit depth considerations is an urban legend, perhaps created on the DPR forums that are colorimetrically false.
My assumption was based on several premises
- That smaller is smaller and is computationally cheaper. CPUs and storage space neither free nor infinite.
- If you do extensive editing in a smaller space there is less danger of posterization in gradients (skies) than if you do extensive editing in a larger space and then export to a smaller space.
- If you do extensive editing in a larger space (that contains colors that any real world monitor can't display) that there is increased danger of hue and saturation shifts when you export to a smaller space.
If you know beforehand how your real world image fits in various color spaces, so you don't have to spend a lot of time with Photoshop's crude gamut clipping soft proofing tools.
First, those who propose it can't tell us how we should determine how the image gamut would fit. Sure, you can render from raw into various color spaces and plot them. Image by image?
I am working on a tool that uses 100% free components and is easy to use. Where is the disadvantage?
More importantly, unnecessary. When I heard this silly idea that you should use a color gamut working space that isn't any larger than the image, I colorimetrically proved it wrong. I took an image that from raw can easily fit into sRGB, did so, then rendered into ProPhoto RGB RGB. There's no difference using either color space!
If an image from raw can't fit into the resulting color space gamut, you clip colors; not good.
If an image easily fits, using something significantly larger makes NO difference.
I already told you I watched your sRGB Myths video. I just watched it again. I obliquely referenced it in my post to Doug when I highlighted CC chart cyan poking out of sRGB. Your video covered a CC chart and an unedited image of a white dog and snow. Neither illustrate the concerns I raised in my list, above.
I also said that I did a random walk examining different spectral analysis tools. Babelcolor CC&T (http://www.babelcolor.com/cta_features.htm) ($125), Robin Myers Imaging SpectraShop 5 (http://www.rmimaging.com/spectrashop.html) ($99), and Chromix ColorThink Pro (https://www.chromix.com/colorthink/) ($399).
I started out using CC&T to compare Lab values of different CC charts (ones I shot under different light against the reference CC Lab values) and hand transcribing Lab values from the PhotoShop info panes into CC&T got tedious real quick. I wanted something that would compare the Lab values from each square of the CC charts as one operation. I examined all three programs and I don't think that any of them could do it. So I returned to my existing methods of running images (and color spaces) through the ArgyllCMS utilities to produce 3D plots. This is quick and easy--all I had to do was add the filenames of the various image files and ICC profiles to a configuration file and run my script. And the interactive HTMLish 3D plots are easily sharable. (ColorThink Pro's...?)
I also read Jim's blog posts
https://blog.kasson.com/the-last-word/the-color-reproduction-problem/ (https://blog.kasson.com/the-last-word/the-color-reproduction-problem/). As I posted earlier "In blog post 14 (https://blog.kasson.com/the-last-word/where-to-go-with-the-color-reproduction-series/) you noted how tedious it was to transcribe Lab numbers from Photoshop" and then the post referenced Matlab. I did a quick check of Matlab and that is "write for a quote", which I assumed didn't mean that it was cheap. So I kept on with my script and ArgyllCMS programs and asked for suggestions on how to improve my methodology.
My random walk continued. Last night I got to GNU Octave (https://www.gnu.org/software/octave/) which is free and claims to be "Drop-in compatible with many Matlab scripts". So possibly this could be used as a free untedious way of comparing CC chart Lab values? (And for doing many other things.)
But, not knowing anything about Matlab, I could use a head start in doing this.
-
Just discovered a cool trick: Make multiple sampler points on an image. Hold down Option key (Mac), toggle to Lab then toggle to 32 bit; all sampler points update to reflect that setting.
Too bad we can't set this in Preferences so it always reflects the settings on all documents. :-[ Everything defaults back to RGB/8-bits expect the 'main' readout. At least it's sticky.
Nice! One caution on the hi-rez (32 bit) Lab readouts. They are wrong when zoom levels are under 100%. Looks like Adobe is grabbing them from downsampled 8 bit data.
I do find the hi-rez stuff useful for things like examining the color shifts over a neutral gradient R=B=G in printer device space to see the sorts of color perturbations from device neutral. Big differences between my 9500 and 9800. The profiling software works harder (and needs more patches) in a few areas of the neutrals.
-
Thanks, good to know. I wonder however, why Adobe provided that option. Is there any reason to use it? It's just a readout that's quite foreign to me.
They've had that forever it seems. I've not found it particularly useful.
What would be nice is to enter fractional Lab values in the color dialog. It seems one is still limited to integer values.
-
I also read Jim's blog posts
https://blog.kasson.com/the-last-word/the-color-reproduction-problem/ (https://blog.kasson.com/the-last-word/the-color-reproduction-problem/). As I posted earlier "In blog post 14 (https://blog.kasson.com/the-last-word/where-to-go-with-the-color-reproduction-series/) you noted how tedious it was to transcribe Lab numbers from Photoshop" and then the post referenced Matlab. I did a quick check of Matlab and that is "write for a quote", which I assumed didn't mean that it was cheap. So I kept on with my script and ArgyllCMS programs and asked for suggestions on how to improve my methodology.
My random walk continued. Last night I got to GNU Octave (https://www.gnu.org/software/octave/) which is free and claims to be "Drop-in compatible with many Matlab scripts". So possibly this could be used as a free untedious way of comparing CC chart Lab values? (And for doing many other things.)
But, not knowing anything about Matlab, I could use a head start in doing this.
MATLAB is expensive. I've used the commercial version since the late 80's for unrelated stuff and the image toolbox add-in since 2003. However, they came out with a "home" version that's about 10% of the price though MATLAB and the Image Toolbox still comes to about $200.
Octave is quite good and continues to improve but, even with the image toolbox, has little for dealing with printer ICC profiles.
https://www.mathworks.com/products/matlab-home.html?s_tid=htb_learn_gtwy_cta4
ICC profile tranformations:
https://www.mathworks.com/help/images/ref/makecform.html
MATLAB makes it fairly easy to write your own functions. For instance this one finds the Lab skin color (70,20,20) printed in Perceptual Intent.
>> ProfileConvert([70 20 20], '9800 Costco Randomized I1P 2871pch.icm', 'f', 0, 'r', 3)
ans =
67.0665 18.1250 17.7305
Which is very close to what is actually printed if read with a spectro. Here ProfileConvert takes a Lab values, converts to device space using Perceptual "-f 0", then back to Lab retaining actual colors using Abs. Col. with the "-r 3"
-
I started out using CC&T to compare Lab values of different CC charts (ones I shot under different light against the reference CC Lab values) and hand transcribing Lab values from the PhotoShop info panes into CC&T got tedious real quick. I wanted something that would compare the Lab values from each square of the CC charts as one operation.
BabelColor PatchTool?
-
BabelColor PatchTool?
Yep, Patchtool works great for comparing any two colors lists in CGATs. Nice histogram and stats too.
-
My assumption was based on several premises
- That smaller is smaller and is computationally cheaper. CPUs and storage space neither free nor infinite.
- If you do extensive editing in a smaller space there is less danger of posterization in gradients (skies) than if you do extensive editing in a larger space and then export to a smaller space.
- If you do extensive editing in a larger space (that contains colors that any real world monitor can't display) that there is increased danger of hue and saturation shifts when you export to a smaller space.
If you know beforehand how your real world image fits in various color spaces, so you don't have to spend a lot of time with Photoshop's crude gamut clipping soft proofing tools.
I am working on a tool that uses 100% free components and is easy to use. Where is the disadvantage?
I already told you I watched your sRGB Myths video. I just watched it again. I obliquely referenced it in my post to Doug when I highlighted CC chart cyan poking out of sRGB. Your video covered a CC chart and an unedited image of a white dog and snow. Neither illustrate the concerns I raised in my list, above.
I also said that I did a random walk examining different spectral analysis tools. Babelcolor CC&T (http://www.babelcolor.com/cta_features.htm) ($125), Robin Myers Imaging SpectraShop 5 (http://www.rmimaging.com/spectrashop.html) ($99), and Chromix ColorThink Pro (https://www.chromix.com/colorthink/) ($399).
I started out using CC&T to compare Lab values of different CC charts (ones I shot under different light against the reference CC Lab values) and hand transcribing Lab values from the PhotoShop info panes into CC&T got tedious real quick. I wanted something that would compare the Lab values from each square of the CC charts as one operation. I examined all three programs and I don't think that any of them could do it. So I returned to my existing methods of running images (and color spaces) through the ArgyllCMS utilities to produce 3D plots. This is quick and easy--all I had to do was add the filenames of the various image files and ICC profiles to a configuration file and run my script. And the interactive HTMLish 3D plots are easily sharable. (ColorThink Pro's...?)
I also read Jim's blog posts
https://blog.kasson.com/the-last-word/the-color-reproduction-problem/ (https://blog.kasson.com/the-last-word/the-color-reproduction-problem/). As I posted earlier "In blog post 14 (https://blog.kasson.com/the-last-word/where-to-go-with-the-color-reproduction-series/) you noted how tedious it was to transcribe Lab numbers from Photoshop" and then the post referenced Matlab. I did a quick check of Matlab and that is "write for a quote", which I assumed didn't mean that it was cheap. So I kept on with my script and ArgyllCMS programs and asked for suggestions on how to improve my methodology.
My random walk continued. Last night I got to GNU Octave (https://www.gnu.org/software/octave/) which is free and claims to be "Drop-in compatible with many Matlab scripts". So possibly this could be used as a free untedious way of comparing CC chart Lab values? (And for doing many other things.)
But, not knowing anything about Matlab, I could use a head start in doing this.
Your assumptions are colorimetricly incorrect and with high bit data, data from all raw capture moot! You seem more concerned with computationally than image data, perhaps a faster more modern computer is necessary. There is zero harm using a wide gamut processing color space which is being used in Adobe raw processed data, and lots of potential data loss encoding after that wide gamut process due to color clipping using a smaller gamut. But it’s your data.
