Pages: [1]   Go Down

Author Topic: Absolute Colormetric with Qimage doesn't go right  (Read 8553 times)

aaronchan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 617
Absolute Colormetric with Qimage doesn't go right
« on: June 29, 2011, 01:41:50 pm »

Dear all,

I've just got the Qimage Ultimate version 2 days ago.
It's really nice program so far with my ICC profiles and the layout features.

But now I've to do a art repro job.
So, under the Printer ICC, I've selected my ICC profile and selected Absolute Colormetric as my rendering intent. This is the way that I do it in Photoshop.

But the softproof and the actual print came out kinda blue, specially in the paper base color of the image. But I've looked at the softproof from PS, it looks perfect.

So does anyone know what's going on with my settings?

Thanks
Aaron

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: Absolute Colormetric with Qimage doesn't go right
« Reply #1 on: June 29, 2011, 02:19:59 pm »

So, under the Printer ICC, I've selected my ICC profile and selected Absolute Colormetric as my rendering intent. This is the way that I do it in Photoshop.

But the softproof and the actual print came out kinda blue, specially in the paper base color of the image. But I've looked at the softproof from PS, it looks perfect.

So does anyone know what's going on with my settings?

Hi Aaron,

Make sure your document has the correct colorspace profile embedded, and that Qimage picks it up. When you then select the correct output profile everything should be correct.

Do note that in Qimage the Absolute Colorimetric rendering intent doesn't do a whitepoint conversion, so a mismatch between the document's whitepoint and the output profile's whitepoint will cause a colorcast. From a recent post on the QI forums it appears that PS does do a whitepoint conversion, so the explanation for different looks might be found there.

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

aaronchan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 617
Re: Absolute Colormetric with Qimage doesn't go right
« Reply #2 on: June 29, 2011, 02:27:22 pm »

My file is embedded with the Adobe RGB.
And I have disabled the Monitor ICC.

yannb

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 63
Re: Absolute Colormetric with Qimage doesn't go right
« Reply #3 on: June 29, 2011, 03:18:23 pm »

My file is embedded with the Adobe RGB.
And I have disabled the Monitor ICC.

Hi,

AdobeRGB has a D65 white point, and your printer profile is probably built with D50 viewing conditions in mind. With an absolute colorimetric rendering intent, this can cause a blueish result on paper, at least with some CMM engines, as BartVanderwolf mentioned. Give ProPhotoRGB (with 16 bit data) or eciRGBv2 (www.eci.org) a try as your working space: these have D50 white points, and your artwork reproduction might benefit from the different gamut volumes as well.

Regards,
Yann
Logged

aaronchan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 617
Re: Absolute Colormetric with Qimage doesn't go right
« Reply #4 on: June 29, 2011, 04:50:39 pm »

Hey Yann

It Works!
But how come when I generate a profile under i1P with D65 as the illuminant, that wouldn't work with the Adobe RGB file.
Strange...

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: Absolute Colormetric with Qimage doesn't go right
« Reply #5 on: June 29, 2011, 06:29:33 pm »

Hey Yann

It Works!
But how come when I generate a profile under i1P with D65 as the illuminant, that wouldn't work with the Adobe RGB file.
Strange...

Hi Aaron,

Wouldn't work in Photoshop or Qimage? Do you create a version 2 or a version 4 ICC output profile? Version 4 is not (yet) supported as well as a version 2 profile.

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

aaronchan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 617
Re: Absolute Colormetric with Qimage doesn't go right
« Reply #6 on: June 29, 2011, 07:13:55 pm »

Hi Aaron,

Wouldn't work in Photoshop or Qimage? Do you create a version 2 or a version 4 ICC output profile? Version 4 is not (yet) supported as well as a version 2 profile.

Cheers,
Bart

I think like what mike, the MAN of Qimage, said that before, PS absolute colormetric does not take it as "absolute" as other program does which will convert the illuminant to the destination profile.

