Perceptual Rendering Intent -- ProPhoto to sRGB

[jc1]

Installation Procedure
1. Download and unzip file. 2 icc files were enclosed in the zip file.
    Download here <ver 1.0>
    a) 0_jc1RGB (beta1).icc    
    b) 0_sRGB_D50_jc1 (beta1).icm
2. Drag or copy the files into the system icc file folder or
    in PC, left click on the icc file and select install.
3. Start Photoshop CS and confirm that the 2 icc files are available for selection.

Perceptual Rendering Intent -- conversion from ProPhoto to sRGB color space
Load your ProPhoto (wide gamut rgb, best rgb or beta rgb ) image file into Photoshop CS.
1. In CS, convert with 0_jc1RGB (beta1).icc
    Rendering intent is always absolute regardless of intent selected.
    You may observe insignificant changes on your display screen if color gamut for your image is well within the color gamut for 0_jc1RGB. Make sure you click OK to complete the conversion.

    Both the downloaded icc files use standard D50 illuminant and Gamma 2.2 and they are of V2 icc specification.
      
2. Next, convert with 0_sRGB_D50_jc1 (beta1).icm
    This icc profile resemblances closely the standard sRGB, but has added saturation and perceptual rendering intents.
    Choose Perceptual rendering for best image reproduction.
    Toggle PREVIEW to detect the differences before and after conversion.
    Depending on the image color gamut, the other rendering intents may work for you as well.

Now, your image has been converted to sRGB color space from ProPhoto color space, perceptually (provided it works!). You can do a profile assign to sRGB or convert to sRGB. The file size for 0_sRGB_D50_jc1 (beta1).icm is about 1.4 Mbytes and hence may increase the storage file size significantly if you choose not to re-assign or convert to standard sRGB.

Examine your converted image and compare it with the original image.

Problem encountered so far:
1) Blue channel for Wide Gamut rgb has knee effect (higher red), but only limited to range  240< B <255. Unsure if this phenomenon is due to highly saturated color being displayed on a standard sRGB monitor.  < Solved >
2) My testings are limited to using simulated color patches. I am unsure how well it works for actual high gamut images.
3) I am using a standard sRGB monitor and hence cannot see the true reproduction color for highly saturated colors. I can only peek and poke the color values ( rgb and Lab).  

Feedback your observation and post your converted image here. You can help me in fine tuning these icc profiles for better color reproduction.

You may send your image file to jcrgb@yahoo.com, tiff file preferred and limited to total 10 files for now.

Thanks.

jc


... Update download link (pm me if you can't download)

[jc1]

Example 1
Converting a ProPhoto GrangerRainbow to sRGB color space with Perceptual Rendering Intent
 
The analog version can be generated with ArgyllCMS. I sampled it with 50x100 pixels and output it to RGB color patches.
 
You can download my digital GrangerRainbo Here. The file was tagged with ProPhoto profile and it can be easily re-assigned to other color spaces with Photoshop.
 
Following images were captured in sRGB. Before capturing, my custom monitor profile was disabled and switched to sRGB color space. So the RGB values for the captured images were truly sRGB.
 
1) Converting using the 2-step workflow


 

 

 

 
 
2) Converting without step1
 


Cheers

jc

... / update download link + file attachment
.../  Disable the tif and change it to .rar, as pc always auto download the tif when I login to LL. (unsure why?)
.../  Problem might be coming from attachment of one of the thread by a forum-er, kindly archive it.
.../  Do not attach jpg or tif file attachment directly. Thanks

[jc1]

 
Upate: jcRGB_beta2 (attachment)
 
a) Revised:  0_jc1RGB (beta2)                 
b) Renamed: 0_sRGB_D50_jc1 (beta2)       
 
Change: Re-locate slightly Blue Primary in 0_jc1RGB
. This greatly improves the Blue rendering when B>240  (R=G=0) when converting from Wide Gamut RGB to sRGB.
. Effect on ProPhoto mapping to sRGB is insignificant.

