Problem reaching 4700K with Solux lamp [solved]

[torger]

Outdoor sunlight is varying heavily over the day, weather, location and time of year. Here up north we have months in a row when experiencing something similar to D50 is just a distant dream .

If your experiments needs a fixed reference for repeatability D50 is there, a standardized average type of daylight. Color science is built around standardized light sources, using simulators more often than real daylight. Without simulators it would be very difficult to make scientific work as it requires repeatability and agreed standards. In the textile and print industry using a fixed repeatable reference which clients agree on is also important, making simulators more suitable than real daylight.

A simulator must replicate the important properties of daylight. Halogen-based simulators lack the spikes you find in fluorescent simulators which can be important in some applications, camera profiling is one. In other applications UV-content is more important (making effect of OBA visible for example) and then fluorescent simulators can be better, although adding a specific UV light source beside your halogen simulator can be even better.

LED is the future in daylight simulators though, high end reference viewing stations from Just Normlicht for example has already LEDs, but they're still extremely expensive. A low cost alternative are UV-pumped white LEDs which look promising (Yuji seems to be the leading company on that tech), I'm not sure they can rival filtered halogens though.

According to data I've seen a voltage-tuned Solux 4700K seems to be performing just as good as a $5000 multichannel LED when it comes to simulating D50. Next step up is stacked daylight filtered halogens, which is $300, probably only mariginally better.

[digitaldog]

No disagreement about what you've written torger in terms of Solux bulbs. I'm suggesting that this bulb is the best man made product I've used and at an affordable price and the there isn't a problem with them whether they reach CTT 4700K or not as the topic title suggests. Again, what multiple people using multiple products and bulbs report in terms of CTT doesn't matter. The plots of it's SPD are far more useful IMHO. As to how well they simulate D50, they key is deciding what one means by simulate and by how far from the goal of an average of real daylight measurements and using a CTT value as reported doesn't tell us that.

[AlterEgo]

According to data I've seen a voltage-tuned Solux 4700K seems to be performing just as good as a $5000 multichannel LED when it comes to simulating D50.

but what about the very uneven illumination from Solux - it seems that if you are experimenting with it than you shall have a very hard time to fit the target within the area where you have consistent parameters...

[Tim Lookingbill]

In regards to hair splitting accuracy of simulated D50 artificial lighting has there been a study measuring the eye's adaptive nature compared to how a sprectro defines color differences in Delta E's?

Years back I found an Eizo pdf sales doc displaying a simple chart demonstrating the level of change between a Delta E 1 vs 2 vs 3 etc. showing the rate of change in what appeared to be a stock photo of Egypt's Sphinx. Delta E 3 I believe showed a marked change from 1 & 2.

Given that additional human influenced variable along with the rest Andrew and torger put forth I'm not seeing this level of accuracy helping photographers, but do kind of see it as necessary for programmers in helping them establish some type of standard for defining color mathematically under a simulated D50 artificial light for writing profiling software algorithms.

[Bart_van_der_Wolf]

In regards to hair splitting accuracy of simulated D50 artificial lighting has there been a study measuring the eye's adaptive nature compared to how a sprectro defines color differences in Delta E's?

I assume that the CMC has published some papers on the subject of Delta E. In principle a delta E of 1 is supposed to be just noticeable. There are several other measurements of Delta E, like Delta E (CIE 1976), or Delta E (CIE 1994), or Delta E (CIE 2000).

Given that additional human influenced variable along with the rest Andrew and torger put forth I'm not seeing this level of accuracy helping photographers, but do kind of see it as necessary for programmers in helping them establish some type of standard for defining color mathematically under a simulated D50 artificial light for writing profiling software algorithms.

Indeed, it is inevitable for programming and research. There are too many human observer differences to pick one and promote it as a standard.

Cheers,
Bart

[digitaldog]

A deltaE of less than 1 between two color values is supposed to be imperceptible but to complicate matters, as Bart points out, there are several formulas for calculating this metric. Further the ability of the eye to distinguish two colors as different and is more limited for yellows but is better for greens and blues.

