Luminous Landscape Forum
Raw & Post Processing, Printing => Colour Management => Topic started by: lightshiner on May 12, 2015, 01:31:38 pm
-
I've been experimenting with the Granger Chart sRGB IEC 61966-2.1 (https://luminous-landscape.com/how-to-create-color-test-charts/) and Adobe Lightroom's Soft Proofing feature. I'm seeing Gamut Warnings when specifying a sRGB soft proofing colour profile (the warning area is much reduced if I switch the colour profile to Adobe RGB),
Is it possible that the sRGB Granger Chart includes colours which are not actually part of the sRGB gamut?
-
The gamut warning isn't accurate, ignore it. Same should be true in Photoshop.
-
Thanks for letting me know that Lightroom gamut warnings may not be accurate.
Such inaccuracy could explain why soft proofing based on a specific printer profile (ICM provided by a professional printing service) indicates that approximately 1/3 of the colours in the (sRGB) Granger chart are outside of the printer's gamut.
-
Thanks for letting me know that Lightroom gamut warnings may not be accurate.
Such inaccuracy could explain why soft proofing based on a specific printer profile (ICM provided by a professional printing service) indicates that approximately 1/3 of the colours in the (sRGB) Granger chart are outside of the printer's gamut.
That an sRGB file soft proofed in sRGB should show out-of-gamut colors in Lr and Ps isn't good, but soft proofing is still useful.
That 1/3 of an sRGB image's colors would be out of a print machine's gamut is another matter. (What printer?). sRGB will have colors that are out of most printer's gamut, but the printer's gamut will also have large areas that are out of the sRGB gamut.
Brian A
-
Thanks Brian for your comments.
The printer profile is from a Fujitsu printer using Fujifilm professional paper. I used Lightroom to reduce saturation to the point that there were no remaining gamut warnings, however, the corrected Soft Proofing image was much less saturated than the actual prints (from two printers). While the prints were not identical. the match between them was much better than either print and the corrected Soft Proofing image.
I've attached a screenshot of the Destination Gamut Warnings using the aforementioned printer profile.
BTW, I posted a similar message on the Photoshop feedback site: http://feedback.photoshop.com/photoshop_family/topics/suspicious-results-using-lightroom-soft-proofing?rfm=1
David S
-
The gamut warning isn't accurate, ignore it. Same should be true in Photoshop.
I constructed Grainger rainbows in sRGB and ProPhotoRGB and soft proofed in PhotoshopCC with the device to simulate set as sRGB. I then turned on the gamut warning after having set the out of gamut warning to bright red to simulate the warning in Lightroom. The soft proof of the sRGB image (shown on the right of the screen capture) showed no out of gamut colors, in contrast to the situation with soft proofing in LR. The soft proof of the ProPhotoRGB image showed extensive out of gamut colors (left in the screen capture). The conclusion is that the gamut warning is more accurate with Photoshop than LR.
The problem with both out of gamut warnings is that an only slightly out of gamut color is not differentiated from a wildly out of gamut color. A color that is out of gamut by one ΔE would not be significant, but one that is out of gamut by 10 ΔEs would be highly significant. A pseudocolor representation of the amount by which a color is out of gamut would make the out of gamut warning more useful. Such a representation can be constructed with ColorthinkPro or Gamutvision. An example of such a Gamutvision gamut plot is shown below in the second image. The scale to the right shows the key for the gamut warning where yellow represents a ΔE of 1 or 2 (insignificant) and purple a ΔE of 7-8 (significant).
Cheers,
Bill
-
Thanks Bill ... very interesting.
I'm using the most recent version of Lightroom (LR6 released April 21st, 2015) so one would hope that the Soft Proofing would be equivalent to that of the current version of Photoshop.
I like your suggestion for determining the degree to which a colour is out of gamut. One simple technique would be to add a matching threshold to LR's Soft Proofing (or perhaps just restrict warnings to "significant", "perceptible", or "exact match" ) which would limit the reporting of out of gamut warnings.