Wide Gamut RGB to sRGB perceptual rendering


jc

[jc1]

Color Matching Test: sRGB_D50_jc1 (AbsCol) vs standard sRGB (RelCol)
Total of 4096 RGB test points with pitch of 15 were created. The corresponding Lab values for the RGB data were generated for the 2 profiles, sRGB and sRGB_D50_jc1, respectively. Theirs color differences in ΔE2000 under D50 illuminant (relative intent for sRGB) were then determined. Results are illustrated as follows.
 

jc
 
.../ RGB-4096_test_patches is attached, in CGATS txt file. Please help to verify/confirm above result.

[jc1]

Example 2
Converting a ProPhoto Ramp Pattern to sRGB

Procedure
1. Read and assign ProPhoto profile to the digital Ramp pattern.  <  DL >
2. Convert with 2-Step approach. Convert again to sRGB and save in Tiff file.
3. Repeat above but with sRGB RelCol.
4. Import Tiff into CTP3.
5. Enjoy.

[Alan Goldhammer]

Maybe I'm a little dense, but could you briefly explain what you are trying to do with these posts.  Thank you.

[jc1]

Hi,

I am inspired and wish I can provide solution after reading http://www.luminous-landscape.com/forum/index.php?PHPSESSID=24205b19480ced199ad3e9e3a10ee9ce&topic=50249.0 and other similar threads.
 
Above illustrations are hypothetical examples which may represent worst case scenario (?) for the converting color space. I believe most photographers who are in need for such a solution may not have the appropriate software tool to visualize and analyze them.
 
Whether the approach is workable, again I am unsure, only some real life examples can prove it.
 
If there is similar solution, kindly enlighten.
 
Thanks.
 
With best regards
jc

[PeterAit]

Maybe I'm a little dense, but could you briefly explain what you are trying to do with these posts.  Thank you.

I am wondering, too. But, the fact is that some people are interested in the technology and gadgets of photography, the art and beauty of it do not matter to them. Pixel peepers! Graphs, technical specs, sharpness, these are the things that matter. Faced with a Weston pepper they would immediately put their nose against the glass and whine about the lack of sharpness.

[louoates]

I'm not sure I can wrap my brain around these conversion procedures. That means I don't know if I can apply those downloaded files correctly. I do know that when I convert from ProPhoto to srgb in "save for web" in CS5, I rarely like the result. But even when I post a regular ProPhoto jpeg shot, unconverted to srgb, onto facebook it looks bad. I suspect they do the srgb conversion there. I probably should go back to the top of this string and study harder.

[digitaldog]

I do know that when I convert from ProPhoto to srgb in "save for web" in CS5, I rarely like the result.

Why? Can you post examples (in some color managed means) of what results you don’t like?

[louoates]

I should have said that while the conversion from ProPhoto to sRGB in "save for web" looked okay in Bridge but awful when viewing in Firefox or in facebook. I've had the same problem with going from ProPhoto to max quality jpgs and then to some of my microstock sites. I believe they convert to sRGB also. Sorry, I don't think I can duplicate them and have them show up the same way here.

[JeffKohn]

I should have said that while the conversion from ProPhoto to sRGB in "save for web" looked okay in Bridge but awful when viewing in Firefox or in facebook. I've had the same problem with going from ProPhoto to max quality jpgs and then to some of my microstock sites. I believe they convert to sRGB also. Sorry, I don't think I can duplicate them and have them show up the same way here.

If they looked OK in Bridge after conversion, the problem wasn't with the conversion to sRGB. More likely, your web browser is not using your monitor profile to display the image, but Bridge is.

FireFox does support ICC profiles, but there are some limitations depending on which version you're running, how it's configured, and what version the ICC profile is.

[digitaldog]

I should have said that while the conversion from ProPhoto to sRGB in "save for web" looked okay in Bridge but awful when viewing in Firefox or in facebook.

As Jeff pointed out, that isn’t due to the conversions from ProPhoto to sRGB. If it looked fine in SFW, it will look that way in other ICC aware applications. The issue is Firefox (probably some older version) and facebook. When you convert using either SFW or Convert to Profile from ProPhoto or for that matter all RGB working spaces, the conversions should look very similar if not identical after the conversion.