-
> That smaller is smaller and is computationally cheaper.
Not in the least, you still need the same number of bits for each pixel, independent of colour space used. 8-bit images are not best practices anyway for image preservation purposes.
More, smaller colour spaces may turn out to be computationally more demanding because of dealing with the edits that result in out-of-gamut colours.
> CPUs and storage space neither free nor infinite.
CPU use have no bearing here, also because modern algorithms operate with images on a GPU, mostly in 16 bits floating point.
Storage space depends on the number of images to be stored, but storing images in 8 bits is not an option for many of us.
> If you do extensive editing in a smaller space there is less danger of posterization in gradients (skies) than if you do extensive editing in a larger space and then export to a smaller space.
I would love to see a demonstration of this. Normally, when colorimetric intent is used to map from larger space to a smaller space, that's were posterisation tends to kick in. Raw processing colour spaces are rather wide.
> If you do extensive editing in a larger space (that contains colors that any real world monitor can't display) that there is increased danger of hue and saturation shifts when you export to a smaller space.
But that means that you are suggesting to work in destination colour space. That was tried for 15+ years and proved to be not satisfactory. Colour management was adopted as a better solution. Dangers are everywhere, sometimes it's good to put a number and an example on a danger, otherwise it is rhetorics a la DPR.
-
Your assumptions are colorimetricly incorrect and with high bit data, data from all raw capture moot! You seem more concerned with computationally than image data, perhaps a faster more modern computer is necessary.
My concern is finding better methods for evaluating light sources. My comment about small vs. large color spaces was an offhand remark and is diverting attention away from this, so let's drop it. It isn't a hill worth fighting over. In this thread.
-
> If you do extensive editing in a larger space (that contains colors that any real world monitor can't display) that there is increased danger of hue and saturation shifts when you export to a smaller space.
But that means that you are suggesting to work in destination colour space. That was tried for 15+ years and proved to be not satisfactory. Colour management was adopted as a better solution. Dangers are everywhere, sometimes its good to put a number and an example on a danger, otherwise it is rhetorics a la DPR.
I was referring to editing in Prophoto and exporting to sRGB. Sorry I wasn't more specific.
For the rest of it, my original comment was offhand and I'll yield to you. My intent in this thread is trying to find better ways of evaluating light sources. Arguing about editing in large vs. small color spaces is a divergence I didn't intend.
-
My concern is finding better methods for evaluating light sources. My comment about small vs. large color spaces was an offhand remark and is diverting attention away from this, so let's drop it. It isn't a hill worth fighting over. In this thread.
Off hand and incorrect. Now you can avoid repeating it! ;D
-
Off hand and incorrect. Now you can avoid repeating it! ;D
For my use case images mostly end up as sRGB, so running my NEC PA241W in sRGB emulation mode, editing in sRGB in 16 bit and only exporting 8 bit sRGB JPEGs as the final step minimizes hue shift and posterization artifacts. Use cases need to be established before making claims.
-
For my use case images mostly end up as sRGB, so running my NEC PA241W in sRGB emulation mode, editing in sRGB in 16 bit and only exporting 8 bit sRGB JPEGs as the final step minimizes hue shift and posterization artifacts. Use cases need to be established before making claims.
Thats all good and fine but doesn’t change your incorrect impression about the size of a color gamut you have provided and believe (believed?). Several posters have told you why.
What color shifts and posterization?
-
For my use case images mostly end up as sRGB, so running my NEC PA241W in sRGB emulation mode, editing in sRGB in 16 bit and only exporting 8 bit sRGB JPEGs as the final step minimizes hue shift and posterization artifacts. Use cases need to be established before making claims.
Before taking any use cases into an account one needs to check how good is the workflow supporting the use case, and how valid is the use case itself.
Before asking a generic question about evaluating light source quality it may be worth asking a couple of different questions, like what is the use case, because when it comes to photography, the properties of sensitive material matter. Same light source may be good for one camera or film and not so good for another. CIPA documents, however, have some recommendations.
-
Thats all good and fine but doesn’t change your incorrect impression about the size of a color gamut you have provided and believe (believed?).
I'm confused. I've been referring to sRGB, Adobe RGB (1998) and Prophoto. These color spaces. (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/xd3-2018-small/sRGB_AdobeRGB_ProPhoto_wrl.x3d.html) And showing samples of images whose gamuts exceeded sRGB (http://www.frogymandias.org/gamuts/gamuts-2018-xd3-html/xd3-2018-small/Set-04-Aputure-Lab_IMG_0081_sRGB_wrl.x3d.html). Presented like this because it is easier to visualize as an interactive 3D plot than it is with Photoshop's soft proof out-of-gamut tool.
Assuming that the tool is quick and easy to use, like mine is.
What color shifts and posterization?
The kind that can happen if an image is edited in a large space, such as Prophoto, and is exported as sRGB without careful examination. If the gamut of the image exceeds sRGB, which can happen to an image that didn't start out exceeding sRGB, but had had, say, the saturation pumped up while the image was in Prophoto (which has colors that aren't visible on any known monitor.)
Anther use case. I've done a lot of camera scanning of negatives and I've found that it is best to use Prophoto throughout when inverting and color processing the negative. Often when I'm done editing, I have to carefully unsaturate a few colors while in ProPhoto before converting it to sRGB. Possibly images that don't originate as film negatives aren't as problematic when converting from Prophoto to sRGB.
My modern "originated as digital" images usually are in sRGB through the whole process, from ACR out to saving as 8 bit JPEG (in addition to as 16 bit TIFF.)
The controversy might be "people who place a lot of emphasis on printing" talking past people that don't. And vice versa.
-
I'm confused.
Let's examine the text that's not Kosher:
Under the assumption that you shouldn't use a color space that is larger than is needed.
Let's go there first (or not) then move onto 'color shifts' that have been requested to be shown. Again, that assumption above isn't correct. I've shown the use of a small color gamut working space and a very large one with an image that easily fits within that smaller color gamut working space. It makes no difference using the smaller or the larger gamut color space; that's the point and the correction. Therefore you can use a color space that's larger than the image and that larger gamut could very well be needed on other images. Doing so is just fine. Not doing so could and often does, depending on that smaller color space (sRGB as an example but Adobe RGB (1998) too) results in color data loss. Clipping of colors. There's no reason to do this. You gain nothing and lose something. The assumption that you shouldn't use a color space that is larger than necessary (and how you'd do so or why) doesn't wash. Just use the largest container, working space gamut, you can in high bit. Forget worrying about what size the image gamut or scene gamut may be, or trying to fit it into something just large enough to contain it. It's work that few can do and it is unnecessary to do anyway. You're shooting raw? Your converter uses a very large processing color space gamut for a reason and encoding that into a smaller working space gamut serves zero purpose and can only result in clipping.
Some people here place a lot of emphasis on not tossing data you can capture and output. Today or in the future.
-
My random walk continued. Last night I got to GNU Octave (https://www.gnu.org/software/octave/) which is free and claims to be "Drop-in compatible with many Matlab scripts". So possibly this could be used as a free untedious way of comparing CC chart Lab values? (And for doing many other things.)
But, not knowing anything about Matlab, I could use a head start in doing this.
If you don't want to pay for the home license of Matlab, Octave is a decent way to go (as is Python). If you do decide to use Matlab, OptProp will make your life easier:
https://www.mathworks.com/matlabcentral/fileexchange/13788-optprop-a-color-properties-toolbox
Unfortunately, it hasn't been updated in a while, and its clever argument-passing methods are likely to soon be obsolete.
It may run in Octave. I never tried.
Jim
-
Some people here place a lot of emphasis on not tossing data you can capture and output. Today or in the future.
Which is why I shoot raw and keep a death grip on my raw files. If I ever need to use a larger color space for an image (i.e., want to have a large print made), then I'd set my PA241W to "native", go back to the raw file and reprocess it from scratch. For that particular printer.
Starting in a large color space and editing colors for an unknown printer's gamut seems to me an awful lot like editing for generic CYMK. Which I believe that you and other members of the Pixel Mafia have cautioned against.
-
Before taking any use cases into an account one needs to check how good is the workflow supporting the use case, and how valid is the use case itself.
Before asking a generic question about evaluating light source quality it may be worth asking a couple of different questions, like what is the use case, because when it comes to photography, the properties of sensitive material matter. Same light source may be good for one camera or film and not so good for another. CIPA documents, however, have some recommendations.
From here? (http://www.cipa.jp/index_e.html) Membership seems to be required to access the papers.
My premise the entire time I've participated in this thread is that the existing metrics of evaluating light (CRI, TLCI, and the R numbers) leave a lot to be desired when evaluating light. And that, even though cameras have the Luther/Ives issue, that, in general, light that renders color well for human eyes also renders light better for cameras.
After learning (on this thread) that there are a lot more nuances of testing camera response than I knew before, I'm especially suspicious of the single scalar TLCI "one size fits all" metric. At least CRI is aligned with the time tested "CIE standard observer."
New topic: while I've been researching during the course of this thread, I rediscovered DCamProf (https://www.ludd.ltu.se/~torger/dcamprof.html) and Lumariver Profile Designer (http://www.lumariver.com/#LumariverPD) (I've been away from LuLa for several years.) Should I spend more time with one or both of those before waiting for better color rendering metrics? In addition to studying ArgyllCMS (and a lot of other things) closer?
-
If you don't want to pay for the home license of Matlab, Octave is a decent way to go (as is Python). If you do decide to use Matlab, OptProp will make your life easier:
https://www.mathworks.com/matlabcentral/fileexchange/13788-optprop-a-color-properties-toolbox
Unfortunately, it hasn't been updated in a while, and its clever argument-passing methods are likely to soon be obsolete.
It may run in Octave. I never tried.
Jim
I just realized that I have already done some baby Matlab programming. I set up an account on ThingSpeak (https://thingspeak.com/) (which is a branch of Matlab) about a year ago to monitor several ESP8266 temp/humidity sensors I put together and adjusted the sample Matlab code in the web interface editor.
