Pages: [1]   Go Down

Author Topic: Perceptual or Relative Colorimetric?  (Read 12901 times)

huguito

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 379
Perceptual or Relative Colorimetric?
« on: March 09, 2015, 04:14:03 pm »

Other than a peak on the proofing function of photoshop before deciding if you will sen the file to your printer processed as either rendering intent.

Do you have any other general criteria that may push you towards one or the other? Do you see a certain area of the image coming alive more in one or the other intent?

Gloss or matte finish seem to make you lean towards more often to one intent than the other one, maybe?

Do you see a certain kind of papers benefiting more from one of the rendering intents than the other?

In my own workflow, once the digital work is almost complete, I switch the softproof view from one to the other, in my case what normally seems to make the case is detail in the deepest shadows.g  The interesting thing is that sometimes, I find similar images, to be printed both in the same paper, seem to look better in different intents, like image 1 in Perceptual and image 2 in Rel. Col.

Just happened with two images I shoot in Joshua Tree Nat Park, that's the trigger of this question.

Hugo

Logged

JRSmit

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 922
    • Jan R. Smit Fine Art Printing Specialist
Re: Perceptual or Relative Colorimetric?
« Reply #1 on: March 09, 2015, 04:31:50 pm »

Hugo,

This is a choice that is simply put only subjective to make. I have not found a clear rule to follow by. Yes if the image has a lot of dark but detailed areas one tend to go to gloss/satin papers to retain all those details. But more than once in case of darish images a matte paper is ultimately chosen by my customers, simply because their subjective experience takes more into account.
It also depends, in my experience, on the quality of the printer used what you will accept subjectively.
As a coarse guide , with gloss/satin papers, the colorspace is in most cases large enough to prevent disturbing clipping effects. My default is Relative. However with matte papers in combination with colorful images i for sure check the relative versus perpetual and propose if needed perceptual.
Logged
Fine art photography: janrsmit.com
Fine Art Printing Specialist: www.fineartprintingspecialist.nl


Jan R. Smit

Paul2660

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4066
    • Photos of Arkansas
Re: Perceptual or Relative Colorimetric?
« Reply #2 on: March 09, 2015, 04:49:35 pm »

For any glossy media, and PK ink  I will always use Relative colormetric, never perceptual.

With Matte Media, and MK ink  where I often see blocking of shadows details, I will use perceptual more often.  I just can't get a smooth shadow with relative colormetric on matte.


9900 and 7880.  printers.  Profiles mine and stock, neither will work well on matte with smooth transition on shadows with matte ink with the relative CSpace.  I even often will push the shadows in LR before I print.  This issue always seems worse on matte canvas than anything else, but for matte paper the same style of profiles. 

This is not the case with glossy and PK, the profiles are much more dependable.

Paul

Logged
Paul Caldwell
Little Rock, Arkansas U.S.
www.photosofarkansas.com

hjulenissen

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2051
Re: Perceptual or Relative Colorimetric?
« Reply #3 on: March 09, 2015, 04:54:43 pm »

I have assumed that "relative" is the best choice for within-gamut images, while relative/perceptual is a subjective choice for images containing outside-of-gamut values.

Sort of like how you would like a perfectly linear amplifier for playing back some song at moderate volume, while at (potentially clipping volumes) you may prefer a soft clipper that sacrifice some linearity at moderate levels for smooth transitions towards signal peaks.

-h
Logged

huguito

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 379
Re: Perceptual or Relative Colorimetric?
« Reply #4 on: March 09, 2015, 05:04:19 pm »

Thanks everyone.

This is the type of answers, where's a reasoning behind it and past experiences to back it up, is what  I was looking for, even in cases like Paul and Jan where their answers as to what they choose are exactly opposite.

Do you find that the choice of rendering intent is the same when you print black and white?  Do you feel that it doesn't matter much?
Logged

hugowolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1001
Re: Perceptual or Relative Colorimetric?
« Reply #5 on: March 09, 2015, 05:48:46 pm »

I am with ‘h’. I look at an image for out of gamut areas, and if there are none I’ll print with relative colorimetric. Even if there are a few scattered pixels out of gamut, I will print relative colorimetric.

If there are relatively large areas of saturated similar colors that are out of gamut, then I will consider perceptual. An example would be a fall foliage shot with saturated yellow foliage, where many of the different yellows could be mapped to a nasty blur of the same yellow.

Brian A
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Perceptual or Relative Colorimetric?
« Reply #6 on: March 09, 2015, 08:17:01 pm »

Use whatever RI looks best to you. ICC profiles don't know anything about color in context, they view this as just a big pile of solid colored pixels.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

dwswager

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1375
Re: Perceptual or Relative Colorimetric?
« Reply #7 on: March 09, 2015, 08:22:34 pm »

Other than a peak on the proofing function of photoshop before deciding if you will sen the file to your printer processed as either rendering intent.

Do you have any other general criteria that may push you towards one or the other? Do you see a certain area of the image coming alive more in one or the other intent?

Gloss or matte finish seem to make you lean towards more often to one intent than the other one, maybe?

Do you see a certain kind of papers benefiting more from one of the rendering intents than the other?

In my own workflow, once the digital work is almost complete, I switch the softproof view from one to the other, in my case what normally seems to make the case is detail in the deepest shadows.g  The interesting thing is that sometimes, I find similar images, to be printed both in the same paper, seem to look better in different intents, like image 1 in Perceptual and image 2 in Rel. Col.

Just happened with two images I shoot in Joshua Tree Nat Park, that's the trigger of this question.

