Colour Management and Profiles for ICM

[l.j.nash]

I'm getting a little confused again with Colour Management in HDR Studio, I'll try my best to explain my confusion and hopefully someone can help me ask the right question and perhaps find an answer!

I've been scanning negatives and positive colour slides as either 48Bit HDR or 64bit HDRi so that I can process them at a later date for both web publication and printing.

I have profiled my scanner using a SilverFast target and hopefully I'm right in thinking that this will result in saved raw image data that has a colour space that matches the colour space of the profiled scanner. So an image subsequently opened in HDR Studio reports an Embedded profile of SFprofT (PerfectionV700).icc. So far so good, hopefully.

I should perhaps point out that I also calibrate my monitor using an X-rite Eye One Display 2.

At this point I expect to perform any dust removal, change output resolution and so on in HDR Studio and then choose an appropriate colour space for the intended output. So if the image is intended only for display on a computer monitor I would choose in the Options CMS tab, Profiles for ICM Internal to be "sRGB IEC61966-2.1" and the embedded ICC profiles section shows this as the profile to embed. Again so far so good. Similarly if I intend to print an image to paper I would like to choose a larger colour space like Adobe RGB (1998) or ProPhoto RGB. This is the point at which my confusion has occurred.

If I select Adobe RGB or ProPhoto RGB and click 'Apply' I don't expect to see any difference on my computer monitor, as it doesn't even cover the Adobe RGB completely but I do!? In fact when ProPhoto is selected the image shown in the main image window becomes slightly 'brighter/lighter' (not a very accurate or technically correct term). What am I missing?

Screen shot of CMS Tab

CMS Tab.png (44.56 KiB) Viewed 6808 times

The exported images when viewed in Photoshop CS3 appear exactly as in the HDR Studio preview.

I've tried to provide these two images as examples. If your using Chrome like me it doesn't support reading embedded profiles so both images look the same! You'll have to download or view in Firefox.

Adobe RGB

Adobe RGB Test.jpg (231.25 KiB) Viewed 6808 times

ProPhoto RGB

ProPhoto RGB Test.jpg (232.42 KiB) Viewed 6808 times

I've also read that ProPhoto RGB uses a different gamma setting to Adobe RGB, which I thought might explain the difference but on further reading I don't think that makes any sense. As long as the software displaying the image uses the embedded profile the resulting image displayed should be the same or very similar.

Thank you in advance for any help.

[LSI_Rossee]

Hi,
before getting into the details I'm a little bit confused by your screenshot from the CMS-settings. Under Input--> Internal it currently says <None>. It should say <Use embedded profile>, which is the scanner/input profile. This way you assigned the working space AdobeRGB or ProPhotoRGB directly. That would be like telling the GPS-system in your car to take you to location X without telling it your current position.
Can you doublecheck the correct settings in CMS?

Best,

JAN

[l.j.nash]

Hi Jan,

Thank you for the reply. I'm away from home at the moment but I will do as you ask as soon as I return (this weekend).

I was also confused by the Input --> Internal displaying <None> but as you can see from the screenshot this option is greyed out in HDR Studio and can't be changed, at least for this image anyway. As the Embedded Profile shows SFprofT(Perfectionv700.icc) I assume that the scanned image does contain an embedded profile?

Best Regards,

Lee.

[l.j.nash]

Hi Jan,

I've uploaded the image original negative scan to Dropbox should you wish to see if you can replicate this behaviour.

Link removed by admin

I've also deleted the configuration file for HDR Studio 6.6 in case this was the issue, but my CMS tab still appears as shown in my original screen capture. I have found however, that if Positve is selected rather than Negative (Pos/Neg:) then the CMS tab appears as below;

CMS Tab When Pos/Neg: Positive

CMS Tab Postive Selected.png (14.03 KiB) Viewed 6771 times

Best Regards,

Lee.

[l.j.nash]

Hi Jan,

It's been three weeks since I posted my last reply. I know it took a little while for the post to be approved (3rd November) but I just wanted to check that you were still looking into my issues.

Best Regards,

Lee.

[LSI_Morales]

Hi Lee

Working with negatives will use a different color profile than that of the scanner (it will use the NegaFix curves).
Positives use the scanner profile for transparencies, or the Kodachrome scanner profile in the case of Kodachrome slides. For reflective scans, SilverFast will use the "Reflective" profile.

The rgb profile is used for the conversion from scanner to internal.