I don't mind paying for Matlab home for my own purposes but I'll like to make tools that can be freely shared with others. But probably the only feasible way to do this is to start with full blooded Matlib (that has support for icc profiles) and then figure out how to achieve the same thing in Octave (ArgyllCMS iccdump (https://argyllcms.com/doc/iccdump.html) (convert an ICC profile to ASCII text) might be a good starting point.
But I'm also losing my momentum. I'm trying to talk about building tools to share but that portion of my posts always gets clipped away in responses. After a certain point...
Thanks for your help.
-
I just realized that I have already done some baby Matlab programming. I set up an account on ThingSpeak (https://thingspeak.com/) (which is a branch of Matlab) about a year ago to monitor several ESP8266 temp/humidity sensors I put together and adjusted the sample Matlab code in the web interface editor.
I don't mind paying for Matlab home for my own purposes but I'll like to make tools that can be freely shared with others. But probably the only feasible way to do this is to start with full blooded Matlib (that has support for icc profiles) and then figure out how to achieve the same thing in Octave (ArgyllCMS iccdump (https://argyllcms.com/doc/iccdump.html) (convert an ICC profile to ASCII text) might be a good starting point.
But I'm also losing my momentum. I'm trying to talk about building tools to share but that portion of my posts always gets clipped away in responses. After a certain point...
Thanks for your help.
Home Matlab is like the full version except it's pretty much limited to non commercial, non institutional uses. The image toolbox is only $45 more and it has full ICC support. Also has a nice editor/debugger and is well integrated with Git source control so easy to keep track of your stuff. Most functions your write can be shared but if you charge money for them or are paid for your work then you need to buy the commercial version. I have both but my commercial version is a few years old.
-
> Membership seems to be required to access the papers.
http://www.cipa.jp/std/documents/e/DC-004_EN.pdf page 19.
> the existing metrics of evaluating light
... are not for photography. Sensor measures light differently from how we perceive it. Sensor / Film metameric error is not accounted for in those metrics.
> light that renders color well for human eyes also renders light better for cameras.
Hm... Some cameras, like Nikon D3..D5 series, are optimized for artificial light common at sports venues. It's a complex process, that optimization, causing a rather significant deviation from L-I condition to present not the colours that we see under that artificial light but rather the colours we would see if the light would be closer to daylight.
> while I've been researching during the course of this thread
I think systematic study starts with textbooks on colour science.
-
> Membership seems to be required to access the papers.
http://www.cipa.jp/std/documents/e/DC-004_EN.pdf page 19.
Thank you. For simulating daylight they list using Hoya LB-120 (https://www.sydor.com/wp-content/uploads/Hoya-LB-120-Light-Balancing-Blue-Filter.pdf) (in front of 3200k tungsten) which isn't available at B&H. I have some Rosco #3202 CTB
(https://static.bhphoto.com/images/multiple_images/images500x500/1231962789000_IMG_66684.jpg)
Is that close enough? (I have a Variac for adjusting a tungsten bulb. And, of course, ColorMeter for reading CCT.) I don't have a spot luminance meter so I need to get one.
I'm not sure where to get an IEC 1146-1 Test chart No. 1 or an ITE Grayscale Chart II.
With said, I'm not sure exactly what is being measured.
> the existing metrics of evaluating light
... are not for photography. Sensor measures light differently from how we perceive it. Sensor / Film metameric error is not accounted for in those metrics.
Yes, this has been covered extensively earlier in this thread. I'm aware of the issues.
> light that renders color well for human eyes also renders light better for cameras.
Hm... Some cameras, like Nikon D3..D5 series, are optimized for artificial light common at sports venues. It's a complex process, that optimization, causing a rather significant deviation from L-I condition to present not the colours that we see under that artificial light but rather the colours we would see if the light would be closer to daylight.
Which is why I asked whether Anders Torgers' DCamProf and/or Lumariver Profile Designer would be helpful. Both seemed to be received well on LuLa forums.
> while I've been researching during the course of this thread
I think systematic study starts with textbooks on colour science.
I've already posted that I've reread
Color Management for Photographers: Hands on Techniques for Photoshop Users (https://www.amazon.com/Color-Management-Photographers-Techniques-Photoshop/dp/0240806492/ref=sr_1_1?ie=UTF8&qid=1529341229&sr=8-1&keywords=andrew+rodney)
Real World Color Management (https://www.amazon.com/Real-World-Color-Management-2nd/dp/0321267222/ref=sr_1_1?ie=UTF8&qid=1529341318&sr=8-1&keywords=real+world+color+management)
Billmeyer and Saltzman's Principles of Color Technology 2nd edition (https://www.amazon.com/Billmeyer-Saltzmans-Principles-Color-Technology/dp/047119459X/ref=sr_1_25?s=books&ie=UTF8&qid=1529341434&sr=1-25&keywords=color+science)
Color Science: Concepts and Methods, Quantitative Data and Formulae 2nd Edition (https://www.amazon.com/Color-Science-Concepts-Quantitative-Formulae/dp/0471399183/ref=sr_1_2?s=books&ie=UTF8&qid=1529341506&sr=1-2&keywords=color+science) (selectively.)
during the course of this thread, along with spidering external references (such as Jim's blog posts) that were posted. Alexey recommended Measuring Colour (https://www.amazon.com/Measuring-Colour-Imaging-Science-Technology-ebook/dp/B005HF2GOQ/ref=sr_1_1?s=books&ie=UTF8&qid=1529341971&sr=1-1&keywords=measuring+color+hunt) but I've been hoping to find a used copy for less than ~$90.
Anders extensively tested light sources for camera profiling (https://www.ludd.ltu.se/~torger/photography/camera-profiling.html#artificial_light) and found even the best LEDs (including Yuji) might not be an improvement over CFLs. (He recommended using Solux bulbs carefully) But Alexey is working on a Yuji COB for Spectron (https://github.com/Alexey-Danilchenko/Spectron). Are there different criteria for camera profiling and for what Spectron is designed to measure?
-
> For simulating daylight they list using Hoya LB-120 (in front of 3200k tungsten) which isn't available at B&H. I have some Rosco #3202 CTB Is that close enough?
No, and Rosco filters are not manufactured to the tolerances needed.
> Anders extensively tested light sources for camera profiling and found even the best LEDs (including Yuji) might not be an improvement over CFLs. But Alexey is working on a Yuji COB for Spectron
Our profiling with Spectron is spectrum-based, using a monochromator. LEDs are not used to illuminate the target is such setups, there is no target present at all. All that is needed is a constant spectrum power distribution (something that you can't achieve with SoLux) and a good coverage of all wavelengths. There is a reason why light sources for monochromators are so expensive. We are making a much cheaper and better suited for photographical purposes one. Shooting targets is a very unreliable profiling method, prone to all kinds of problems with flare, uneven illumination, uneven "white balance", etc. No serious setup should use target shooting (yes, it is much better than nothing, but a guaranteed setup needs very stable set of four high-grade studio flash units and a lot of additional equipment, making the whole thing economically inefficient), and in fact in many cases target shooting is already replaced with spectral profiling. Transition started 15+ years ago.
-
Anders extensively tested light sources for camera profiling (https://www.ludd.ltu.se/~torger/photography/camera-profiling.html#artificial_light)
To be honest nothing on that link suggests Anders actually tested all of the lights. He also explicitly states he has not used them all and simply considered their respective SPD and decided to stay with filtered tungsten halogen. It also quite specifically indicated he had not considered LEDs (due to their "bumpy spectrum"), not even Yuji. So in reality no practical testing has been done to draw any conclusions.
-
We are making a much cheaper and better suited for photographical purposes one.
Bearing in mind that the standard for "photographical purposes" on this thread is being able to buy it at B&H (http://forum.luminous-landscape.com/index.php?topic=123799.msg1033505#msg1033505), it looks like you have a ways to go. (https://github.com/Alexey-Danilchenko/Spectron) :)
That aside, I'd like to see your project succeed. If, as you say, that evaluating light spectrums and profiling cameras is mostly futile. I think that I can manage building one. There are a lot of pieces in the project. I was planning on starting by experimenting with a Particle Photon to calibrate myself on managing this particular microcontroller.
I'm guessing that a lot of the rest of the project is in flux while Alexey finishes the COB project, correct?
-
Bearing in mind that the standard for "photographical purposes" on this thread is being able to buy it at B&H (http://forum.luminous-landscape.com/index.php?topic=123799.msg1033505#msg1033505), it looks like you have a ways to go. (https://github.com/Alexey-Danilchenko/Spectron) :)
That aside, I'd like to see your project succeed. If, as you say, that evaluating light spectrums and profiling cameras is mostly futile. I think that I can manage building one. There are a lot of pieces in the project. I was planning on starting by experimenting with a Particle Photon to calibrate myself on managing this particular microcontroller.
I'm guessing that a lot of the rest of the project is in flux while Alexey finishes the COB project, correct?
The idea is to share what we know and thoroughly tested, to provide some help and reliable reference (schematics, PCB design, etc) to those who may need it; and of course to make it reasonably easy for DIY and not overly expensive (comparable or cheaper than the regular profiling setup). The method by itself is not a unique and a very logical one, it became a relative commonplace in 2009. Far from being the first, in 2006 I used off-the-shelf equipment (monochromator, integrating sphere, spectrometer, and light source) for this (because I can afford the expense). Resulting colour transforms were incorporated in RPP about the same time.
LEDs are the best solution that we found (for multiple reasons, including being DIY-friendly), but halogen lamps and xenon flashlamps in strobe mode were also tested to work well. As to COBs, we don't expect any surprises with the recent designs we tried, so it is mostly a matter of testing reliability now, and it takes time. One can come up with his own design, so I don't see how what we are doing in terms of light source design is holding things off for those who are comfortable with EE ;)
B&H: they sell some of our software, I don't see why they wouldn't be interested in selling hardware for spectral profiling in some future. It may be based on this project, or on some other, including by a different team. If one needs to profile lens+camera combos, creating multiple profiles, he may be interested to have it available. But mostly it is about renting and service.
