Pages: 1 [2]   Go Down

Author Topic: Colomunki - edit profiles?  (Read 12840 times)

Paul Ozzello

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 262
    • www.paulozzello.com
Re: Colomunki - edit profiles?
« Reply #20 on: October 17, 2013, 01:31:40 pm »

What is the source for the grayscale? Gray Gamma, dot gain?
Gray Gamma 2.2

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Colomunki - edit profiles?
« Reply #21 on: October 17, 2013, 01:33:24 pm »

Gray Gamma 2.2
So any difference when you convert to say sRGB (which has a 2.2 gamma) vs. ColorMatch RGB which has 1.8?
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Paul Ozzello

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 262
    • www.paulozzello.com
Re: Colomunki - edit profiles?
« Reply #22 on: October 17, 2013, 01:40:32 pm »

So any difference when you convert to say sRGB (which has a 2.2 gamma) vs. ColorMatch RGB which has 1.8?

I don't remember using sRGB, only AdobeRGB and ProPhotoRGB. I'm at work but I'll try it again tonight !

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Colomunki - edit profiles?
« Reply #23 on: October 17, 2013, 02:47:10 pm »

I don't remember using sRGB, only AdobeRGB and ProPhotoRGB. I'm at work but I'll try it again tonight !
You want to match the gray gamma to the RGB color working space gamma.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

NeroMetalliko

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 78
Re: Colomunki - edit profiles?
« Reply #24 on: October 17, 2013, 06:10:24 pm »

So any difference when you convert to say sRGB (which has a 2.2 gamma) vs. ColorMatch RGB which has 1.8?

Hello,
I'm sure you perfectly know that sRGB has not exactly a 2.2 gamma, it is overall similar but it is not a pure gamma curve: according to colorspace definition the very deep shadows are linear.

In addition all gamma curved colorspaces in Adobe software have a numerical implementation that limit the maximum slope, this affect the first 14 RGB triplets in AdobeRGB (the junction point being 14) and GrayGamma 2.2 and the first 4 RGB triplets in ProPhoto.
sRGB is not affected from this slope limit because the linear shape is already part of the colorspace definition itself and not an external implementation applied by the software color management engine.

For the sake of precision the exact gamma exponent in the AdobeRGB and Graygamma 2.2 is 2+(51/256)=2,19921875, as visible looking to the icc with a profile inspector.

Now, assuming an AdobeRGB and GrayGamma 2.2 working spaces in PS,
I bet the little spike you see growing in the histogram after Gray/RGB conversion is exactly at 14,14,14 (which is shifted toward 13,13,13 after the conversion).
You can easily verify it by doing a RGB/gray/RGB loop using a full wedge grayscale.

If this is the case, this is probably the slope limit implementation of Adobe ACE in action, it happens in 16 bit too obviously (which in reality are 15 bit in PS) and this is very curious because in theory the gamma curve (and the limited slope) should be exactly the same for the AdobeRGB and Graygamma 2.2 colorspaces.

Other than this, which is a numerical artifact, you should not see noticeable variations in the RGB values outside the slope limited region above mentioned.

I hope this could be useful,
Ciao.

Andrea :)
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Colomunki - edit profiles?
« Reply #25 on: October 17, 2013, 06:50:22 pm »

I'm sure you perfectly know that sRGB has not exactly a 2.2 gamma
Yes. The OP has to convert to an RGB working space so the gamma should match (closely). One could build a simplified 2.2 sRGB profile directly in Photoshop's Color Setting too. The original data is in color so it seems best to leave it that way but for the time being, the converted grayscale files need to be converted into some RGB color space.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Paul Ozzello

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 262
    • www.paulozzello.com
Re: Colomunki - edit profiles?
« Reply #26 on: October 18, 2013, 12:56:56 pm »

Hello,
I'm sure you perfectly know that sRGB has not exactly a 2.2 gamma, it is overall similar but it is not a pure gamma curve: according to colorspace definition the very deep shadows are linear.

In addition all gamma curved colorspaces in Adobe software have a numerical implementation that limit the maximum slope, this affect the first 14 RGB triplets in AdobeRGB (the junction point being 14) and GrayGamma 2.2 and the first 4 RGB triplets in ProPhoto.
sRGB is not affected from this slope limit because the linear shape is already part of the colorspace definition itself and not an external implementation applied by the software color management engine.

For the sake of precision the exact gamma exponent in the AdobeRGB and Graygamma 2.2 is 2+(51/256)=2,19921875, as visible looking to the icc with a profile inspector.

Now, assuming an AdobeRGB and GrayGamma 2.2 working spaces in PS,
I bet the little spike you see growing in the histogram after Gray/RGB conversion is exactly at 14,14,14 (which is shifted toward 13,13,13 after the conversion).
You can easily verify it by doing a RGB/gray/RGB loop using a full wedge grayscale.