- David S
-
Try this to get around the "anomolies" in Adobe softproofing (in Photoshop)
1. Dupe your larger colorspace image, then CONVERT the dupe to the smaller colorspace (Prophoto to sRGB).
2. Copy the dupe, paste it back onto the original. The paste will convert the smaller back to larger.
3. Put the new layer in Difference mode.
4. Add a layer to desaturate.
What you get is a grayscale view of out-of-gamut areas and a representation of how far things are out of gamut. Also, that grayscale can be used as a mask for adjustments if desired.
Doing this with Bill Atkinson's 28 balls is more dramatic than with a Granger Rainbow.
-
Try this to get around the "anomalies" in Adobe softproofing (in Photoshop)
1. Dupe your larger colorspace image, then CONVERT the dupe to the smaller colorspace (Prophoto to sRGB).
2. Copy the dupe, paste it back onto the original. The paste will convert the smaller back to larger.
3. Put the new layer in Difference mode.
4. Add a layer to desaturate.
What you get is a grayscale view of out-of-gamut areas and a representation of how far things are out of gamut. Also, that grayscale can be used as a mask for adjustments if desired.
Doing this with Bill Atkinson's 28 balls is more dramatic than with a Granger Rainbow.
Thanks Redcrown for the suggestion. Unfortunately, I don't have access to Photoshop but it is generally possible to change the colour space in LR while exporting (original image has an embedded sRGB colour profile). I could import an exported Pro Photo version of the (original sRGB) image and finally re-export using sRGB.
BTW, I like the idea of using the out of gamut warning regions as a mask to help bring such regions back into the available gamut (although different colours may possibly require different adjustments).
- David S
-
1. Dupe your larger colorspace image, then CONVERT the dupe to the smaller colorspace (Prophoto to sRGB).
Why sRGB of all color spaces?
-
Why sRGB of all color spaces?
Web publishing, on-line printing services, ....?
Cheers,
Bart
-
Web publishing, on-line printing services, ....?
So convert to sRGB and you're done. You are forced into a RelCol conversion, the colors are going to clip to sRGB as expected. Why the mask, editing the image with OOG overlay if you're going to clip to sRGB?
-
So convert to sRGB and you're done.
Not really, and you know it Andrew.
There may be issues that can be addressed while in the larger gamut space. Perhaps a targeted desaturation, or a hue shift, or a luminance adjustment will reduce the amount of clipping (and thus loss of distinction or posterization). Therefore a direct comparison, also of thee magnitude of clipping, and the ability to create a graduated mask to isolate the problem areas, is very useful. I also use the method that 'Redcrown' described, I have it in a Photoshop Action.
Cheers,
Bart
-
Not really, and you know it Andrew.
For Web publishing, on-line printing services, ....? Really?
-
Jeeze, Andrew.. I respect your knowledge and bow to your experience. But I wish you would stay on your drugs. Or go back to grade school and re-learn how to count to 10 before you pop off.
-
This is complicated by the conversion being relative
-
Why sRGB of all color spaces?
Because the OP was soft proofing with sRGB as the target. Redcrown's method could be used with a printer profile as the target, in which case the perceptual rendering intent could be used. He specifically refers to the smaller color space, and sRGB was only an example of a smaller space. I haven't yet tried the method, but it appears interesting.
ColorthinkPro is perhaps the gold standard for these gamut comparisons, but try feeding it a 100 MB 16 bit file. One has to drastically downsize such a file and then generate a color list--a very time consuming process. Gamut vision can be used directly with large files, but it is Windows only and hasn't been updated for a while and will not even install on my Windows 8.1 desktop.
Bill
-
Redcrown's method could be used with a printer profile as the target, in which case the perceptual rendering intent could be used. He specifically refers to the smaller color space, and sRGB was only an example of a smaller space. I haven't yet tried the method, but it appears interesting.
Pointless, we have soft proofing, we can and should edit via soft proofing.
The conversion made for brining the OOG comparison was made using, oh my, a profile! We can use it for soft proofing (since 1998) and for affecting OOG colors for the conversion, exactly what the technique does before all the futzing around manually. Doing this for sRGB to pop onto the web? Seriously?