-
... is holding things off for those who are comfortable with EE ;)
"comfortable with EE" will probably considerably reduce the pool of photographers that would benefit. I'd like to increase the pool size. To at least include me.
The Particle Photon part is probably the easiest part of the project (to implement, not to design!) I looked at the PP code and I'm impressed with the clarity and professionalism. (Much better than any of my own code.) The Particle Photon documentation (https://docs.particle.io/guide/getting-started/intro/photon/) indicates that flashing the firmware should be easy. After getting the build environment set up and doing the "Hello world" thing, it looks like all the user needs to do with the code is transcribe the calibration data for his/her own Hamamatsu sensor for the "#define CALIBRATION_15F00163" statement.
Based on my reading several issues of Make magazine (https://makezine.com/), probably getting the PCBs made will be even easier (similar to having photo prints made by a printing service.)
Sourcing components isn't difficult--just tedious. But soldering surface mount components is difficult. I've never done it myself but I think I can persuade somebody I know to show me how. Ditto with fabricating COB mounting plates and heatsinks.
A good next step for the project would be to separate the easy-but-tedious portions from the difficult-no-matter-what portions. With the goal of encapsulating and reducing the latter. And (big gulp) make it understandable to non-EE photographers.
Any sufficiently advanced technology is indistinguishable from magic.
Clarke's third law
Please make it a bit less magical.
-
"comfortable with EE" will probably considerably reduce the pool of photographers that would benefit. I'd like to increase the pool size. To at least include me.
The Particle Photon part is probably the easiest part of the project (to implement, not to design!) I looked at the PP code and I'm impressed with the clarity and professionalism. (Much better than any of my own code.) The Particle Photon documentation (https://docs.particle.io/guide/getting-started/intro/photon/) indicates that flashing the firmware should be easy. After getting the build environment set up and doing the "Hello world" thing, it looks like all the user needs to do with the code is transcribe the calibration data for his/her own Hamamatsu sensor for the "#define CALIBRATION_15F00163" statement.
Based on my reading several issues of Make magazine (https://makezine.com/), probably getting the PCBs made will be even easier (similar to having photo prints made by a printing service.)
Sourcing components isn't difficult--just tedious. But soldering surface mount components is difficult. I've never done it myself but I think I can persuade somebody I know to show me how. Ditto with fabricating COB mounting plates and heatsinks.
A good next step for the project would be to separate the easy-but-tedious portions from the difficult-no-matter-what portions. With the goal of encapsulating and reducing the latter. And (big gulp) make it understandable to non-EE photographers.
Any sufficiently advanced technology is indistinguishable from magic.
Clarke's third law
Please make it a bit less magical.
If you are going to use SMT components, an excellent video instruction is https://youtu.be/M0wI-5YZQm4 (other Marc's videos on his channel, too). He shows different tools and explains what each is used for, and how to operate it. Equipment varies, technique stays the same. The secret to good and reliable soldering is understanding and feeling the surface tension, it does all the job, even small component alignment. You may want to use some SMT training boards.
Some PCB manufacturing services also offer the assembly.
Sourcing components is a matter of a shopping list.
Not sure I understand what exactly you mean by "easy-but-tedious portions" and "difficult-no-matter-what portions".
-
Not sure I understand what exactly you mean by "easy-but-tedious portions" and "difficult-no-matter-what portions".
"Easy-but-tedious" would be going through all the steps that are required for things like flashing the Particle Photons. And figuring out where to buy all the components and having the PCBs made. It is easy (but maybe tedious) to write a tutorial to walk people through doing this. (People don't need to learn everything there is to know about Particle Protons to just flash the firmware with the software you have already written. A tutorial makes the process easier.)
Easy-but-tedious can be explained with a digestible tutorial.
"Difficult-no-matter-what" is reading schematics and populating the PCBs with SMT components and fabricating aluminum parts for the COBs. And EE. These are more like ten-thousand-hour rule (https://www.newyorker.com/news/sporting-scene/complexity-and-the-ten-thousand-hour-rule) things.
An example of difficult-no-matter-what is playing an advanced musical piece. There is no way to write a digestible tutorial. But an advanced musical piece can be encapsulated (with a recording.)
I believe that you have explained that evaluating the color rendering ability of light sources and shooting targets for profiling cameras are both very problematic and that spectral profiling is a much better approach for photographers that want to accurately capture color. If my interpretation is correct, then I want make an effort to make spectral profiling be more accessible. If possible.
-
> "Difficult-no-matter-what" is reading schematics and populating the PCBs with SMT components
But as I said, the populating of PCBs can be outsourced.
> An example of difficult-no-matter-what is playing an advanced musical piece.
Populating PCBs for Spectron and light sources is nowhere near playing an advanced musical piece, those PCBs are intentionally designed to be cheap and simple to solder and use.
-
Populating PCBs for Spectron and light sources is nowhere near playing an advanced musical piece, those PCBs are intentionally designed to be cheap and simple to solder and use.
OK, I'll have to wait until I have everything in front of me.
My copy of Hunt and Pointer's Measuring Color just arrived. Quickly skimming through it, what you have been talking about makes a lot more sense. None of the other color science books I described earlier covered any of this (what is covered starting in chapter nine of Measuring Color.) The applied science portions of the older books were more about training chemists how to choose pigments so that plastic buttons matched sweater fabrics. It was difficult to extrapolate that into spectral profiling of cameras.
Sometimes it is difficult to reduce unknown unknowns to known unknowns.
-
The Particle Photon documentation indicates that flashing the firmware should be easy. After getting the build environment set up and doing the "Hello world" thing, it looks like all the user needs to do with the code is transcribe the calibration data for his/her own Hamamatsu sensor for the "#define CALIBRATION_15F00163" statement.
Flashing is easy and can be done from any build environment, including Particle's online web IDE.
The calibration data is stored in an array internally but because it needs to be passed to controlling software as well, initialisation is done as a string from a single sysdef which you can override with your data. The future versions will include a changeable calibration data to allow updating it fro outside (following a spectrometer calibration externally).
But soldering surface mount components is difficult. I've never done it myself but I think I can persuade somebody I know to show me how. Ditto with fabricating COB mounting plates and heatsinks.
This is what this project is all about to come up with fairly easy to repeat alternative. After I tried a few variations, I came up with the fairly easy one - 4 x LED COBs mounted on commonly available fanned heatsink (this one from Amazon (https://www.amazon.co.uk/gp/product/B01CW881TO/ref=oh_aui_detailpage_o01_s00?ie=UTF8&psc=1)) by means of Arctic Thermal Epoxy. A bit of soldering to connect LED COBs in pairs and the board on top and all is done. Takes a day (for epoxy to dry) to get the fully operational solution.
As for the light sources - I am about to submit all of that (having tested all the LEDs to death now and finishing Xenon pulsed light source testing) but it will take me a few days to finish write ups. Hopefully will be on GitHub by the end of upcoming week.
-
My copy of Hunt and Pointer's Measuring Color just arrived. Quickly skimming through it, what you have been talking about makes a lot more sense. None of the other color science books I described earlier covered any of this (what is covered starting in chapter nine of Measuring Color.)
This book is essential if you want to venture into colour measurements. It was not leaving my desk pretty much for the last two years when Iliah and I started working on Spectron. The light sources section there is very interesting. Also the chapter about Precision and Accuracy of the measurements if of particular interest to me.
-
OK, I'll have to wait until I have everything in front of me.
My copy of Hunt and Pointer's Measuring Color just arrived. Quickly skimming through it, what you have been talking about makes a lot more sense. None of the other color science books I described earlier covered any of this (what is covered starting in chapter nine of Measuring Color.) The applied science portions of the older books were more about training chemists how to choose pigments so that plastic buttons matched sweater fabrics. It was difficult to extrapolate that into spectral profiling of cameras.
Sometimes it is difficult to reduce unknown unknowns to known unknowns.
Measuring Color is the first step. Preferred reproduction of colour to follow. For a studio scene under halogen lights, in double blind test, participants consistently select a reproduction that resulted from a colour transform with mean dE 5.5, max dE 11 over the one resulting from colour transform with mean dE 3, max dE 5
-
Measuring Color is the first step. Preferred reproduction of colour to follow.
This? The Reproduction of Colour 6th Edition (https://www.amazon.com/Reproduction-Colour-R-W-Hunt/dp/0470024259/ref=pd_sim_14_4?_encoding=UTF8&pd_rd_i=0470024259&pd_rd_r=db95ea25-7803-11e8-b8f8-cb66dcdfd92f&pd_rd_w=bDCnr&pd_rd_wg=lhonb&pf_rd_i=desktop-dp-sims&pf_rd_m=ATVPDKIKX0DER&pf_rd_p=7967298517161621930&pf_rd_r=Y08DN4DYKPZ4KPGEKFFQ&pf_rd_s=desktop-dp-sims&pf_rd_t=40701&psc=1&refRID=Y08DN4DYKPZ4KPGEKFFQ)
Or this? Colour Reproduction in Electronic Imaging Systems: Photography, Television, Cinematography 1st Edition (https://www.amazon.com/Colour-Reproduction-Electronic-Imaging-Systems/dp/1119021766/ref=pd_sim_14_2?_encoding=UTF8&pd_rd_i=1119021766&pd_rd_r=db95ea25-7803-11e8-b8f8-cb66dcdfd92f&pd_rd_w=bDCnr&pd_rd_wg=lhonb&pf_rd_i=desktop-dp-sims&pf_rd_m=ATVPDKIKX0DER&pf_rd_p=7967298517161621930&pf_rd_r=Y08DN4DYKPZ4KPGEKFFQ&pf_rd_s=desktop-dp-sims&pf_rd_t=40701&psc=1&refRID=Y08DN4DYKPZ4KPGEKFFQ)
Or something different?