If this is the case, this is probably the slope limit implementation of Adobe ACE in action, it happens in 16 bit too obviously (which in reality are 15 bit in PS) and this is very curious because in theory the gamma curve (and the limited slope) should be exactly the same for the AdobeRGB and Graygamma 2.2 colorspaces.

Other than this, which is a numerical artifact, you should not see noticeable variations in the RGB values outside the slope limited region above mentioned.

I hope this could be useful,
Ciao.

Andrea :)


Hi Andrea, your knowledge of colorspaces is much better than mine, I'm still trying to wrap my brain around everything you said :) I'll have to check again but I believe 14 => 13 is where it's happening, the rest of tones are the same - only very deep shadows are being affected. The histogram changes after the conversion but is not visible on screen until I Softproof, and it is also visible in the final print. What can I do to counteract this ? Is there an alternate method of converting the file ? Unfortunately I have several grayscale file originals that were intended to be printed in QTR using piezography inks. But now I'm using my own printer and want to apply a split tone; aside from this shadow problem Andrew's toning suggestions gives a really nice result.

I tried sRGB but no difference... I can attach a crop of the file tonight if it might be helpful.

Paul

Paul Ozzello

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 262
    • www.paulozzello.com
Re: Colomunki - edit profiles?
« Reply #27 on: October 18, 2013, 01:03:02 pm »

Yes. The OP has to convert to an RGB working space so the gamma should match (closely). One could build a simplified 2.2 sRGB profile directly in Photoshop's Color Setting too. The original data is in color so it seems best to leave it that way but for the time being, the converted grayscale files need to be converted into some RGB color space.


Some are RGB originals but I have several grayscale originals as well :( and months' worth of work tied into these images (I tried copying all my layers to an untouched RGB original but it's way off I really want to avoid having to redo all my work). How do I go about creating a 'simplified' 2.2 sRGB profile in Photoshop ?

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Colomunki - edit profiles?
« Reply #28 on: October 18, 2013, 04:55:09 pm »

How do I go about creating a 'simplified' 2.2 sRGB profile in Photoshop ?
If you go into the Color Settings, RGB working space drop down menu and select sRGB, save that out, the new profile (which you should rename) Photoshop will use a simplified 2.2 gamma curve to create this ICC profile. Not sure that will help or is necessary.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Paul Ozzello

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 262
    • www.paulozzello.com
Re: Colomunki - edit profiles?
« Reply #29 on: October 18, 2013, 07:57:19 pm »

If you go into the Color Settings, RGB working space drop down menu and select sRGB, save that out, the new profile (which you should rename) Photoshop will use a simplified 2.2 gamma curve to create this ICC profile. Not sure that will help or is necessary.

Didn't help...

Here is the grayscale image softproof:


And the the same softproof after converting to RGB:


And a small crop of the original image:
test.psb

We're looking at the bottom left section of



PS - sorry Jeff for hijacking your thread :(
« Last Edit: October 18, 2013, 08:16:28 pm by Paul Ozzello »
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Colomunki - edit profiles?
« Reply #30 on: October 18, 2013, 08:48:17 pm »

But do the two appear the same prior to the soft proof?
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

NeroMetalliko

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 78
Re: Colomunki - edit profiles?
« Reply #31 on: October 19, 2013, 04:44:03 am »

But do the two appear the same prior to the soft proof?

Hello Andrew and Paul,
from what I have seen the two appear nearly the same prior the soft proofing,
so this is something related to the soft proofing calculations/visualization imho...
I have tried several different paper profiles and got every time a different behavior on RGB and GrayGamma2.2 images.
The results are different even if you soft proof choosing perceptual rendering intent without BPC.
So at this point my opinion is to... don't trust the soft proof visualization too much for critical BW shadows appearance.
I know it could seem weird, but I think that if you print the little test crop for real in the RGB and Gray version, it does not show noticeable difference or at least not of the degree you see in the simulated soft proof visualization.

Regarding the not perfect conversion between Gray and RGB:
this is the result of the ACE numerical implementation of the slope limit. This occurs every time you convert between colorspaces based on gamma curve, and  you can easily spot it by a simple RGB->Lab->RGB loop.
If you open a gray full wedge in AdobeRGB and then you make a Lab mode and then a RGB mode again, you will see the weird spike appearing in the histogram around RGB 14.
If you do the same in sRGB you will see that the spike does not appear, because there is not the slope limit active there.
If you set as working spaces sRGB and sGray and you make the RGB/Gray conversion you will see that no spike occurs.

Unfortunately I'm not aware of a simple conversion method in PS that could allow you to avoid this numerical artifact.
(note that this was one of the reason why I have decided to implement my Linearization curves in a DeviceLink icc by doing all the math in floating point: this allows me to make a virtual RGB->Lab->RGB trip with zero errors).

In any case I don't think that in real world prints the numerical mismatch of a single RGB/Gray conversion in the RGB 14 zone should be so noticeable, so my suggestion is to don't worry it too much and to try a test print of the crop without trusting the soft proof visualized mess too much...

Just my 2 cents...