[Tim Lookingbill]

I assume that the CMC has published some papers on the subject of Delta E. In principle a delta E of 1 is supposed to be just noticeable. There are several other measurements of Delta E, like Delta E (CIE 1976), or Delta E (CIE 1994), or Delta E (CIE 2000).

Indeed, it is inevitable for programming and research. There are too many human observer differences to pick one and promote it as a standard.

Cheers,
Bart

Knowledgeable finds, Bart, but I was hoping to seeing the rate of change in the form of a visual sample and then show the adaptive change most humans see out in nature.

I often test my own eye's adaptive nature looking at the warm color of asphalt pavement lit by the open sun and then switch to checking the hue of a pecan tree bark in shade which is close to neutral and then look up at a blue sky that has a noticeable magenta tinge and look back at the tree bark and notice it looks more green.

What Delta E is that? Landscape photographers might depend on it when trying to remember what they saw out in nature during post processing.

[digitaldog]

What Delta E is that?

dE is derived by measurements. Of two solid samples. So the question can't be answered because the process you 'tested' doesn't fall into colorimetry but rather your subjective perception. As such we could both conduct your test, even side by side at the same time and come up with quite different conclusions.

[Tim Lookingbill]

dE is derived by measurements. Of two solid samples. So the question can't be answered because the process you 'tested' doesn't fall into colorimetry but rather your subjective perception. As such we could both conduct your test, even side by side at the same time and come up with quite different conclusions.

It's not my own subjective perception I can assure you. It happens viewing prints under a static artificial print viewing light after staring for an extended period of time at an open window during the day and looking back at both my display and prints. I shut my eyes for a while and the hue of both print and display returns back to the way they were during editing. Scenes captured and viewed on my display induce similar adaptive effects especially with greens lit by sunset.

It's a natural optical effect every human being lives with. I should know. I used to paint photorealistic renderings on illustration board and canvas. I'ld spend entire summers just looking at how daylight renders colors and noticing the differences between the mixed lighting of direct sunlight vs outdoor shade. It's not my objective perception. It's a real human condition and many don't bother to notice this when shooting outdoors.

But all this is just as an example of the many variables involved with how to define color perception mathematically to some standard. Maybe everyone's rate of change influenced by this adaptive effect is different between each individual but it still remains a real phenomenon and not some subjective judgement call on color perception.

[digitaldog]

It's not my own subjective perception I can assure you.

Of course it is. That's how our color vision works Tim.

It's a natural optical effect every human being lives with.

Yes but that's not something we can measure and produce a dE report. How would someone in her 20's differ from a male in his 60's and someone who suffers some degree of color blindness or has cataracts or had a couple glasses of wine or a few bong loads perceive this?

[Tim Lookingbill]

Of course it is. That's how our color vision works Tim. Yes but that's not something we can measure and produce a dE report. How would someone in her 20's differ from a male in his 60's and someone who suffers some degree of color blindness or has cataracts or had a couple glasses of wine or a few bong loads perceive this?

My comment and point is in reference to torger's requirements for exact Kelvin measurements from a D50 simulated artificial light (Solux) in order to attain a more accurate to scene editing starting point derived from a camera profile he's writing software for whose color target is lit by said D50 simulator.

Considering editing involves human visual intervention that exhibits adaptive visual behavior as I've described, makes this "accurate" starting point even more subjective since we humans don't rely on Delta E numbers measured from a spectro that records color with a static non-adaptive behavior. IOW what's accurate for one viewer may not be so for others.

But what I described about how I my eyes adapt viewing a landscape is real for me and I've seen this with others especially those asking for a by the numbers system for judging proper saturation levels editing images since their eyes take longer to adapt to newer edits.

[digitaldog]

But what I described about how I my eyes adapt viewing a landscape is real for me and I've seen this with others especially those asking for a by the numbers system for judging proper saturation levels editing images since their eyes take longer to adapt to newer edits.

I never said it wasn't real. I said it can't be described using dE which you asked about. Because apparently you don't yet understand, despite my posts, what you describe has nothing to do with dE or colorimetry.