[Peter_DL]

Load your ProPhoto (wide gamut rgb, best rgb or beta rgb ) image file into Photoshop CS.
1. In CS, convert with 0_jc1RGB (beta1).icc
    Rendering intent is always absolute regardless of intent selected.
    You may observe insignificant changes on your display screen if color gamut for your image is well within the color gamut for 0_jc1RGB. Make sure you click OK to complete the conversion.

    Both the downloaded icc files use standard D50 illuminant and Gamma 2.2 and they are of V2 icc specification.
      
2. Next, convert with 0_sRGB_D50_jc1 (beta1).icm
    This icc profile resemblances closely the standard sRGB, but has added saturation and perceptual rendering intents.
    Choose Perceptual rendering for best image reproduction.
    Toggle PREVIEW to detect the differences before and after conversion.
    Depending on the image color gamut, the other rendering intents may work for you as well.

Now, your image has been converted to sRGB color space from ProPhoto color space, perceptually (provided it works!). You can do a profile assign to sRGB or convert to sRGB. The file size for 0_sRGB_D50_jc1 (beta1).icm is about 1.4 Mbytes and hence may increase the storage file size significantly if you choose not to re-assign or convert to standard sRGB.

Please kindly explain the components of these profiles (matrix primaries, white point, TRC / gamma, ...)
and the idea behind the double conversion starting with AbsCol.

Thanks.
& Best regards, Peter

--

[jc1]

Update: Beta2.1  [ DL  ]
Changes: minor adjust made to RGB primaries.

Hi Peter,

Please kindly explain the components of these profiles (matrix primaries, white point, TRC / gamma, ...)
and the idea behind the double conversion starting with AbsCol.

Thank you for your response. It may be easier if I answer briefly with some graphs which are self explanatory.

1) jc1RGB.icc (Intermediate space for color conversion, Step 1)

2) sRGB_D50_jc1.icm (Perceptual mapping, Step2)
Supports rendering intents (AbsCol / RelCol, Saturation and Perceptual).
Following graph is a repeat of a previous post and more details can be found there.

3) Why AbsCol?
All with D50 illuminant, RelCol is thus irrelevant.

4) Why double step conversion? Why not direct Space-to-Space gamut mapping? Why not use BetaRGB, RonRGB, Wide_Gamut_RGB or other standard and available wide spaces as intermediate conversion space?

Before an answer can be given, let me first shoot some photos with real objects (colorful and saturated objects  ) to confirm if above conversion can be applied practically and satisfactory  . I shall organize my observation before back to here. (Sorry if more graphs and photos be posted)

Thanks

jc

.../spelling correction

[jc1]

Test Image 1   Image ref: Dsc_2933.nef

Setup: Nikon 300s with 60mm macro lens, 
          Indoor, available light from window, custom WB with ColorRight Pro.
          Use in-camera matrix metering -0.7 stop, manual mode.
          Set camera picture mode to Neutral and shoot with 12-bit RAW.
Import: Read with Capture NX2 and assigned color profile to ProPhoto.
Export: Image was exported with 16-bit Tiff.
 
Duplicated 3 Tiff. Set Monitor profile to sRGB. Start CS.
Read all 3 files into CS.
1) No change
2) Conversion with 2-Step work flow, then to sRGB.
3) Conversion with RelCol sRGB.
 
The screen images were captured, re-arranged, & saved in png.
 
Converted sRGB files (2 files) can be downloaded here: 2933_converted_jpeg_sRGB_RelCol & Perceptual.rar

Below is the highlight.

 
To be convincing, I need more images to test . If you have for instance, flower shot with highly saturated color and lot of fine details, please give it a go. If possible, post and share your picture here.

jc
 
 

[jc1]

 
Comparison between various options for ProPhoto to sRGB conversion
 
For the following illustration, the colors for the sRGB image were intentionally saturated by assigning it to ProPhoto profile before subjecting it to the conversion.
 
A) Color Reproduction
 

 
B) Brief Summary

1) RelCol sRGB
    .Work fine for image without excessive saturated color.
    .Unpleasant color and loss of details for highly saturated image after conversion.
 
2) Pro2sRGB
    .Dedicated source-to-destination pairwise mapping from ProPhoto to sRGB.
    .Smooth perceptual rendering, always.
    .Can result in excessive color shift and dull color.
    .Perceptual rendering only works for L*a*b* where a* and b* are within the range (-127 to +128).
 