-
Thanks for all the comments...
At the risk of unleashing a firestorm, I chose the sRGB since both my camera and my intended printer support this colour space. As was mentioned, probably both devices may support a gamut that is actually wider than sRGB (e.g. the camera also supports Adobe RGB). I thought that using sRGB would help ensure that image colours would be faithfully reproduced (especially by using Soft Proofing with the colour profile provided by the printing service).
I view/edit images using Lightroom 6 on a 32" LED monitor (actually a Samsung 720p TV connected via HDMI) that has been calibrated using the Datacolor Spyder4 Express device/utility software. According to this utility, the monitor covers 98% of the sRGB gamut. So from my naive colour management perspective, it seems like sRGB is a good choice given that the camera, printer, and monitor all cover a large portion of the SRGB colour space.
BTW, I recall reading that Lightroom itself uses Adobe RGB internally (except for the develop module where Pro Photo is used).
- David S
-
BTW, I recall reading that Lightroom itself uses Adobe RGB internally (except for the develop module where Pro Photo is used).
Internally it uses ProPhoto RGB primaries with a linear TRC. For previews outside Develop, it's Adobe RGB (1998) unless something has changed. But the processing is done with ProPhoto RGB gamut.
-
Pointless, we have soft proofing, we can and should edit via soft proofing.
The conversion made for brining the OOG comparison was made using, oh my, a profile! We can use it for soft proofing (since 1998) and for affecting OOG colors for the conversion, exactly what the technique does before all the futzing around manually. Doing this for sRGB to pop onto the web? Seriously?
Editing via soft proofing is fine, but problems can arise when one is editing colors that are out of the gamut of the monitor. Shown below is the gamut of my monitor (calibrated with Spectraview and the NEC colorimeter) and the gamut of my printer with Epson Premium Glossy paper as described by the Epson profile. Some yellows and teal-greens are out of the gamut of the monitor but within that of the printer. The monitor is advertised as wide gamut. I've read your previous posts where you advise that if one is editing the colors of the file and the changes are not shown on the monitor, the gamut of the monitor may have been exceeded.
Lightroom does try to help in this situation by also giving a gamut warning for the monitor. What do you think of that?
Regards,
Bill
-
Lightroom does try to help in this situation by also giving a gamut warning for the monitor. What do you think of that?
I think it's useful and could be far more useful.
-
Internally it uses ProPhoto RGB primaries with a linear TRC. For previews outside Develop, it's Adobe RGB (1998) unless something has changed. But the processing is done with ProPhoto RGB gamut.
Andrew,
Thank you for clarifying. Actually, this makes perfect sense since I understand that Pro Photo is the widest gamut (compared to sRGB and Adobe RGB). BTW, I had to look up TRC (Tone Response Curve).
I'd be curious to know if the wide gamut of the 4k Ultra HD specification will necessitate a revision to Pro Photo as the gold standard for photography.
Regarding my original query, I've sent a note to Adobe asking if someone could have a look at this apparent anomaly.
- David S
-
...
Lightroom does try to help in this situation by also giving a gamut warning for the monitor. What do you think of that?
Bill,
I've also noticed some unexpected behaviour with LR's Monitor Gamut Warning. I'd be curious to know if you're observing the same with your (advanced) setup.
Steps:
- select image (e.g. Granger chart). turn on Soft Proofing with Monitor and Destination Gamut Warnings both enabled
- select the monitor's colour profile as the Destination profile (use Other with Include Display Profiles checked)
Now since the same colour profile is used for both gamut warnings, colours outside of both gamuts should be highlighted in pink (i.e. red for destination, blue for monitor). However, the image below shows only Destination Gamut Warnings.
- David S
-
Bill,
I've also noticed some unexpected behaviour with LR's Monitor Gamut Warning. I'd be curious to know if you're observing the same with your (advanced) setup.