For a studio scene under halogen lights, in double blind test, participants consistently select a reproduction that resulted from a colour transform with mean dE 5.5, max dE 11 over the one resulting from colour transform with mean dE 3, max dE 5
Meaning two different color transforms? The first one provides more pleasing results even though it has a higher dE?
-
Hunt's The Reproduction of Colour is IMHO very worth having.
> Meaning two different color transforms? The first one provides more pleasing results even though it has a higher dE?
Exactly.
-
This? For a studio scene under halogen lights, in double blind test, participants consistently select a reproduction that resulted from a colour transform with mean dE 5.5, max dE 11 over the one resulting from colour transform with mean dE 3, max dE 5
Or something different?
Meaning two different color transforms? The first one provides more pleasing results even though it has a higher dE?
Good grief! You might as well have a forger reproduce the paintings considering how many folks have lost millions thinking they had the real thing.
Those crappy Delta E numbers as the accepted color match pretty much makes this thread a waste of time.
-
Or something different?
Meaning two different color transforms? The first one provides more pleasing results even though it has a higher dE?
Good grief! You might as well have a forger reproduce the paintings considering how many folks have lost millions thinking they had the real thing.
Those crappy Delta E numbers as the accepted color match pretty much makes this thread a waste of time.
Reproducing an artwork to make a copy that looks the same side by side, is not the same thing as picking a reproduction on it's own that has the most appealing appearance. Nothing at all surprising about that. Quite a different issue than forgery.
There's a study out there using tunable, bumpy, illuminants which has the effect of generally increasing saturation over a Planckian locus or D source at the same CCT. It shows people finding the illuminated objects generally more appealing.
-
Reproducing an artwork to make a copy that looks the same side by side, is not the same thing as picking a reproduction on it's own that has the most appealing appearance. Nothing at all surprising about that. Quite a different issue than forgery.
There's a study out there using tunable, bumpy, illuminants which has the effect of generally increasing saturation over a Planckian locus or D source at the same CCT. It shows people finding the illuminated objects generally more appealing.
Doug, a "Pleasing Appearance" was not what I got out of this entire discussion going by the ton of technical minutiae generated in this thread on getting precise color match from paintings viewed and photographed under a given light spectra quality. This was what Wayne was going for, the ultimate light that would give a color match. So a halogen will suffice for getting pleasing color but not a color match according to Delta E numbers.
If pleasing color was the goal in this thread, could you point me to the posts that indicated this was the main objective?
-
Something else I just noticed about this thread on light spectra for photographic reproduction of paintings...
No one mentioned the camera profile as a factor affecting Delta E numbers.
-
Doug, a "Pleasing Appearance" was not what I got out of this entire discussion going by the ton of technical minutiae generated in this thread on getting precise color match from paintings viewed and photographed under a given light spectra quality. This was what Wayne was going for, the ultimate light that would give a color match. So a halogen will suffice for getting pleasing color but not a color match according to Delta E numbers.
If pleasing color was the goal in this thread, could you point me to the posts that indicated this was the main objective?
It wasn't and no one here is saying it was. It's just an interesting side note. I was responding to the comment reply #254.
-
This? The Reproduction of Colour 6th Edition.
Or this? Colour Reproduction in Electronic Imaging Systems: Photography, Television, Cinematography 1st Edition.
This, (https://www.amazon.com/Measuring-Colour-R-W-Hunt/dp/1119975379/ref=sr_1_1?ie=UTF8&qid=1529914888&sr=8-1&keywords=measuring+color) Hunt's 'Reproduction' book is outdated imho.
-
No one mentioned the camera profile as a factor affecting Delta E numbers.
Hi Tim, that's what the transforms mentioned above are all about.
Jack
-
This, (https://www.amazon.com/Measuring-Colour-R-W-Hunt/dp/1119975379/ref=sr_1_1?ie=UTF8&qid=1529914888&sr=8-1&keywords=measuring+color) Hunt's 'Reproduction' book is outdated imho.
They are about completely different topics though.
-
Hi Tim, that's what the transforms mentioned above are all about.
Jack
Sad that once again, we have to point out the obvious.... ;)
Those crappy Delta E numbers as the accepted color match pretty much makes this thread a waste of time.
Please move on and stop waiting your precious time... :P
-
This, (https://www.amazon.com/Measuring-Colour-R-W-Hunt/dp/1119975379/ref=sr_1_1?ie=UTF8&qid=1529914888&sr=8-1&keywords=measuring+color) Hunt's 'Reproduction' book is outdated imho.
But a great way to get a historical perspective from the end of the film era. First color science book I bought, in 1990, after Kodak paid a visit to my lab and asked if we were interested in some joint work on color file format standards. It felt kind of outdated (and wandering) even then. Measuring Colour was the Jack Webb, "just the facts, ma'am" alternative (though, as Alexey points out, the thrusts of the two books are different), and I found that a few weeks later.
Jim
-
This, (https://www.amazon.com/Measuring-Colour-R-W-Hunt/dp/1119975379/ref=sr_1_1?ie=UTF8&qid=1529914888&sr=8-1&keywords=measuring+color) Hunt's 'Reproduction' book is outdated imho.
I modified my post to include the 2nd book while Iliah was responding to my original post that only mentioned The Reproduction of Colour 6th Edition. I don't think that Iliah saw Colour Reproduction in Electronic Imaging Systems: Photography, Television, Cinematography 1st Edition.
[Edit. I thought you were referring to Colour Reproduction in Electronic Imaging Systems: Photography, Television, Cinematography 1st Edition. I just clicked your link. I just got Measuring Colour 4th Edition (https://www.amazon.com/Measuring-Colour-R-W-Hunt/dp/1119975379/ref=sr_1_1?ie=UTF8&qid=1529914888&sr=8-1&keywords=measuring+color) (because of Alexey's recommendation earlier in the thread) last Saturday.
I've found that reading multiple books that converge around the same subject helps. Convergence is sublime.]
Wayne
-
I've found that reading multiple books that converge around the same subject helps. Convergence is sublime.]
Wayne
Absolutely! I found that approach much faster overall as far back as when learning calculus and linear systems. One learns by making connections and sometimes a text will poorly communicate some ideas where another will do an excellent job. Each has their strengths and weaknesses. So I've used the same technique ever since. You wouldn't believe the number of books I have on error correcting codes, programming, signal processing, and general tech stuff.
-
But a great way to get a historical perspective from the end of the film era. First color science book I bought, in 1990, after Kodak paid a visit to my lab and asked if we were interested in some joint work on color file format standards. It felt kind of outdated (and wandering) even then. Measuring Colour was the Jack Webb, "just the facts, ma'am" alternative (though, as Alexey points out, the thrusts of the two books are different), and I found that a few weeks later.
Jim
Nice to hear from someone in the color film business giving some background on the progress of photographic color reproduction.
Jim, do you see a huge improvement in the digital world? I do. And I don't need all this splitting hair measuring and use of obscure definers to determine color is precisely being reproduced digitally?
At least one thing that helped me in this thread is Wayne's posting of the French's mustard bottle images going from greenish to reddish as it related to the measured CRI numbers using a spectro. It at least showed a connection that the measuring device was actually showing the effects of the spikes in various LED spectra that's difficult to see perceptually.
And we don't judge the accuracy of color reproduction of images by comparing the original side by side with the reproduction like checking Delta E/Lab numbers A/B'ing individual color patches. So I don't see the usefulness of checking for precise by the numbers comparisons this way. It's a waste of time.
I'm surrounded by a community of fine art painters who have outside vendors provide reproduction prints of their paintings from artist provided digital captures, some with cellphones, others with P&S and DSLR's. None of these artists complained about color matching to the original.
-
And we don't judge the accuracy of color reproduction of images by comparing the original side by side with the reproduction like checking Delta E/Lab numbers A/B'ing individual color patches. So I don't see the usefulness of checking for precise by the numbers comparisons this way. It's a waste of time.
I suspect most people who don't understand basic colorimetry or have the tools to evaluate it, nor measure an illuminant and analyze it (the discussion here) would feel the same way.
-
Jim, do you see a huge improvement in the digital world? I do. And I don't need all this splitting hair measuring and use of obscure definers to determine color is precisely being reproduced digitally?
I'm not knocking the talents of the brilliant chemical and industrial engineers at Kodak, Agfa, Fuji, and many other companies. It's just that it is really hard to produce a film/processing -- in the case of 'chromes -- and film/processing/paper/processing -- in the case of color negative films and papers -- system that produces accurate color. At SPIE meetings Ed Giorgianni, a Kodak engineer of some repute, used to give seminars on the details of the color processing of film-based systems. I would walk out shaking my head, amazed that film did as well as it did, considering all the really tough chemical problems involved.
I once did an experiment where I took 25 different 'chrome emulsions (this was in the 90s, when there were lots to choose from), shot a Macbeth chart, and read all the resultant patches with a spectrophotometer. The average CIEL*a*b* delta-Es were in double digits. By the way, the great thing about doing that experiment was smelling film canisters right after I opened them. I loved that smell, and miss it in the digital age.
The engineers working on film systems weren't dummies. They knew they couldn't make it accurate. There was a body of opinion that stated that you didn't want it accurate even if you could do it. So they came up with many clever and artful inaccuracies that we call the film look. But it's not one look, it's a bunch of looks, and none of them are very accurate. Now we have digital systems, and we have less constrictions on what we can do. It's still impossible to build practical systems for color photography with perfect accuracy, but what we have now is much better than with film.
Jim
-
I once did an experiment where I took 25 different 'chrome emulsions (this was in the 90s, when there were lots to choose from), shot a Macbeth chart, and read all the resultant patches with a spectrophotometer. The average CIEL*a*b* delta-Es were in double digits. By the way, the great thing about doing that experiment was smelling film canisters right after I opened them. I loved that smell, and miss it in the digital age.