3) 1-Step sRGB perceptual with (0_sRGB_D50_jc1)
    .Works satisfactorily if color gamut for the image is not bigger than that for BetaRGB, for instance.
 
4) 2-Step sRGB perceptual with (0_jc1RGB + 0_sRGB_D50_jc1)
    .Need 2 steps (unsure if collink is applicable for combining them to a single profile)
    .Works satisfactorily for all cases,  including a* and b* values that are out of range of (-127 to +128)
    .Good color saturation can still be realised.
    .Perceptual and pleasant color rendering at the expense of colorimetric accuracy*
 
* This applies to perceptual rendering intent in general.   
     
jc
 

[Peter_DL]

For the following illustration, the colors for the sRGB image were intentionally saturated by assigning it to ProPhoto profile before subjecting it to the conversion.

By means of such pRGB assignment it can easily happen that "imaginary colors" are created which do not exist in reality. Then, it could be a kind of creating a problem first, before solving it. But I don’t want to sound negative at all.

If I get you right, suggested approach is as follows:

regular ProPhoto RGB w/ D50 white, 1.8 gamma, pRGB gamut
-- Abs/RelCol -> jc1RGB (beta2.1) w/ D50 white, 2.2 gamma, gamut: somewhat smaller but still large
-- Perceptual -> sRGB_D50_jc1 (beta 2.1) w/ D50 white, 2.2 gamma, sRGB gamut (?)
-- Abs/RelCol -> sRGB IEC61966-2.1

Right ?

The file offered for download of the profiles seems to be a .rar, not a .zip
I for one am not familiar with this format, although Google points me to some freeware for decoding.

JC, - while I have a true respect for people who can "mangle" profiles the way you do,
my impression is that you don’t make it easy for us / people to follow your thoughts.

Best regards, Peter

--

[jc1]

Hi,

By means of such pRGB assignment it can easily happen that "imaginary colors" are created which do not exist in reality. Then, it could be a kind of creating a problem first, before solving it. But I don’t want to sound negative at all.

If I get you right, suggested approach is as follows:

regular ProPhoto RGB w/ D50 white, 1.8 gamma, pRGB gamut
-- Abs/RelCol -> jc1RGB (beta2.1) w/ D50 white, 2.2 gamma, gamut: somewhat smaller but still large
-- Perceptual -> sRGB_D50_jc1 (beta 2.1) w/ D50 white, 2.2 gamma, sRGB gamut (?)
-- Abs/RelCol -> sRGB IEC61966-2.1

Sorry if I did not make myself clear to you.

Imaginary values may be produced via mathematical transformation.
Adobe color model is based on CIELAB 50 hence imaginary color cannot  be created in that way.

You are correct and thank for the clarification.
In practice, rendering intent can be set to Perceptual as follows.
-- Perceptual -> jc1RGB 
-- Perceptual -> sRGB_D50_jc1 
-- Perceptual -> sRGB IEC61966-2.1

By the way, the intermediate color space was chosen after tedious trial and error as I can't find a easy way out.
 
With best regards
jc

[Nigel Johnson]

jc

AFIK, when you assign a profile in Photoshop (versus Convert to a profile) CIE LAB conversion is not used, rather the RGB values in the file are left unchanged and only the associated profile information (colour primaries and tone curve) is changed. Hence, as Peter stated, it is possible to produce imaginary colours in the image.

Here is a test I have done to confirm this (all files are 8-bit TIFF).

• Firstly I created an sRGB file and filled it with 0-R 255-G 0-B a real colour in sRGB (file: sRGB.tif).
• I then assigned the ProPhoto RGB profile to the file (file: ProPhoto_assigned.tif). When the colour is checked it is still 0-R 255-G 0-B which in ProPhoto RGB is entirely imaginary.
• Finally I took the original sRGB file and converted it to ProPhoto RGB (file: Converted_to_ProPhoto.tif). As expected this did not change the appearance of the colour on-screen but it did change the colour values to 138-R 237-G 78-B which in ProPhoto RGB is the same colour as 0-R 255-G 0-B in sRGB.

This confirms that Assigning can create imaginary colours.

Regards
Nigel