As you see in your own CMS dialogue (in positive mode), the embedded profile is the one of the scanner

CMS Embedded profile.jpg (62.32 KiB) Viewed 6742 times

Assigning different rgb profiles will cause some colors to fall out of the color space, it means, those colors will have to interpreted to be placed somewhere within the available color space.
Since Adobe RGB and ProPhoto RGB are different color spaces, the colors of the one picture might be in the limits which fall outside of the other color space, hence the differences

Cheers

[l.j.nash]

Hi,

Thanks for the reply but I'm still not convinced something isn't wrong, I'll try to explain and would request you carefully read my response as it is somewhat long winded. Please ignore Positive scans and concentrate on the negative scanning.

From what I understand of the functioning of SilverFast and from what has been said in replies here the following situation appears to be what is supposed to be happening;

(1) A negative is scanned and the scanners RGB data is recorded unaltered (for an HDRi).

(2) Negafix profiles are used to map the scanned RGB values to positive RGB colour values internally (within SilverFast) and an image is then displayed (in HDR Studio). The displayed colours in the image are limited by the chosen colour space.

(3) The colour values in the saved image are limited by the chosen colour space (chosen in "Profile to embed").

The most obvious problems are these; as I am viewing images displayed in HDR Studio on a monitor that has a colour gamut more limited than the Adobe RGB colour space (probably only 85% of Adobe RGB) it seems reasonable to assume that little or no obvious difference would be perceived when the colour space is changed from Adobe RGB to ProPhoto RGB (only colours at the extremes of the gamut). This is not the case. The most obvious change is that the image actually becomes lighter (again perhaps not the best descriptive term) as well as colour being altered.

Further to the above I have also discovered that when an image saved using ProPhoto RGB is "assigned" the Adobe RGB profile (in say photoshop, but any colour managed application will do) that this image has exactly the same RGB values as an image saved directly from HDR Studio using the Adobe RGB colour space. In other words, rather than those colours being interpreted to be placed somewhere within the available color space (as you suggest), it seems that HDR Studio is not doing any colour space conversion. Instead it appears to be writing the same RGB values and simply assigning which ever RGB profile you have chosen. This would also seem to explain why images saved as ProPhoto RGB appear lighter, as without colour space conversion the different gamma value used by ProPhoto RGB (1.8 rather than 2.2) is causing the displayed image to appear lighter.

In fact if colour space conversion was taking place if I "convert" a ProPhoto RGB exported file into an Adobe RGB colour space (again using a colour management aware piece of software) this image should appear quite similar to an image directly exported from HDR Studio using the Adobe RGB colour space (allowing for some compounding of errors by making the double conversion). However if I do so the resulting images do not match, in fact the converted ProPhoto exported image still appears noticeably lighter.

This behaviour is unchanged in the beta of HDR Studio 8. Has anyone else noticed this issue?

Regards,

Lee.

[LSI_Morales]

Hi lee

From what I understand of the functioning of SilverFast and from what has been said in replies here the following situation appears to be what is supposed to be happening;

(1) A negative is scanned and the scanners RGB data is recorded unaltered (for an HDRi).

(2) Negafix profiles are used to map the scanned RGB values to positive RGB colour values internally (within SilverFast) and an image is then displayed (in HDR Studio). The displayed colours in the image are limited by the chosen colour space.

(3) The colour values in the saved image are limited by the chosen colour space (chosen in "Profile to embed").

So far correct!

The most obvious problems are these; as I am viewing images displayed in HDR Studio on a monitor that has a colour gamut more limited than the Adobe RGB colour space (probably only 85% of Adobe RGB) it seems reasonable to assume that little or no obvious difference would be perceived when the colour space is changed from Adobe RGB to ProPhoto RGB (only colours at the extremes of the gamut)

Not really, since you do not actually see which colors have fallen out the color gamut it would not be right to assume that you should not perceive any differences.

The most obvious change is that the image actually becomes lighter (again perhaps not the best descriptive term) as well as colour being altered.

In fact there are differences.

Further to the above I have also discovered that when an image saved using ProPhoto RGB...

An image saved using Photphoto RGB using which application?

At this point I will have to abruptly change the direction of this conversation.
If you can send us the HDRi file used for this example (using an ftp server or a dropbox account) we will be able to better follow the whole issue.