Steps:
- select image (e.g. Granger chart). turn on Soft Proofing with Monitor and Destination Gamut Warnings both enabled
- select the monitor's colour profile as the Destination profile (use Other with Include Display Profiles checked)
Now since the same colour profile is used for both gamut warnings, colours outside of both gamuts should be highlighted in pink (i.e. red for destination, blue for monitor). However, the image below shows only Destination Gamut Warnings.
- David S
David,
I performed your suggested test with my setup (which is no more advanced than yours, I suspect).
With the Grainger in sRGB and soft proofing with the monitor profile as the destination, no out of gamut is observed for the monitor, but the destination shows some out of gamut colors. Since my monitor is wide gamut (approximately AdobeRGB), there should be no out of gamut at all.
With the Grainger in ProPhotoRGB and soft proofing with the monitor profile as the destination, severe out of gamut is present in the destination, but the monitor shows no out of gamut, which is impossible.
I did check to make sure that the system was using the proper profile for my display, and it was as shown in the third image below. I conclude that the monitor gamut warning is seriously defective.
Cheers,
Bill
-
David,
I performed your suggested test with my setup (which is no more advanced than yours, I suspect).
With the Grainger in sRGB and soft proofing with the monitor profile as the destination, no out of gamut is observed for the monitor, but the destination shows some out of gamut colors. Since my monitor is wide gamut (approximately AdobeRGB), there should be no out of gamut at all.
With the Grainger in ProPhotoRGB and soft proofing with the monitor profile as the destination, severe out of gamut is present in the destination, but the monitor shows no out of gamut, which is impossible.
I did check to make sure that the system was using the proper profile for my display, and it was as shown in the third image below. I conclude that the monitor gamut warning is seriously defective.
Cheers,
Bill
Thanks Bill for confirming that a single colour profile produces vastly different results when specified as a Destination Profile or when specified as a Monitor Profile. For the time being, I think it's safer to ignore the gamut warnings than to modify (e.g. desaturate) the image based on these warnings.
Apart for Soft Proofing gamut warnings, has it been your experience that the Soft Proofing preview gives a reasonable rendition of the eventual print? If I compare the attached image to the corresponding print of the Granger Chart, Soft Proofing seems very approximate. Of course, this could reflect an inaccurate profile as opposed to a deficiency in Soft Proofing.
Regards,
David
-
For the time being, I think it's safer to ignore the gamut warnings than to modify (e.g. desaturate) the image based on these warnings.
Amen to that brother! Huge waste of time, spend it more productively elsewhere.
-
Thanks Bill for confirming that a single colour profile produces vastly different results when specified as a Destination Profile or when specified as a Monitor Profile. For the time being, I think it's safer to ignore the gamut warnings than to modify (e.g. desaturate) the image based on these warnings.
...
Regards,
David
An explanation of the Monitor Gamut Warnings behaviour ...
From Suspicious Results using Lightroom? (http://feedback.photoshop.com/photoshop_family/topics/suspicious-results-using-lightroom-soft-proofing?rfm=1)
LRuserXY
The monitor gamut warning is implemented _after_ the conversion to the target color space, so if both are equal, there will be no monitor warning regarless of the original saturations.
This was changed in some early beta (?) stage when the softproof was introduced, and I think there was an explanation by Eric Chan why this was changed in the Adobe forums. Unfortunetely, this was never explained in Adobes official Lightroom help/documentation.
Original forum thread (https://forums.adobe.com/thread/947403?start=40&tstart=0) (Jao vdLs post and Eric Chans answer "The current behavior is a bug and will be fixed").
It seems logical, because if I want to softproof using a printer profile, I am not interested in the areas that exceed the monitor gamut in the _original_ image. I am only interested in the areas that exceed the monitor gamut in the _converted_ image - because only these areas will _not_ look the same on the monitor vs. printed output.
-
An explanation of the Monitor Gamut Warnings behaviour ...
From Suspicious Results using Lightroom? (http://feedback.photoshop.com/photoshop_family/topics/suspicious-results-using-lightroom-soft-proofing?rfm=1)
LRuserXY
The monitor gamut warning is implemented _after_ the conversion to the target color space, so if both are equal, there will be no monitor warning regarless of the original saturations.