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Absolute Colormetric with Qimage doesn't go right
« Reply #7 on: June 29, 2011, 07:40:35 pm »

With an Absolute colorimetric intent, its the paper white (which is measured to build a profile) that is being affected in a different method than Relative. An Abs Colorimetric intent simulate the paper color. As such, the comment “Do note that in Qimage the Absolute Colorimetric rendering intent doesn't do a whitepoint conversion“ or “PS absolute colormetric does not take it as "absolute" as other program does which will convert the illuminant to the destination profile” confuses me.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

aaronchan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 617
Re: Absolute Colormetric with Qimage doesn't go right
« Reply #8 on: June 29, 2011, 07:44:33 pm »

Sorry about my sentence, I'm not really good in words but here is what I've read from ddisoftware forum:

ProPhoto RGB has a white point of D50 and Adobe RGB has a white point of D65.  I don't see how you can convert from one to the other without getting a color cast if you use Absolute Colorimetric!  The very word "absolute" means you are not "converting" from one white point to another.  Not sure why Adobe's conversion doesn't do the same: they appear to be making a "relative" conversion when it should be absolute.  Simply put, you should get a warm cast if you do an absolute conversion from a D50 space to a D65 space because D50 light (from ProPhoto) will look warm when viewed at D65 (from Adobe).  If you don't want that absolute conversion, then use RC or Perceptual.

Why Adobe is doing it wrong, I'm not sure... I suspect they're taking a "non standard" approach and converting the white point because they know that Absolute Colorimetric intent is of little use to photographers DUE to the way it handles white point.  But I'll stick to the way I do it because doing it the Adobe way is wrong: it's no longer Absolute Colorimetric if you "scale" the white point!

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Absolute Colormetric with Qimage doesn't go right
« Reply #9 on: June 29, 2011, 07:52:01 pm »

Sorry about my sentence, I'm not really good in words but here is what I've read from ddisoftware forum:
ProPhoto RGB has a white point of D50 and Adobe RGB has a white point of D65.  I don't see how you can convert from one to the other without getting a color cast if you use Absolute Colorimetric! 

So you are using working spaces for such conversions, not a paper (output profile)?
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Absolute Colormetric with Qimage doesn't go right
« Reply #10 on: June 29, 2011, 07:58:28 pm »

ProPhoto RGB has a white point of D50 and Adobe RGB has a white point of D65.  I don't see how you can convert from one to the other without getting a color cast if you use Absolute Colorimetric! 

Which is exactly what happens in Photoshop when you conduct such a conversion from ProPhoto to Adobe RGB (1998) using Abs Colorimetric. Photoshop is doing what it should correctly.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

aaronchan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 617
Re: Absolute Colormetric with Qimage doesn't go right
« Reply #11 on: June 29, 2011, 08:33:07 pm »

So you are using working spaces for such conversions, not a paper (output profile)?

This is just something that I've read online which gave me the idea of what I might have done wrong.

At the beginning, my file's working space was in Adobe RGB and my profile which is generated in i1P with the D50 setting under the Lighting section.
And I am using Qimage to make the print.
When I use Qimage to soft proof and do the actual print, the image came out very cold compare to the original artwork ( and I know the file is good without that much problem)
So then I took Yann's suggestion, covert the file working space from Adobe RGB to ProPhoto RGB.
And suddenly it works perfectly.

So, I assuming the problem was about the white point of my file working space.


But one of my confusion is:
When I first got the problem of the cold look, and I use the same ICC file and do the soft proof in Photoshop,
the soft proof looks fine.
Is that mean Qimage takes absolute colormetric differently than Photoshop?

Thanks
Aaron

Ernst Dinkla

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4005
Re: Absolute Colormetric with Qimage doesn't go right
« Reply #12 on: June 30, 2011, 02:53:10 am »

So you are using working spaces for such conversions, not a paper (output profile)?

The quote describes a comment of Mike on my report of what I thought was a bug in Qimage's CM. And that was an Absolute Colormetric conversion on assigned color spaces. I try to avoid any conversion on color spaces but in this case it happened for some reason and compared to PS it delivered another outcome in Qimage. Whether Mike is right in his approach or not is something I do not worry about, I was not using the right route anyway.


met vriendelijke groeten, Ernst

New: Spectral plots of +250 inkjet papers:

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
Logged

Ernst Dinkla

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4005
Re: Absolute Colormetric with Qimage doesn't go right
« Reply #13 on: June 30, 2011, 03:25:02 am »

Which is exactly what happens in Photoshop when you conduct such a conversion from ProPhoto to Adobe RGB (1998) using Abs Colorimetric. Photoshop is doing what it should correctly.