Thanks in advance

[l.j.nash]

Hi,

I don't mean to sound pedantic but I did say

little or no obvious difference

rather than no difference. You certainly don't expect a gamma difference in the outputted files do you?


To respond to your questions:

An image saved using Photphoto RGB using which application?

HDR Studio 6.6 (and HDR Studi 8 beta)

If you can send us the HDRi file used for this example (using an ftp server or a dropbox account) we will be able to better follow the whole issue.

I did upload the image used in this example and it was downloaded by LSI_Rossee. Since then I have deleted this file from DropBox, could you please request this file from your colleague in the first instance. Otherwise I will have to re-upload this evening.

Although I can't be certain I suspect the behavior I'm seeing is not particular to my files.

Just to reiterate, my main suspicion is that HDR Studio rather than mapping the RGB values derived internally (from applying the negafix transformation) to somewhere within the available color space, it seems that HDR Studio is not doing any colour space conversion. Instead it appears to be writing the same RGB values (to file and screen) and simply assigning (tagging if you like) whichever RGB profile has been chosen.

Hopefully we can get to the bottom of this issue.

Best Regards,

Lee.

[l.j.nash]

Hi LSI_Morales,

I've just realised that I can undelete the requested file, it can be donwloaded from here;

http://dl.dropbox.com/u/47396447/D-U.tif

Best Regards,

Lee.

[l.j.nash]

Hi,

I was wondering if anyone at LaserSoft has been able to replicate this major flaw that I seem to have uncovered in HDR Studios RGB colour space transformation (or lack of it)?

Over the last week I have continue to experiment with HDR Studio and am more certain than ever that HDR Studio is not performing the transformation from the internal Negafix derived RGB values to the chosen output colour space (e.g. Adobe RGB, ProPhoto RGB).

In support of this statement I provide the following in addition to all the information previously submitted.

I should start by acknowledging Gary G. Ballard the creator of the 'Whacked RGB' colour space which I shall be using to illustrate my point further. The Whacked RGB colour space has been specifically designed to help root out colour management issues. A particularly enlightening web page is http://www.gballard.net/psd/go_live_pag ... files.html. The Whacked RGB ICC profile can also be downloaded from this website.

As I suspect that HDR Studio is not converting its internal RGB working colour space to the chosen output colour space (that which is embedded) the Whacked RGB colour space is a great test. If a colour space is being converted then you will expect that red will still be red and blue will still be blue etc... after conversion ignoring subtle differences.

The images below show the CMS tab settings in HDR Studio 6.6 and the resulting output image;

CMS Tab with Whacked RGB colour space selected

Whacked SilverFast HDR CMS Tab.png (14.2 KiB) Viewed 6654 times

Whacked RGB

Whacked RGB Test.jpg (202.78 KiB) Viewed 6654 times

I should again point out that the above image will look quite normal if viewed in a web browser (or any program) that does not support colour management (this points to the fact the RGB values are somewhat similar to the sRGB colour space). What you should see is that the dresses are bright green!

In contrast the image below was saved from HDR Studio with an embedded profile of Adobe RGB and then the colour space was converted to the Whacked RGB colour space in Photoshop CS3;

External conversion to Whacked RGB

Adobe RGB Output converted to Whacked RGB.jpg (210.81 KiB) Viewed 6654 times

In this instance a colour managed browser (or viewer) will show fairly normal looking colours whilst a non-colour managed browser will show the dresses to be blue along with the skin tones!

I would really appreciate a considered response, as if I'm correct then HDR Studio has a serious internal flaw at present. If you need more information please let me know, I want to get to the bottom of this issue.

Best Regards,

Lee.

[LSI_Morales]

Hi Lee,

thanks again for your posts.
Which version number of SilverFast HDR Studio are you using?

Cheers

[l.j.nash]

Hi,

I'm using Silverfast HDR Studio v6.6.2r5, but I've also found the same behaviour when testing the beta release of Silverfast HDR Studio v8.0.0r2.

Regards,

Lee.

[LSI_Morales]

Hi lee,

I have searched for the "whacked" color profile to try to reproduce the problem here but could not find any link to download it from the site you recommended (probably my failure). Is it possible for you to send me the link to download the mentioned profile?

Thanks

[l.j.nash]

Hi,

The file can be downloaded from the following link;

http://www.gballard.net/photoshop/pdi_d ... kedRGB.zip

You will need to download and extracted the file "WhackedRGB.icc".

Regards,

Lee.