This was changed in some early beta (?) stage when the softproof was introduced, and I think there was an explanation by Eric Chan why this was changed in the Adobe forums. Unfortunetely, this was never explained in Adobes official Lightroom help/documentation.
Original forum thread (https://forums.adobe.com/thread/947403?start=40&tstart=0) (Jao vdLs post and Eric Chans answer "The current behavior is a bug and will be fixed").
It seems logical, because if I want to softproof using a printer profile, I am not interested in the areas that exceed the monitor gamut in the _original_ image. I am only interested in the areas that exceed the monitor gamut in the _converted_ image - because only these areas will _not_ look the same on the monitor vs. printed output.
I had not looked at the issue that way, but the observed behavior does make the most sense and I must retract my statement that the monitor gamut warning is seriously flawed. Eric Chan understands the issues better than I, and the monitor gamut warning could be useful.
Bill
-
I have a hypothesis to explain my original observations but need the group's expertise as to whether this explanation is, in fact, plausible.
Let's assume that the sRGB Granger chart has been properly generated and that its embedded sRGB profile correctly matches the colour space used to generate the image. Also recall that there is no control for the sensitivity of Soft Proofing gamut warnings. I assume then that even minor (perhaps barely perceptible) colour mismatches will be flagged with Destination gamut warnings.
I understand Lightroom performs all its internal calculations using the (very wide gamut) Pro Photo colour space. Thus both the specified soft proofing sRGB profile and the embedded image sRGB profile would have to be converted to Pro Photo.
I suspect that small differences in the sRGB profiles (or possibly LR is using a slightly different conversion algorithm in each case) could account for the observed gamut warnings. To verify this explanation, one would have to be able to measure the deviation between the flagged image colours and the corresponding reference colours used by Soft Proofing.
Comments?
- David
-
I have a hypothesis to explain my original observations but need the group's expertise as to whether this explanation is, in fact, plausible.
Let's assume that the sRGB Granger chart has been properly generated and that its embedded sRGB profile correctly matches the colour space used to generate the image. Also recall that there is no control for the sensitivity of Soft Proofing gamut warnings. I assume then that even minor (perhaps barely perceptible) colour mismatches will be flagged with Destination gamut warnings.
I understand Lightroom performs all its internal calculations using the (very wide gamut) Pro Photo colour space. Thus both the specified soft proofing sRGB profile and the embedded image sRGB profile would have to be converted to Pro Photo.
I suspect that small differences in the sRGB profiles (or possibly LR is using a slightly different conversion algorithm in each case) could account for the observed gamut warnings. To verify this explanation, one would have to be able to measure the deviation between the flagged image colours and the corresponding reference colours used by Soft Proofing.
Comments?
- David
I've used the ExifTool to extract the following information directly from Granger Chart sRGB IEC 61966-2.1.jpg (http://www.pbase.com/dougj/image/112380832/large) (Copyright 2011© Douglas Janson) and from an image exported from Lightroom.
