Tool for comparing gamuts

[avogra]

Hello,

I had a discussion with a member of the official capture one forum and I couldn't answer him, how to evaluate if one gamut is completely inside some other gamut.

As here are a few colour experts: Do you know of any free or cheapt tools to visually compare gamuts? Usage could be for example to compare a printer gamut to say adobeRGB so that you can determine if adobeRGB reduces the printable colour. Another example: compare a monitor gamut to a printer gamut to evaluate if the monitor is suitable for soft proofing.

I already discovered https://www.iccview.de/ which would be perfect if it only supported v4 ICCs. Most print services seem to offer v4 profiles only.
Another tool is http://www.gamutvision.com/ but I was not able to get that running. Maybe someone has hints? Does it support v4?

Apart from that I only know of commercial tools that you probably know better than me.

Is it possible to use Argyll for that task?

Best regards,
Alex

[kers]

if you are on a mac you can use the colorsync helper program in the program/utlilities..
don't know how reliable it is, since it there since a long time and have not seen any changes...

you can compare the gamut of adobe RGB ( tranparant) with HP- prof satin paper for instance:

as you can see  some turqoise and reds are better covered on the paper .

[Doug Gray]

Hello,

I had a discussion with a member of the official capture one forum and I couldn't answer him, how to evaluate if one gamut is completely inside some other gamut.

As here are a few colour experts: Do you know of any free or cheapt tools to visually compare gamuts? Usage could be for example to compare a printer gamut to say adobeRGB so that you can determine if adobeRGB reduces the printable colour. Another example: compare a monitor gamut to a printer gamut to evaluate if the monitor is suitable for soft proofing.

I already discovered https://www.iccview.de/ which would be perfect if it only supported v4 ICCs. Most print services seem to offer v4 profiles only.
Another tool is http://www.gamutvision.com/ but I was not able to get that running. Maybe someone has hints? Does it support v4?

Apart from that I only know of commercial tools that you probably know better than me.

Is it possible to use Argyll for that task?

Best regards,
Alex

A few odd facts:

1. No printer can print all of sRGB's gamut.
2. Even the cheapest inkjets can print many colors outside of Adobe RGB just like sRGB, though there's even more unprintable colors in Adobe RGB.
3. ICC profiles are not capable of representing all of Adobe RGB in PCSLAB. For example RGB (0,255,0) has an a* of < -129 which exceeds ICC's L*a*b* range
4. Only printer Colorimetric intents have a defined gamut.

[HerveCornette]

Copra is a German software where the function of displaying gamuts and their comparison 2 to 2 and which can include the colors of the photo is free.
https://colorlogic.de/en/copra/

[avogra]

Thank you for the great input!

@Doug: That are really some interesting facts. especially #3 is puzzling. Why would you restrict the available values especially for a color space that is able to encode all visible colors?

@Kers: I'm on windows, so no option for me 

@Herve: That sounds great! I have already registered. Will download and try it as soon as I'm at home. From the installation instructions, one can activate a 14-day trial, I hope that you can use it after that for inspection?

[avogra]

CoPra works perfectly, thank you for the great hint :-)

[avogra]

Doug,
I still can't wrap my head around the connection between printer gamut and intent. Can you or someone else enlighten me?

My understanding was, that an ICC profile for a printer contians 2 core ingredients:
1. it characterises the gamut of all colours that can be reproduced by the profiled printer, inks and paper.
2. the mapping between numerical values (e.g. rgb or cmyk) fed into the printer and the resulting colours.
In addition, there is a bunch of supplemental information cotained.
I thought that rendering intent is just an instruction for the CMM which would compare the source and target (printer) gamut, and then map the source colours to the target colours according to the intent.

Apparently, my idea is not how it works. I already found out, that the printer profile contains different tables for each intent. But how does it work then? How can an intent be baked into an output profile without knowing, how the source looks like?

[digitaldog]

Doug,
I still can't wrap my head around the connection between printer gamut and intent. Can you or someone else enlighten me?

Gamut mapping (and how) and overall gamut defined by the profile (which differ depending on the software; not all do this equally**).
** http://www.colorwiki.com/wiki/Color_Management_Myths_26-28#Myth_26
As for the intent, depends on if one is providing gamut compression or gamut clipping as seen below.

[Doug Gray]

In addition, there is a bunch of supplemental information cotained.
I thought that rendering intent is just an instruction for the CMM which would compare the source and target (printer) gamut, and then map the source colours to the target colours according to the intent.

Apparently, my idea is not how it works. I already found out, that the printer profile contains different tables for each intent. But how does it work then? How can an intent be baked into an output profile without knowing, how the source looks like?

ICC profiles and CMMs don't use knowledge of the image source gamut for anything other than getting the image colors in PCSLAB or XYZ.

For instance the RGB (255,255,0) in sRGB is the same color as RGB (255,255,60) in Adobe RGB. These will print the same for any Intent though the specific color that gets printed will vary between Perceptual, Relative, and Absolute.

Another way to look at is is that any image in sRGB, when converted to the larger gamut ProPhoto RGB or Adobe RGB, will print the same.

[GWGill]

I still can't wrap my head around the connection between printer gamut and intent. Can you or someone else enlighten me?

I go into it in some detail here, <https://www.argyllcms.com/doc/iccgamutmapping.html>, with specific reference to ArgyllCMS behaviour.

[avogra]

Thank you all for the great input!

The link to your article about gamut mapping helped a lot. Especially the last sentence about how doing the mapping at link time whould solve some problems. I imagined that this is what actually happens. And in that case, a single cLUT to transform into / from PCS and a LUT to describe the input / output gamut would be sufficient right?