Ciao.

Andrea :)


« Last Edit: October 19, 2013, 04:45:50 am by NeroMetalliko »
Logged

Paul Ozzello

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 262
    • www.paulozzello.com
Re: Colomunki - edit profiles?
« Reply #32 on: October 22, 2013, 12:18:25 am »

But do the two appear the same prior to the soft proof?

Yes, they do.

Paul Ozzello

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 262
    • www.paulozzello.com
Re: Colomunki - edit profiles?
« Reply #33 on: October 22, 2013, 01:11:11 am »

Hello Andrew and Paul,
from what I have seen the two appear nearly the same prior the soft proofing,
so this is something related to the soft proofing calculations/visualization imho...
I have tried several different paper profiles and got every time a different behavior on RGB and GrayGamma2.2 images.
The results are different even if you soft proof choosing perceptual rendering intent without BPC.
So at this point my opinion is to... don't trust the soft proof visualization too much for critical BW shadows appearance.
I know it could seem weird, but I think that if you print the little test crop for real in the RGB and Gray version, it does not show noticeable difference or at least not of the degree you see in the simulated soft proof visualization.

Regarding the not perfect conversion between Gray and RGB:
this is the result of the ACE numerical implementation of the slope limit. This occurs every time you convert between colorspaces based on gamma curve, and  you can easily spot it by a simple RGB->Lab->RGB loop.
If you open a gray full wedge in AdobeRGB and then you make a Lab mode and then a RGB mode again, you will see the weird spike appearing in the histogram around RGB 14.
If you do the same in sRGB you will see that the spike does not appear, because there is not the slope limit active there.
If you set as working spaces sRGB and sGray and you make the RGB/Gray conversion you will see that no spike occurs.

Unfortunately I'm not aware of a simple conversion method in PS that could allow you to avoid this numerical artifact.
(note that this was one of the reason why I have decided to implement my Linearization curves in a DeviceLink icc by doing all the math in floating point: this allows me to make a virtual RGB->Lab->RGB trip with zero errors).

In any case I don't think that in real world prints the numerical mismatch of a single RGB/Gray conversion in the RGB 14 zone should be so noticeable, so my suggestion is to don't worry it too much and to try a test print of the crop without trusting the soft proof visualized mess too much...

Just my 2 cents...

Ciao.

Andrea :)




Thank you both for all the efforts trying to help me out :)

I did some further testing this weekend and went through some of my original scans. Turns out this particular image was scanned in AdobeRGB (I had converted it to grayscale myself). So without modifying the RGB file I softproofed it... and it showed the exact same problems in the shadows ! And stranger still after conversion TO Grayscale the softproof is fine ! So the problem appears to be elsewhere...

 ???

Unfortunately the effect is also noticeable in the final print (the above crop of is 4.5" x 11" of the final 44" print)

Paul

NeroMetalliko

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 78
Re: Colomunki - edit profiles?
« Reply #34 on: October 22, 2013, 05:21:32 am »

Hello Paul,
There are two things:

- the not perfect RGB/Gray conversion done by ACE for the above explained reasons, affecting the very deep shadows. There is no easy way to avoid this, but the visual real effect could be considered negligible for a single conversion in my opinion.
This is somewhat confirmed by the fact that your crops are looking nearly the same before soft proofing and you can spot the slight difference only by looking at the histograms.
So in any case I think this aspect is not something to be really worried about.

- the noticed difference in soft-proofing (and real prints to some degreee, as you have found) in the shadows zone between the same RGB and Gray converted images, even with colorspaces matching the same gamma value.
In addition the results are varying depending on the ICC profile used.
This is very tricky, and the fact that this is not only a visual effect but it is somewhat reflected in the real print is not so easy to explain, given that the source files are looking nearly identical on screen before soft proofing.

I hope someone more skilled than me cold put some light over this aspect.

In the meantime the further only thing I can suggest you is to try to print the RGB/Gray test crops even using "Perceptual" intent and NO black point compensation, and see if the printed results could be better matching the original file as seen on screen before soft proofing.
Maybe this workaround could mitigate the difference to an acceptable degree, let we know.

As last resort, if you could consider feasible avoiding the possibility to have split toned images and to soft proof the preview, try to bypass the ACE at all and print the RGB/Gray test crops letting the printer to manage colors and engaging ABW (neutral dark setting to start) on the Epson (I don't know if there is something equivalent on the big Canon). This will give you a bit of higher dmax too. it could be interesting to see if the printed crops in this case are better matching the files on screen. If the case you have at least another workaround available, I know that this is not ideal but it's still better than none.

Any further comment/opinion is welcome.

Ciao.

Andrea :)

EDIT: I'm suggeting to use "Perceptual" without BPC and not "Relative colorimetric" with BPC,
as wrongly written previously and as now corrected. Sorry for the typo.
« Last Edit: October 22, 2013, 09:00:49 am by NeroMetalliko »
Logged
Pages: 1 [2]   Go Up