[Tim Lookingbill]

I never said it wasn't real. I said it can't be described using dE which you asked about. Because apparently you don't yet understand, despite my posts, what you describe has nothing to do with dE or colorimetry.

dE from what I've understood it to be is the measure of color difference from a known and measured target. The rate of difference as it appears to a human has not been clearly defined. It would be helpful if photographers knew or had some kind of idea whether differences in dE as measured from a spectro is something to be concerned on how much it affects their editing and if having a more "accurate" to scene starting point is beneficial.

It doesn't appear that it's all that important in measuring dE differences caused by Kelvin number variations measured from a Solux light. But also it has not been shown how it affects the results derived from software driven camera profiles whose algorithms have these Solux induced Kelvin variations encoded in the creation of the camera profile.

[digitaldog]

dE from what I've understood it to be is the measure of color difference from a known and measured target.

No, it's the measure of difference between two solid colors. Take a pile of two sets of such readings, you can produce an average dE report. But that doesn't change the fact that the metric isn't based on anything you've described you perceive you've seen!

It would be helpful if photographers knew or had some kind of idea whether differences in dE as measured from a spectro is something to be concerned on how much it affects their editing and if having a more "accurate" to scene starting point is beneficial.

If they are concerned between the differences of two solid color samples, measured with a fixed illuminant in a Spectrophotometer, that's doable and easy. Again, it has no relationship to what you've described:

I often test my own eye's adaptive nature looking at the warm color of asphalt pavement lit by the open sun and then switch to checking the hue of a pecan tree bark in shade which is close to neutral and then look up at a blue sky that has a noticeable magenta tinge and look back at the tree bark and notice it looks more green.
What Delta E is that?

You're making a statement of what you visually perceive and then ask what the differences are based on the differences of two solid colors. You can't get an answer in dE!
It is like saying you intend to drive from California to Arizona and then asking what is the weight of the two states you hope to visit? The two have nothing to do with each other and further, how do you expect someone to answer the question about the weight?

It doesn't appear that it's all that important in measuring dE differences caused by Kelvin number variations measured from a Solux light.

The K numbers express a range of colors and we're not measuring two samples so the statement you've made doesn't make sense. Sorry. Two different software products, hooked up to the same instrument measuring the same Solux could very likely produce two different CCT values. Now what?

[Tim Lookingbill]

Exactly.

[AlterEgo]

No, it's the measure of difference between two solid colors. Take a pile of two sets of such readings, you can produce an average dE report. But that doesn't change the fact that the metric isn't based on anything you've described you perceive you've seen!

but ultimately that is based on experiments on human sight... now the mere fact that those experients are far (timewise) removed from today and luckily (or not ?) were not conducted by Tim Lookingbill still does not change the fact about the origins of CIE color spaces, no ? 

[digitaldog]

but ultimately that is based on experiments on human sight... now the mere fact that those experients are far (timewise) removed from today and luckily (or not ?) were not conducted by Tim Lookingbill still does not change the fact about the origins of CIE color spaces, no ?  

Indeed, it IS based on experiments of human vision. OF TWO SOLID color samples in very ridged and controlled environments. What Tim is asking about his far, far more than two samples in such a condition, no?  

[AlterEgo]

Indeed, it IS based on experiments of human vision. OF TWO SOLID color samples.

popular sources saying a mixure of 3 illuminants

[digitaldog]

popular sources saying a mixure of 3 illuminants

Ah OK, but what's this have to do with Tim's request for a dE value based on: I test my own eye's adaptive nature looking at the warm color of asphalt pavement lit by the open sun and then switch to checking the hue of a pecan tree bark in shade. What Delta E is that?

[AlterEgo]

Ah OK, but what's this have to do with Tim's request for a dE value based on: I test my own eye's adaptive nature looking at the warm color of asphalt pavement lit by the open sun and then switch to checking the hue of a pecan tree bark in shade. What Delta E is that?

I can only say that it was a blessing that Tim was not part of those experiments otherwise we 'd be still living w/o dEs   or may  be we 'd have a better system - who knows