There is quite a difference between the Qimage and PS CS4 results using that kind of conversions both ways, AdobeRGB <>ProPhoto.
I would think PS4 compensates to make it more or less practical, Qimage does not.

Edit: picture sRGB original converted to ProPhoto, Absolute Colormetric rendering, PS4 and Qimage Ultimate. ACE as the color engine for PS4, LCMS for Qimage.


met vriendelijke groeten, Ernst

New: Spectral plots of +250 inkjet papers:

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm


« Last Edit: June 30, 2011, 10:20:34 am by Ernst Dinkla »
Logged

yannb

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 63
Re: Absolute Colormetric with Qimage doesn't go right
« Reply #14 on: June 30, 2011, 09:03:17 am »

There is quite a difference between the Qimage and PS CS4 results using that kind of conversions both ways, AdobeRGB <>ProPhoto.
I would think PS4 compensates to make it more or less practical, Qimage does not.

Hello,

Even in Photoshop, this depends on the used CMM to do the conversion. Only with 'Adobe (ACE)' I get no blue cast. With either 'Microsoft CMM' or 'Apple CMM', the result is blue. If I use PS5 on my Mac and use 'Adobe CMM', I get the blue cast as well. On Windows I don't. Strange. So anyway: unless Qimage has access to the Adobe (ACE) conversion method or simular when using absolute colorimetric conversions, you'll end up with a blue cast going from D65 to D50.

This is a nice thing to try in Photoshop: set your conversion options->rendering intent in Color Settings to absolute colorimetric and pick the 'Adobe (ACE)' cmm. Next, look at the Lab values in the Info Palette when you hover over pure white in your AdobeRGB image. The Lab values will be L100, a0, b0. Next, change to Apple CMM or Microsoft CMM in Color Settings. The same RGB white will now read L100, a -2, b -19. Aha, that's blueish white according to Lab D50! This is how most CMM's other than Adobe (ACE) convert from D65 to D50 workspace. Do the same with ProPhotoRGB or ecrRGBv2, and they'll be L100, a0, b0 with either CMM.

I often use Photoshop this way to learn about printer profiles. When using the absolute colorimetric rendering intent in Color Settings, you can very easily 'measure' the Lab values of paper white, solids etc. by assigning different media profiles to images. If you see an 'a' and 'b' of zero, and the paper is not creamy white, then probably UV cutoff filters are used to measure the target, etc etc. Of course, remember to revert this setting if you're fond of doing color conversions using Image->Mode->CMYK...

Regards,
Yann
Logged

Ernst Dinkla

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4005
Re: Absolute Colormetric with Qimage doesn't go right
« Reply #15 on: June 30, 2011, 10:19:13 am »

Hello,

Even in Photoshop, this depends on the used CMM to do the conversion. Only with 'Adobe (ACE)' I get no blue cast. With either 'Microsoft CMM' or 'Apple CMM', the result is blue. If I use PS5 on my Mac and use 'Adobe CMM', I get the blue cast as well. On Windows I don't. Strange. So anyway: unless Qimage has access to the Adobe (ACE) conversion method or simular when using absolute colorimetric conversions, you'll end up with a blue cast going from D65 to D50.
Regards,
Yann


Only checked ACE in PS4 on Windows Vista 64, I see Microsoft CMM shows the blue cast in the sRGB>ProPhoto conversion. I can no longer run Kodak's color engine but I expect ACE is the exception in all cases. Qimage uses LCMS.


met vriendelijke groeten, Ernst

New: Spectral plots of +250 inkjet papers:

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm





Logged

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
Re: Absolute Colormetric with Qimage doesn't go right
« Reply #16 on: July 02, 2011, 05:31:14 pm »

It's a complicated story, if you start in a workspace that is at 5000K whitepoint you should be ok, if not it depends on the CMM etc etc.

Edmund

Dear all,

I've just got the Qimage Ultimate version 2 days ago.
It's really nice program so far with my ICC profiles and the layout features.

But now I've to do a art repro job.
So, under the Printer ICC, I've selected my ICC profile and selected Absolute Colormetric as my rendering intent. This is the way that I do it in Photoshop.

But the softproof and the actual print came out kinda blue, specially in the paper base color of the image. But I've looked at the softproof from PS, it looks perfect.

So does anyone know what's going on with my settings?

Thanks
Aaron
Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald
Pages: [1]   Go Up