I share your love for the smell of opening a fresh film canister, but back in the '80's I had to often hold my breath to avoid the musty old chemical smell cracking open the film advance compartment in my used Yashica SLR I bought at a pawn shop in Austin, Tx. There was some weird kind of chemical/mixed metal corrosion going on that had the foam seal crumbling in my hands every time I dropped in a role of film.
I do think digital might be too good at color reproduction that raise unrealistic expectations for reproduction on inkjet prints for some "Southwest Art" style artists who want to reproduce the brilliant oranges, yellows and tans in their acrylic paintings viewed under halogen lights. It's sort of a fluoresce effect that I had to assume some inkjet printed inks can't reproduce.
A local art gallery curator emailed me back in 2006 to help him get a print match from the scans of these Southwest Art paintings and I didn't know enough about how the quality of light one views paintings under has limitations when digitally captured. The scans were just butt ugly and so were the prints compared to how the original looked under the halogen. The curator resorted to using a DSLR but he lit the painting with a daylight balance CFL which still couldn't reproduce the eye popping vibrance of the originals provided by the halogen lights.
I've posted this example before in other discussions on Raw capture with DSLR's but it explains the fluoresce effect I'm talking about I had to edit in order to emulate the effect.
Thinking back to the art gallery curator I wonder if he could've achieved the same effect shooting the paintings under the halogen instead of the daylight CFL. Of course Delta E number measuring wouldn't have helped because the edits were all based on perception, not exactly measured spectra
-
I do think digital might be too good at color reproduction that raise unrealistic expectations for reproduction on inkjet prints for some "Southwest Art" style artists who want to reproduce the brilliant oranges, yellows and tans in their acrylic paintings viewed under halogen lights. It's sort of a fluoresce effect that I had to assume some inkjet printed inks can't reproduce.
With proper colorimetry and deltaE reporting, no need to assume; provide some measurement data of the acrylic painting, in areas from these English words to describe a sensation (brilliant oranges, yellows and tans), a good ICC profile from any ink jet printer, some here can tell you with absolute certainty if they can or cannot be reproduced on that printer! :-*
-
Yes, digital is better. In 80s it was absolutely impossible to reproduce a painting accurately enough, a print compared to the original was visibly very different, even with no limits to budget (like when working for the Kremlin museums). But the criteria in museums were also relaxed. Now to get the work accepted I need to pass "deltaE" test. The bar is now much higher, and "just digital capture" doesn't cut it.
-
Yes, digital is better. In 80s it was absolutely impossible to reproduce a painting accurately enough, a print compared to the original was visibly very different, even with no limits to budget (like when working for the Kremlin museums). But the criteria in museums were also relaxed. Now to get the work accepted I need to pass "deltaE" test. The bar is now much higher, and "just digital capture" doesn't cut it.
Iliah, do you ever get comments or criticisms about the print reproduction of paintings not looking eye popping vibrant when viewed under the same light as the original that makes the original look this way?
The Soraa 5000K Vivid spot LED I lit the Frenches Mustard bottle and shot Raw with my DSLR I can't visually duplicate accurately with edits in ACR even using a daylight DNG profile. An Xrite CC chart is no problem.
I probably should try to make a custom camera profile and see if that makes a difference. The Frenches Mustard bottle looks very vivid under the Soraa light and no ACR saturation slider of any kind can make it match. Of course it also could be my sRGB gamut display that's the limiter.
-
I probably should try to make a custom camera profile and see if that makes a difference.
Indeed!
-
The Frenches Mustard bottle looks very vivid under the Soraa light and no ACR saturation slider of any kind can make it match. Of course it also could be my sRGB gamut display that's the limiter.
Yes, many saturated yellows even in well known studio scenes are typically way out of gamut (https://www.dpreview.com/forums/thread/4247965) (e.g. out of Adobe RGB).
Jack
-
Hi Jim,
What I might conclude from Jack's and your work is that there may be a lot of look to raw processing profiles. It may seem that a simple compromise matrix conversion may produce quite accurate colour, but it may be that colour profiles shift colours a bit for pleasantness giving up on accuracy?
What is your take?
Best regards
Erik
I'm not knocking the talents of the brilliant chemical and industrial engineers at Kodak, Agfa, Fuji, and many other companies. It's just that it is really hard to produce a film/processing -- in the case of 'chromes -- and film/processing/paper/processing -- in the case of color negative films and papers -- system that produces accurate color. At SPIE meetings Ed Giorgianni, a Kodak engineer of some repute, used to give seminars on the details of the color processing of film-based systems. I would walk out shaking my head, amazed that film did as well as it did, considering all the really tough chemical problems involved.
I once did an experiment where I took 25 different 'chrome emulsions (this was in the 90s, when there were lots to choose from), shot a Macbeth chart, and read all the resultant patches with a spectrophotometer. The average CIEL*a*b* delta-Es were in double digits. By the way, the great thing about doing that experiment was smelling film canisters right after I opened them. I loved that smell, and miss it in the digital age.
The engineers working on film systems weren't dummies. They knew they couldn't make it accurate. There was a body of opinion that stated that you didn't want it accurate even if you could do it. So they came up with many clever and artful inaccuracies that we call the film look. But it's not one look, it's a bunch of looks, and none of them are very accurate. Now we have digital systems, and we have less constrictions on what we can do. It's still impossible to build practical systems for color photography with perfect accuracy, but what we have now is much better than with film.
Jim
-
Hi Jim,
What I might conclude from Jack's and your work is that there may be a lot of look to raw processing profiles. It may seem that a simple compromise matrix conversion may produce quite accurate colour, but it may be that colour profiles shift colours a bit for pleasantness giving up on accuracy?
What is your take?
Kodak in the old days when they did produce DSLR cameras and PhotoDesk, solved it in the similar way to RPP - raw conversion profile provides accurate colour and on top of that a "look" profile can be applied when generating the output conversion (applied as one of the last conversion steps if I recall correctly). That separates pleasant and inaccurate looks from profiling the camera and supposedly make it easier to get more consistent results.
-
Hi Jim,
What I might conclude from Jack's and your work is that there may be a lot of look to raw processing profiles. It may seem that a simple compromise matrix conversion may produce quite accurate colour, but it may be that colour profiles shift colours a bit for pleasantness giving up on accuracy?
What is your take?
I'll speak to the Adobe camera profiles, since I have more experience with them than with, say, C1. They tend to introduce nonlinearities in the tone curve that impact accuracy but increase midtone contrast and flatten near the white and black points. That adversely affects color accuracy and is undoubtedly deliberate. In addition, Caucasian flesh tones are biased away from green. I'm pretty sure that is deliberate, too.
A year or so ago, Jack and I did a lot of work over on DPR on compromise matrix selection and its effect on accuracy with various cameras and found no camera that could produce 1 delta-E errors across the whole Macbeth chart (which is a very small sample space). So I don't think that you can get great accuracy with commercial cameras and compromise matrices. With LUTs, it should be possible to nail all 24 Macbeth samples, but at the expense of colors not in the training set.
Jim
-
> do you ever get comments or criticisms about the print reproduction of paintings not looking eye popping vibrant when viewed under the same light as the original that makes the original look this way?
The lights in my case are mostly of SoLux type or natural daylight, so no such problem. Comparing the look of the original under different lights it is easy to see it changes. Monet's painting of Houses of Parliament at Sunset looks rather acidy under FL lights.
> Soraa 5000K Vivid spot LED
In no particular order:
- "bumpy" light spectrums tend to oversaturate certain colours;
- colour transforms, especially LUT-based, tend to reproduce targets that were used to create those transforms in the first place better than arbitrary colours, particularly the colours that are out of the target gamut (another reason to switch to spectral characterization and profiling);
- sRGB may be limiting.
-
Yes, many saturated yellows even in well known studio scenes are typically way out of gamut (https://www.dpreview.com/forums/thread/4247965) (e.g. out of Adobe RGB).
Jack
That's a surprise because I always thought yellow was more hobbled by the display gamut than its encoded color space but then for me it's quite often difficult to determine perceptually until I hit a brick wall getting the yellows to look correct either from my memory or comparing to a lit target.
The custom DNG Soraa profile didn't make much of a dent in making the yellow look bright, vibrant and in the correct hue compared to shooting the French's Mustard bottle outdoors in direct sunlight which does render the correct hue but not in vibrance or luminance as rendered by the Soraa which of course is an artificial light. To get as close as possible to the actual Soraa lit target scene shown in the example below I just used the dual illuminant made using daylight lit and tungsten lit CCchart. It corrected the appearance of the blue on the mustard bottle.
The sun's rendering should be the standard for accurate color appearance but if an artificial light as the Soraa LED plays perceptual tricks employing colorant and filtering mechanisms that enhance what the sun can't, I don't see much use of a custom profile for that particular light.
The yellow that is off is the ink printed on top of the yellow plastic bottle. It's not commercial offset CMYK ink (no halftone dots) and appears to be a custom mix similar to Pantone and probably a proprietary formulation. None of my edits in the Raw capture could duplicate the vibrance, brightness and hue.
To show where this yellow resides on the luminance scale which has a major influence on color gamut size the point curve inset shows I'ld have to clip the heck out of the yellow and still not make it look correct.
-
That's a surprise because I always thought yellow was more hobbled by the display gamut than its encoded color space but then for me it's quite often difficult to determine perceptually until I hit a brick wall getting the yellows to look correct either from my memory or comparing to a lit target.
W??T??F? What encoded color space??? :o
Well at least using proper tools, I see my display color gamut exceeds an sRGB encoding (of anything) in yellows. And more. Encode in ProPhoto RGB, something that exceeds Adobe RGB (1998), of course it is 'hobbled' (it's called clipped) of my display color gamut.
Wireframe is sRGB. Encode any image in sRGB, you can't go (color) outside the lines, just like when you play with crayons ;D
Solid and colored is my display profile color gamut. Larger in yellow! Sure, green is massively larger but none the less, yellow is too.
It does help to actually have tools to see this stuff instead of guessing.
-
Let me reiterate for those that focus on the wrong point I'm making.