Hugo

When I make my own profiles, or if I know how other generic/custom profiles I use were created, I utilize the same rendering intent as the profile, even though the profile itself doesn't really have anything to do with it.  But if I print a target using a Rendering Intent when making a profile, I tend to use that same Rendering Intent for print.  For unknown profiles, usually either Relative Colorimetric or Perceptual will look better.

As has been indicated, if the image has out of gamut colors, then Perceptual is usually going to give a better image than Relative Colorimetric.
« Last Edit: March 09, 2015, 08:24:59 pm by dwswager »
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Perceptual or Relative Colorimetric?
« Reply #8 on: March 09, 2015, 08:28:50 pm »

But if I print a target using a Rendering Intent when making a profile, I tend to use that same Rendering Intent for print.  
The target used to create the profile should undergo no rendering intent, it's printed without color management (or should be).
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

hugowolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1001
Re: Perceptual or Relative Colorimetric?
« Reply #9 on: March 09, 2015, 08:37:09 pm »

The target used to create the profile should undergo no rendering intent, it's printed without color management (or should be).

There is rendering intent field in profiles. I don't know what it is supposed to do, but I have seen many new-to-profiles users think it is the suggested intent to use.

Brian A
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Perceptual or Relative Colorimetric?
« Reply #10 on: March 09, 2015, 08:43:48 pm »

There is rendering intent field in profiles. I don't know what it is supposed to do, but I have seen many new-to-profiles users think it is the suggested intent to use.
There's a place to provide a preferred RI for use in applications that don't ask the user. It's usually set for Perceptual. Few products ask you what to place in there and in any situation where  you can pick, that's what is used.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

artobest

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 287
Re: Perceptual or Relative Colorimetric?
« Reply #11 on: March 11, 2015, 09:46:27 am »

Sometimes Perceptual intent can have a disturbing effect on colour relationships within the image. In some cases, if a particular colour is wildly out of gamut, I will try pulling that colour back within the gamut boundaries manually rather than subject the entire image to the black-box Perceptual transformation.
Logged

dwswager

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1375
Re: Perceptual or Relative Colorimetric?
« Reply #12 on: March 11, 2015, 10:51:54 am »

The target used to create the profile should undergo no rendering intent, it's printed without color management (or should be).

Correct.  I have not made profiles for awhile and was thinking of printing calibration images as a point of comparision instead of profiling targets!
Logged

dwswager

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1375
Re: Perceptual or Relative Colorimetric?
« Reply #13 on: March 11, 2015, 10:56:35 am »

Sometimes Perceptual intent can have a disturbing effect on colour relationships within the image. In some cases, if a particular colour is wildly out of gamut, I will try pulling that colour back within the gamut boundaries manually rather than subject the entire image to the black-box Perceptual transformation.

That is my preferred method as well.  I will check what is out of gamut and bring them back.  However, if you do print an image with colors that are significantly outside the printer gamut, Perceptual will usually render a better overall images because it is my experience that Relative Colorimetric will just grossly desaturate those colors.  I once printed a skier with a red hat.  The hat was out of gamut and the entire image looked great except for the almost grey hat!
Logged

aaronchan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 617
Re: Perceptual or Relative Colorimetric?
« Reply #14 on: March 11, 2015, 11:28:12 am »

This is only my personal reference:

I used to like relative colormetric render intent for all of my print work because it gives the exact look of what it turns out of my original digital image. But now, I mostly print with perceptual. P RI gives a better overall contrast and make the image pop a bit more comparing to RC RI. But there is an exception, when I do need to do a toned B&W print, let's say sepia, I still use RC RI, because P RI will shift the color tone by some how.

aaron

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Perceptual or Relative Colorimetric?
« Reply #15 on: March 11, 2015, 11:51:44 am »

I will check what is out of gamut and bring them back. 
IMHO, largely a huge waste of your time. The OOG overlay is buggy and inaccurate. The profile will do a much better, through job of mapping OOG into gamut as illustrated here:
http://digitaldog.net/files/LR4_softproof2.mov
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

dwswager

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1375
Re: Perceptual or Relative Colorimetric?
« Reply #16 on: March 11, 2015, 02:55:12 pm »

IMHO, largely a huge waste of your time. The OOG overlay is buggy and inaccurate. The profile will do a much better, through job of mapping OOG into gamut as illustrated here:
http://digitaldog.net/files/LR4_softproof2.mov

Having watched your video, I have 2 issues.

When converting to another color space, one of the issues is selecting the rendering intent!  As opposed to printing, I would almost always use perceptual intent which is the out of the box default for Photoshop (I assume lightroom as well).  There are times another intent would be better, especially if you have target colors you are trying to hold.

This leads to the 2nd issue.  In the video, you compared a desaturated image to that made by converting to the target sRGB profile by I will assume perceptual intent.  Desaturation alone is usually the worst way to bring an out of gamut color into gamut.  I have found using all 3 HSL sliders in combination, without pushing the hue much at all can bring the colors into gamut with almost an imperceptible change in look of the image. 
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Perceptual or Relative Colorimetric?
« Reply #17 on: March 11, 2015, 03:02:58 pm »

"Issue" one as I understand it isn't an issue. You have to view two, maybe three rendering intents (nothing wrong with Saturation depending on the profile) for printing. The one that looks best is the right answer.
The only time I'd convert to sRGB is when sRGB is the appropriate output color space and there's only one intent to select. No matter if the output is print or sRGB, I soft proof and I can if I so desire, make output specific edits (in LR that be on a Proof Copy) and make that iteration look as good as I can. None of this has anything to do with OOG overlays.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".
Pages: [1]   Go Up