Still I don't fully understand, how it works with pre-computed mappings Another guess on my side: The saturation cLUT of the input device maps the input gamut to the full gamut of the PCS. Then the saturation cLUT of the output device maps the PCS gamut to the output gamut. Then saturated colours of the input device would be mapped to saturated colours of the output device. I'm certainly still wrong, am I?

Thank you for your patience by the way!

[Doug Gray]

Still I don't fully understand, how it works with pre-computed mappings Another guess on my side: The saturation cLUT of the input device maps the input gamut to the full gamut of the PCS. Then the saturation cLUT of the output device maps the PCS gamut to the output gamut. Then saturated colours of the input device would be mapped to saturated colours of the output device. I'm certainly still wrong, am I?

You need to wrap your head around the following.

1. The conversions done by a printer ICC profile to/from the printer have no idea what the image profile or gamut is.
2. The conversions done by an ICC profile to/from the image's ICC profile have no idea what the printer profile or gamut is.
3. The conversions to/from the profile PCS and the monitor's profile have no idea what either the printer or image profiles are.

This is the organization that defines ICC specs. There's a lot of details in there but it can be a bit geeky.
https://color.org/index.xalter

[digitaldog]

1. The conversions done by a printer ICC profile to/from the printer have no idea what the image profile or gamut is.
2. The conversions done by an ICC profile to/from the image's ICC profile have no idea what the printer profile or gamut is.
3. The conversions to/from the profile PCS and the monitor's profile have no idea what either the printer or image profiles are.

Well it could, and there is a 'design' to kind of do this, but I know of now product that does.
This is 12 years old, I still don't know what products use the PRGM as outlined.
https://www.color.org/specification/ICC1v43_2010-12.pdf

[avogra]

You need to wrap your head around the following.

1. The conversions done by a printer ICC profile to/from the printer have no idea what the image profile or gamut is.
2. The conversions done by an ICC profile to/from the image's ICC profile have no idea what the printer profile or gamut is.

This one I have understood. What I don't understand yet: How else can the printer profile achieve a gamut mapping then? In my understanding you always need to know two forms to define a mapping between them  If the profile wants to achieve a smooth transition of colours outside the printer gamut, It has to assume a source gamut doesn't it? As a last resort, that could be the maximum gamut the PCS is capable of.
It feels like a very simple knot in my brain that you lough about after untangling it. I should make a small drawing to illustrate where exactly my problem lies. Hopefully that helps you to find my mistake. I will try to find the time later today.

[Doug Gray]

Well it could, and there is a 'design' to kind of do this, but I know of now product that does.
This is 12 years old, I still don't know what products use the PRGM as outlined.
https://www.color.org/specification/ICC1v43_2010-12.pdf

Same here, Andrew. It's kind of like things have just frozen for 2 decades. Sad really. But then there is so much bad info out there even for those things old and established.

[GWGill]

If the profile wants to achieve a smooth transition of colours outside the printer gamut, It has to assume a source gamut doesn't it?

Yes. And if the actual source gamut is not the same as the one assumed when the profile was constructed, then the gamut mapping doesn't quite work as intended. That's the thing about pre-calculated gamut mappings.

As a last resort, that could be the maximum gamut the PCS is capable of.

L*a*b* PCS is intentionally very large. XYZ PCS is insanely large. If you take these as source gamuts you will be compressing all the color out of your images.

An approach that I haven't pursued, but in principle others may, is to make an assumption about typical gamuts of source images (i.e. photographs), and implement a generic and progressive form of compression based on that assumption. By progressive, I mean that rather than attempting to make all of your source gamut fit within the destination, you take a statistical approach, i.e. you create your source gamut as a probability cloud and apply more compression to higher probability source colors, and less (i.e. allow more clipping) to low probability colors. Your gamut mapping characteristic would be defined by a curve mapping probability to amount of clipping (under compression) or equivalently mapping probability to amount of over compression.

[JuanCarlosHernandez]

Hello,

I had a discussion with a member of the official capture one forum and I couldn't answer him, how to evaluate if one gamut is completely inside some other gamut.

As here are a few colour experts: Do you know of any free or cheapt tools to visually compare gamuts? Usage could be for example to compare a printer gamut to say adobeRGB so that you can determine if adobeRGB reduces the printable colour. Another example: compare a monitor gamut to a printer gamut to evaluate if the monitor is suitable for soft proofing.

I already discovered https://www.iccview.de/ which would be perfect if it only supported v4 ICCs. Most print services seem to offer v4 profiles only.
Another tool is http://www.gamutvision.com/ but I was not able to get that running. Maybe someone has hints? Does it support v4?

Apart from that I only know of commercial tools that you probably know better than me.

Is it possible to use Argyll for that task?

Best regards,
Alex

Hi Alex, Beyond the free tools you could explore ColorThink Pro (requires license: https://www.chromix.com/colorthink/) but the beauty of this software is that not only you will be able to compare color profiles i.e. AdobeRGB vs Your monitor profile, but you will aso be able to compare image color against a profile i.e. Paper ICC / Monitor Profile vs Your Image File. To some extent you can already do this by using the gamut warning tool in Lr or Ps, but ColorThink Pro will graph it for you. I think they have a free trial.... Hope it helps....

[avogra]

Thank you for the suggestion. Colorthink looks like an awesome tool. Sadly 150$ is too much to just satisfy my curiosity. I didn't find a demo version.