I'll put it simply that the printed yellow on the French's Mustard bottle lit by the Soraa LED spot light is glowing bright in luminance (maintaining hue) compared to the yellow plastic surround of the bottle much like the fluoresce effect of the green leaves I demonstrated several posts back.
Anyone who wants to split hairs explaining this in an argumentative manner using obscure color science jargon that no one can connect to a real and visual practical example as I demonstrated using edits of actual images, I consider not helpful.
-
Let me reiterate for those that focus on the wrong point I'm making.
We can only focus on your actual writings.... By all means reiterate and clarify. You haven't yet. You actually typed text combing one color space (I always thought yellow was more hobbled by the display gamut) with something that's the same (than its encoded color space) which makes no sense. Do you understand what encoding is? Are you talking about two color spaces, the display and? ??? If so, what's the 2nd color space that's 'hobbled' (a word that has no meaning in the discussion of color)?
hobbled; past participle: hobbled
1.
walk in an awkward way, typically because of pain from an injury.
"he was hobbling around on crutches"
synonyms: limp, walk with difficulty, walk lamely, move unsteadily, walk haltingly; More
2.
tie or strap together (the legs of a horse or other animal) to prevent it from straying.
Anyone who wants to split hairs explaining this in an argumentative manner using obscure color science jargon that no one can connect to a real and visual practical example as I demonstrated using edits of actual images, I consider not helpful.
We can only take your text as written, no argumentative manner (confusing by your own use of non color science jargon language). "No one can connect" only means, unless everyone posts, that you cannot connect. Sorry. If you want your readers to take your questions/writings seriously, ask them/comment without such ambiguities and language that has so little to do with this topic of color.
Meanwhile, the facts about a display color gamut, the color yellow, clipping compared to sRGB has been show visually, without any jargon.
-
Let me reiterate for those that focus on the wrong point I'm making.
Yes, I agree about that. ;D
-
The yellow that is off is the ink printed on top of the yellow plastic bottle. It's not commercial offset CMYK ink (no halftone dots) and appears to be a custom mix similar to Pantone and probably a proprietary formulation. None of my edits in the Raw capture could duplicate the vibrance, brightness and hue.
And yet, when I measure the French's Classic Yellow mustard bottle (plastic, of course no halftone dot), I get these Lab values from an i1Pro using M0:
Lstar 81.09 astar 6.38 bstar 67.57
And now let's plot that lab value against sRGB in 2D (all we need) and we can see, it falls easily within sRGB. Examine one yellow 'dot' within the sRGB gamut triangle. Again, no assumptions, actual colorimetry at play.
Further, I don't find this mustard to be very tasty but that's a subjective analysis. The colorimetry however, isn't!!!
Once again we see posting that illustrates that if you've only imagined it, you haven't experienced it. Now that I've experienced the mustard and it's actual measurement, we can move on to fact based discussions of color. ::)
-
Horrors: the actual label, which under a loupe does show a dot pattern, is just outside sRGB color gamut. Which is why I don't suggest ENCODING raw captures into sRGB!
OK, done with Mustard and issues of color gamut of French's for those who insist on using sRGB.
Lab values of label is 85.07/14.31/96.50
"I often say that when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind".
-Lord Kelvin
To show where this yellow resides on the luminance scale which has a major influence on color gamut size the point curve inset shows I'ld have to clip the heck out of the yellow and still not make it look correct.
Only if you keep assuming it does.
-
Horrors: the actual label, which under a loupe does show a dot pattern, is just outside sRGB color gamut. Which is why I don't suggest ENCODING raw captures into sRGB!
OK, done with Mustard and issues of color gamut of French's for those who insist on using sRGB.
Lab values of label is 85.07/14.31/96.50
"I often say that when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind".
-Lord Kelvin
Only if you keep assuming it does.
The measured color would apply to scene referred images. Of course output referred images will often be more saturated hence the need for wider gamuts than sRGB.
-
The measured color would apply to scene referred images. Of course output referred images will often be more saturated hence the need for wider gamuts than sRGB.
Which again is why I don't suggest ENCODING raw captures into sRGB!
A few folks posting here will recall this discussion:
https://www.dpreview.com/forums/post/59559609
-
Which again is why I don't suggest ENCODING raw captures into sRGB!
A few folks posting here will recall this discussion:
https://www.dpreview.com/forums/post/59559609
:)
-
Which again is why I don't suggest ENCODING raw captures into sRGB!
A few folks posting here will recall this discussion:
https://www.dpreview.com/forums/post/59559609
Oh yes! The unknown member!
-
Oh yes! The unknown member!
You blind too? Do you know how to read a post there, examine the sig? Seems not. Thanks for the thread hijack. Another person who hasn't the colorimetric tools or understanding to use em! :o \
Figure out yet how to read an Epson Nozzle check? (http://forum.luminous-landscape.com/index.php?topic=124550.msg1042484#msg1042484) ;)
-
You blind too? Do you know how to read a post there, examine the sig? Seems not.
The joke was apparently lost on you.
-
Horrors: the actual label, which under a loupe does show a dot pattern, is just outside sRGB color gamut. Which is why I don't suggest ENCODING raw captures into sRGB!
You've got the wrong French's Mustard Bottle (mine does not show a dot pattern in the printed yellow field: see image below shot through a loupe).
Also you need to light it with the Soraa to see that the light and ink combo is beyond reproduction through normal digital capture and processing even when applying edits.
Stay focused on my points and you won't have to do so much work.
Just to make it crystal clear since I didn't mention it earlier, I process/edit Raw captures in 16 bit ProPhotoRGB.
-
You've got the wrong French's Mustard Bottle (mine does not show a dot pattern in the printed yellow field: see image below shot through a loupe).
Nonsense, now you're going to tell us that the plastic and the label (which does have a dot pattern) vary based on the bottle size? What Lab value measurements you get from YOUR bottle? Oh, you can't conduct such measurements! :o
Also you need to light it with the Soraa to see that the light and ink combo is beyond reproduction through normal digital capture and processing even when applying edits.
No, I don't. But it's possible to comment on your inability to edit your raws! No matter the working space.
Stay focused on my points and you won't have to do so much work.
Dismissing (http://forum.luminous-landscape.com/index.php?topic=123799.msg1052286#msg1052286) your admitted fiction (and the admitted wrong points your making) on the topic of color isn't at all work and rather fun!
Rule on holes Tim, when you're deep in one, stop digging.
-
Compare apple to apples.
I'm not the one who finds it necessary to argue something I see with my eyes which is what most folks use to check a color match from the original to the final print viewed under lighting.
I'm not even seeing a match on my display with the Soraa light.
And for the love of god, Andrew, could you please learn some decent processing skills? That French's bottle is off in all colors. Why so much red in the plastic bottle. That's not even a match on a display.
-
> you need to light it with the Soraa to see that the light and ink combo is beyond reproduction through normal digital capture and processing even when applying edits
I'm not sure what is it that are you suggesting, is it that Soraa 5000K Vivid spot LED are not suitable for photographic reproduction purposes?
-
Compare apple to apples.
I'm not the one who finds it necessary to argue something I see with my eyes which is what most folks use to check a color match from the original to the final print viewed under lighting.
I'm not even seeing a match on my display with the Soraa light.
And for the love of god, Andrew, could you please learn some decent processing skills? That French's bottle is off in all colors. Why so much red in the plastic bottle. That's not even a match on a display.
I haven’t processed anything here Tim, I did what you cannot; measured the bottle and label to prove one falls completely into sRGB and the other just outside it. As to what you can and cannot match on your display, I suspect you may be struggling with display calibration as much as you struggle with editing your raws.
It's your burden if you insist on using an illuminant that kind of sucks, that you cannot colorimetrically evaluate (which IS the discussion here for those that can), that you assume a color falls outside a color gamut you cannot analyze, that you think the display color gamut is an encoding and all the rest of the silly text you've provided here that have been shown to be wrong or at best, massive assumptions on your part. Must you dug that hole deeper? Do you really think other than maybe one guy who just hijacked this thread yesterday, equally unable to measure color, anyone is taking you seriously? Please, stop flogging yourself in public. You're hobbling along here. ;D
-
> you need to light it with the Soraa to see that the light and ink combo is beyond reproduction through normal digital capture and processing even when applying edits
I'm not sure what is it that are you suggesting, is it that Soraa 5000K Vivid spot LED are not suitable for photographic reproduction purposes?
What ever he's suggesting, it's massively based on assumptions. I'm not sure he knows what he's suggesting either.
He's assumed the shot of my mustered bottle is processed when it's a JPEG straight from an iPhone simply to show the bottle next to a device he doesn’t own or know to use, that I used to measure the plastic and label.
-
> you need to light it with the Soraa to see that the light and ink combo is beyond reproduction through normal digital capture and processing even when applying edits
I'm not sure what is it that are you suggesting, is it that Soraa 5000K Vivid spot LED are not suitable for photographic reproduction purposes?
Not with all colors if one wants an exact match especially with paintings in combination with the lighting viewed under. I've had the same issues with the 6 pigment phosphor technology of the 5000K 95CRI Ottlight 100wt HD CFL but at least with that I could generate a color match on my display with common objects like an old deck of cards, OB paper but I didn't have the French's Mustard bottle at the time.
And that's really been my main point which is there needs to be a broader choice of color targets when using any brand of lighting with fine art painting repro both artificial and natural. Just because someone comes out with a brand new wizbang full spectrum artificial white light and it measures perfectly with some spectro it doesn't mean it's going to reproduce exactly all possible paint formulations used in paintings.
It might even look great as a fine art painting viewing light but not necessarily be ideal for capturing and processing digitally what we see in combination with the light and the painting.
-
Need to make a correction on the Ottlight. I meant to say that I didn't get a match to what it was suppose to look like under normal white light because the Ottlight made warm colors appear more saturated but at least I got a match on my display of this saturated color.
I ran into a brick wall with the Soraa and French's Mustard bottle. First time I've ever seen this.
Here's the image I've posted before of the Ottlight compared to the Solux 4800K. The Ottlight makes the deck of cards too clean and new looking. The Solux reproduces it exactly like it appears to me.
-
Danger Will Robinson, someone here is Lost in color space. ;D
-
Not with all colors if one wants an exact match especially with paintings in combination with the lighting viewed under. I've had the same issues with the 6 pigment phosphor technology of the 5000K 95CRI Ottlight 100wt HD CFL but at least with that I could generate a color match on my display with common objects like an old deck of cards, OB paper but I didn't have the French's Mustard bottle at the time.
And that's really been my main point which is there needs to be a broader choice of color targets when using any brand of lighting with fine art painting repro both artificial and natural. Just because someone comes out with a brand new wizbang full spectrum artificial white light and it measures perfectly with some spectro it doesn't mean it's going to reproduce exactly all possible paint formulations used in paintings.
It might even look great as a fine art painting viewing light but not necessarily be ideal for capturing and processing digitally what we see in combination with the light and the painting.
Read, learn:
https://blog.kasson.com/the-last-word/metameric-failure/ (https://blog.kasson.com/the-last-word/metameric-failure/)
http://www.colorwiki.com/wiki/Metamerism_Article
http://www.johnpaulcaponigro.com/blog/234/metamerism-metameric-failure/ (http://www.johnpaulcaponigro.com/blog/234/metamerism-metameric-failure/)
https://www.homecinemaguru.com/metameric-failure-in-display-color-calibration/ (https://www.homecinemaguru.com/metameric-failure-in-display-color-calibration/)
-
https://www.homecinemaguru.com/metameric-failure-in-display-color-calibration/
Interesting ..
“ Forbidden - Visitors from countries outside of North America are not permitted to browse this site because of changes in European privacy laws.”
-
Interesting ..
“ Forbidden - Visitors from countries outside of North America are not permitted to browse this site because of changes in European privacy laws.”
Weird! Pretty good display centric piece on metameric error but plenty of good in the others.
-
Interesting ..
“ Forbidden - Visitors from countries outside of North America are not permitted to browse this site because of changes in European privacy laws.”
Wonder if that explains DigitalDog.net script among other tracker indicators showing up in the lower left corner of Firefox browser. I'm getting tired of all these scripts slowing my downloading something as simple as a web page with the indicator "waiting to connect"...xxx whatever.
-
Wonder if that explains DigitalDog.net script among other tracker indicators showing up in the lower left corner of Firefox browser. I'm getting tired of all these scripts slowing my downloading something as simple as a web page with the indicator "waiting to connect"...xxx whatever.
More assumptions, this time outside the realm of color. Hole 3 feet deeper.
Got Safari bud? Runs fine, gets to that site. But start with Jim's article although JP's is probably maybe aimed at your level of understanding. :P
-
I'm not assuming. I can see your script with your name on it, Andrew.
LuLa loads quite quickly for me. I just don't like seeing scripts in general because some do hang for awhile waiting to connect. Google now does this which is my home page and now I've switched it to "The Fastest Loading Page On The Internet". https://varvy.com/pagespeed/wicked-fast.html
I'm going to take advice from the IT specialist on that site before I take advice from a Color Management Consultant.
-
I'm not assuming. I can see your script with your name on it, Andrew.
I'm going to take advice from the IT specialist on that site before I take advice from a Color Management Consultant.
It isn't my script. Just like I didn't process any image here, of mustard bottles or anything else.
You don't take advise from color management experts so I can't fathom you'd find then take advise from an IT specialist either.
All I did was copy and paste a URL from a site I can navigate to, as the screen capture proves. Don't go there, it is likely over your head anyway. Start with JP's piece, then Jims, see if can connect the dots such you begin to understand issues with color matching and illuminants. Or don't.
-
As for the light sources - I am about to submit all of that (having tested all the LEDs to death now and finishing Xenon pulsed light source testing) but it will take me a few days to finish write ups. Hopefully will be on GitHub by the end of upcoming week.
All submitted to GitHub (https://github.com/Alexey-Danilchenko/Spectron) now - the hardware (and mostly firmware) part of the project is complete now.
But soldering surface mount components is difficult. ... Ditto with fabricating COB mounting plates and heatsinks.
LED light source as well as some details of constructing LED COB head are on this page (https://github.com/Alexey-Danilchenko/Spectron/tree/master/hardware/Light_Source/LEDs) if you are interested
-
Horrors: the actual label, which under a loupe does show a dot pattern, is just outside sRGB color gamut. Which is why I don't suggest ENCODING raw captures into sRGB!
OK, done with Mustard and issues of color gamut of French's for those who insist on using sRGB.
Lab values of label is 85.07/14.31/96.50
Interesting. That Lab value is also outside of AdobeRGB from illuminant A to D65 with 2 deg CIE XYZ observers. What observer did you use in your conversion?
-
All submitted to GitHub (https://github.com/Alexey-Danilchenko/Spectron) now - the hardware (and mostly firmware) part of the project is complete now.
LED light source as well as some details of constructing LED COB head are on this page (https://github.com/Alexey-Danilchenko/Spectron/tree/master/hardware/Light_Source/LEDs) if you are interested
Excellent Alexey. How about adding some images of the light as set up for its intended purpose?
-
Excellent Alexey. How about adding some images of the light as set up for its intended purpose?
Sorry not quite following the question. The light head mounted to face monochromator entry slit at a distance that keeps light in a small condensed circle. Nothing special around that especially since monochromatots do differ. I'll eventually post photos of all my setup and some details about construction but not yet.
-
> How about adding some images of the light as set up for its intended purpose?
But it's hard to see, Alexey feeds light directly into monochromator, I use a fiber, everything is pretty standard.
-
> How about adding some images of the light as set up for its intended purpose?
But it's hard to see, Alexey feeds light directly into monochromator, I use a fiber, everything is pretty standard.
To add to to Iliah's comment, my monochromator came without entry slit but with entry attachment for optical cable. I use one of the Luxtron/Lumasense/Xinix monochromators found on eBay in large quantities - in those industrial Luxtron boxes (if disassembled) are usually packaged Mini Chrom compact series monochromators for visible range (considering the target usage of Luxtron devices) but with entry slits fed from remote sensing chamber via optical cable. To make it more compact and avoid any light losses I always attempted to feed the light directly into monochromator (initial version with tungsten halogen light was using optical cable with collimating lens at the end - I posted pictures of that setup here in dcamprof thread). After lots of experimentation with achieving good focused light on the monochromator output in the integrating sphere, I made up my own entry - essentially optical cable entry that came with it fused with single element collimating lens of the small focal distance. The position was chosen such that it focuses the light falling on the lens into integrating sphere producing quite clear projection of output slit (I think it's the best so far that I could do with DIY approach without any special tools).
From there on the rest is simple - light source directed pretty much on entrance (the collimating lens) and quite close to it. The diameter of this lens is about 4mm and it is fused into plastic entrance block by Sugru and sealed, no photo will show it clearly enough unfortunately (tried that before).
-
To add to to Iliah's comment, my monochromator came without entry slit but with entry attachment for optical cable. I use one of the Luxtron/Lumasense/Xinix monochromators found on eBay in large quantities
One like this? (https://www.ebay.com/itm/LUXTRON-MODEL-0201-MONOCHROMATOR-w-Fiber-Optic-Cable-5001-0201-01-00/283020440523?hash=item41e55533cb:g:A~UAAOSwuAVWzVnv)
All of this has been, um, educational. I've got a small stack of textbooks I am working through (Hunt's Measuring Color and Reproducing Color, several "Introduction to Matlab" books (and learning enough math to do at least some of the math in the above textbooks), several Particle Photons to wake up, following Iliah's suggestions on working with SMT components (http://forum.luminous-landscape.com/index.php?topic=123799.msg1051440#msg1051440), etc.
This thread has been very rich for me. I'm just poking my head up to thank everybody for all the valuable information. Now back to reading.
-
One like this? (https://www.ebay.com/itm/LUXTRON-MODEL-0201-MONOCHROMATOR-w-Fiber-Optic-Cable-5001-0201-01-00/283020440523?hash=item41e55533cb:g:A~UAAOSwuAVWzVnv)
Yes. Inside this one is one of these https://www.dynasil.com/product-category/mini-chrom-monochromators/digital-mini-chrom-monochromator/ together with photomultiplier and supporting circuitry. Those industrial second hand blocks are great way of getting monochromator for a fraction of its original price.
My one comes from Luxtron/Lumasense 1108 (https://www.ebay.co.uk/itm/Luxtron-LumaSense-Model-1108/170625275474)
-
Sorry not quite following the question. The light head mounted to face monochromator entry slit at a distance that keeps light in a small condensed circle. Nothing special around that especially since monochromatots do differ. I'll eventually post photos of all my setup and some details about construction but not yet.
Right. Share a picture so we can see what you are doing and how you are doing it :)
Great stuff Alexey and Iliah!
Jack
-
Right. Share a picture so we can see what you are doing and how you are doing it :)
When its ready perhaps, but not yet. Monochromator is a box with input and output so all photo will show you is a light next to a box. Similar to this for example:
(http://zhang-nano.gatech.edu/Replace%20sphere%20-%20serve%20as%20both%20monochromator%20and%20integrating%20sphere.jpg)
The generic article for Monochromator lights - this for example (https://www.newport.com/t/getting-light-into-a-monochromator) would perhaps be of some use.
-
> avoid any light losses
Exactly. It is may sound trivial: when a monochromator is used as a tunable filter in such an application, light losses are something to avoid, because with insufficient light the sensor will be characterized for untypical integration time. A procedure for calculation is presented here: https://www.newport.com/t/calculating-monochromator-throughput Because monochromator output is partially polarized, some further losses on sensor sandwich are possible.
-
Has anyone made a spectral data file of Imaging resources lightning that I can use in Lumariver Profile Desinger?