Granger sRGB Profile
Device Manufacturer : IEC
Device Model : sRGB
Device Attributes : Reflective, Glossy, Positive, Color
Rendering Intent : Media-Relative Colorimetric
Connection Space Illuminant : 0.9642 1 0.82491
Profile Creator : HP
Profile ID : 0
Profile Copyright : Copyright (c) 1998 Hewlett-Packard Company
Profile Description : sRGB IEC61966-2.1
Media White Point : 0.95045 1 1.08905
Media Black Point : 0 0 0
Red Matrix Column : 0.43607 0.22249 0.01392
Green Matrix Column : 0.38515 0.71687 0.09708
Blue Matrix Column : 0.14307 0.06061 0.7141
Device Mfg Desc : IEC http://www.iec.ch
Device Model Desc : IEC 61966-2.1 Default RGB colour space - sRGB
Lightroom sRGB Profile
Device Manufacturer : IEC
Device Model : sRGB
Device Attributes : Reflective, Glossy, Positive, Color
Rendering Intent : Perceptual
Connection Space Illuminant : 0.9642 1 0.82491
Profile Creator : HP
Profile ID : 0
Profile Copyright : Copyright (c) 1998 Hewlett-Packard Company
Profile Description : sRGB IEC61966-2.1
Media White Point : 0.95045 1 1.08905
Media Black Point : 0 0 0
Red Matrix Column : 0.43607 0.22249 0.01392
Green Matrix Column : 0.38515 0.71687 0.09708
Blue Matrix Column : 0.14307 0.06061 0.7141
Device Mfg Desc : IEC http://www.iec.ch
Device Model Desc : IEC 61966-2.1 Default RGB colour space - sRGB
Update 1: I tried to remove the Destination Gamut Warnings by using the Primary Saturation sliders. Contrary to my hypothesis, the gamut warnings are highlighting colours that Lightroom believes are well outside the gamut that Soft Proofing is testing for (see attached screenshot).
Update 2: Photoshop Feedback user LRuserXY (http://feedback.photoshop.com/photoshop_family/topics/suspicious-results-using-lightroom-soft-proofing?rfm=1) noted that turning the GPU on or off affects the gamut warnings displayed. I confirmed this for the Granger chart. The Destination Gamut Warning area is larger when the GPU is enabled. This behaviour may be confirmation that colour characterization is a highly sensitive to calculation errors.
Update 3: When I extracted LR's sRGB profile from an JPEG exported RAW file and viewed the Exif data, I noticed a difference in the Rendering Intent (see above).
Q: Could this difference possibly impact the display gamut warnings?
A: No, gamut warnings are independent of the Rendering Intent.
See Introduction to Rending Intents (http://www.wasatch.com/iccintent.html) and Understanding Rendering Intents (http://www.steves-digicams.com/knowledge-center/understanding-rendering-intents.html#b) for more information.
Update 4: Using an ICC profile viewer, the SRGB and embedded Granger Chart gamuts seem to be identical (see attached plot).
Update 5: Final nail in the coffin => sRGB Granger Chart soft proofed with its own (embedded) colour profile still displays identical gamut warnings (see Gamut Warning for Granger Chart Soft Proofed with own Colour Profile (https://drive.google.com/file/d/0BynlzzsBxchZX3FRMmJiZkRxUzA/view?usp=sharing)). I am convinced that observed behaviour is due to a deficiency within Lightroom.
- David
-
I'd be curious to know if the wide gamut of the 4k Ultra HD specification will necessitate a revision to Pro Photo as the gold standard for photography.
In a word, no.
The Rec 2020 gamut is contained within the PPRGB gamut.
(http://www.kasson.com/ll/res2020pprgb.PNG)
Jim
-
In a word, no.
The Rec 2020 gamut is contained within the PPRGB gamut.
Jim
Thanks Jim ... That's reassuring.
- David
-
As mentioned, there are related LR6 issues with the GPU enabled. See Suspicious Results using Lightroom Soft Proofing (http://feedback.photoshop.com/photoshop_family/topics/suspicious-results-using-lightroom-soft-proofing?rfm=1) and LightRoom 6: Inaccurate Develop Preview Colour Mapping for ColorChecker Chart (http://feedback.photoshop.com/photoshop_family/topics/lr6-inaccurate-develop-preview-colour-mapping-for-colorchecker-chart?rfm=1) for more details.
- David
-
In a word, no.
The Rec 2020 gamut is contained within the PPRGB gamut.
(http://www.kasson.com/ll/res2020pprgb.PNG)
Jim
Jim,
How does the gamut of Rec 2020 compare to AdobeRGB? I understand that it is somewhat wider; if so, are there monitors capable of displaying this gamut?
Bill
-
How does the gamut of Rec 2020 compare to AdobeRGB?
Here you go:
(http://www.kasson.com/ll/mongamuts.PNG)
I understand that it is somewhat wider; if so, are there monitors capable of displaying this gamut?
I don't know.
Jim