• Global community
    • Language:
      • Deutsch
      • English
      • Español
      • Français
      • Português
  • 日本語コミュニティ
    Dedicated community for Japanese speakers
  • 한국 커뮤니티
    Dedicated community for Korean speakers
Exit
0

JPEG and color management

Explorer ,
Dec 08, 2008 Dec 08, 2008

Copy link to clipboard

Copied

Hi. Is it impossible to get 100% identical colors (if we leave aside the jpeg compression artefacts) when saving a .jpeg file with Photoshop using the highest quality setting (12) ?

Views

11.8K

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Community Beginner ,
Dec 10, 2008 Dec 10, 2008

Copy link to clipboard

Copied

"Peter Figen, I just happened to stumble across some. Look here:

http://www.colourmanagement.ca/resources.shtml "

Mark,

As far as I can see, the images there are all test images, not profile targets - a big difference, and their CMYK test image happens to omit an embedded profile, making it next to useless. Oh well.

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Explorer ,
Dec 10, 2008 Dec 10, 2008

Copy link to clipboard

Copied

Peter,
I just happened to see the links and since you asked for it ...

And I think colour reference targets like this:
http://www.colourmanagement.ca/Images/Colour%20Reference%20Target%20RGB.jpg
should have accurate colours, too.

Anyway,
do you feel the conclusions ( Mark J Peterson, "JPEG and color management" #39, 9 Dec 2008 11:06 pm) I have drawn from your previous posting are correct?

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Community Beginner ,
Dec 10, 2008 Dec 10, 2008

Copy link to clipboard

Copied

I think two things here. One, I've still never seen a profiling target as jpeg - these links are to test images, not profile targets. Second - even if there may be the slightest color shift in some areas in some images, it's not an issue. In dealing with tens of thousands of images over the years, many of them jpegs, I've never come across a circumstance where there was a noticeable color shift. There is far more variability in any printing process than color shift in jpegs.

There has been a single instance where the compression of a CMYK jpeg caused a gradation problem in a very subtle Cyan to paper white gradation that we traced to the file being a jpeg, but that's a single image since 1995.

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Explorer ,
Dec 10, 2008 Dec 10, 2008

Copy link to clipboard

Copied

Thanks Peter Figen for commenting on this from a practical point of view.

So I guess the bottom line of this thread is, that - independently of any JPEG artefacts near edges or in areas with high local contrasts - there can be very subtle shifts in color, even for very large single-colored areas. However the color shift we could demonstrate was only (0,+1,+2) and a professional confirmed that he has known a single incidence of perceptible color shift for real-world images. A JPEG-Expert has confirmed, that neutral grays are much less likely to be affected by any color shifts.

For me, all this was unexpected though, because I thought that for maximum quality JPEGs, large single-colored areas would always have 100% identical colors.

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Engaged ,
Dec 10, 2008 Dec 10, 2008

Copy link to clipboard

Copied

Mark,

My tests agree with what Gernot stated (if I understood him correctly)....that blocks of neutral gray areas remain unchanged, even with high jpg compression levels.

Did you try the test that I ran? Creat a NEW file, sRGB, 200 ppi with a black to white gradient. Posterize it to give you 21 steps of gray. Use this as your "control". Now duplicate this file and save the duplicate as a jpg. Close and reopen the jpg file (just to be sure you are opening jpg compressed data) and look at it under high magnification. Go over it with the eye dropper set to point sample. I don't find ANY portions of ANY gray patch that are not exactly the same as the original file, with the exception of artifacts at or very close to the edges. The grays do NOT shift at all, at least in my experiment.

So, unless I am missing something, I cannot agree with your conclusion that the colors shift in the gray patches.

Lou

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Explorer ,
Dec 10, 2008 Dec 10, 2008

Copy link to clipboard

Copied

Lou Dina, I ran your test and confirm your results.

>So, unless I am missing something, I cannot agree with your conclusion that the colors shift in the gray patches.

I think we misunderstand each other. I can't see where I said in my conclusion, there would be "color shifts in gray patches"

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Engaged ,
Dec 10, 2008 Dec 10, 2008

Copy link to clipboard

Copied

Mark,

Thanks for the feedback. I must have misunderstood what you were saying. Just wanted to make sure we are on the same page. Seems we are.

Lou

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Guide ,
Dec 10, 2008 Dec 10, 2008

Copy link to clipboard

Copied

Mark,

You remind me of one of the two bald men fighting over a comb . :)

But I must admit you've held my attention for a lot longer that I would have ever thought I would spend on the subject of JPEGs. :D

Peter Figen must find this amusing. :)

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Explorer ,
Dec 10, 2008 Dec 10, 2008

Copy link to clipboard

Copied

So everyone accepts and agrees with the conclusions in ( post #53) ?

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Explorer ,
Dec 11, 2008 Dec 11, 2008

Copy link to clipboard

Copied

> One, I've still never seen a profiling target as jpeg - these links are to test images, not profile targets.

It's not as if we wouldn't know the difference. But test images should have accurate colors, too. Anyway, here is a profile target in JPEG format (as you asked for a link, in case one comes my way):

http://www.drycreekphoto.com/images/Charts/MacbethCC-sRGB.jpg

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Engaged ,
Dec 11, 2008 Dec 11, 2008

Copy link to clipboard

Copied

An uncompressed version of the ColorChecker is here
on p.10:
http://www.fho-emden.de/~hoffmann/camcal17122006.pdf

Also available on request as sRGB-EPS or Lab-EPS.
This may help to avoid discussions about the accuracy
of the JPG version.

Best regards --Gernot Hoffmann

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Explorer ,
Dec 11, 2008 Dec 11, 2008

Copy link to clipboard

Copied

Thank you :)

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Guide ,
Dec 11, 2008 Dec 11, 2008

Copy link to clipboard

Copied

LATEST
>So everyone accepts and agrees with the conclusions in (post #53) ?

"Everyone"?

Well, I, for one, am not going back to see what #53 says. Even so, I don't think you've made a single meaningful point so far. 😕 Does it even matter?

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines