Pages: [1]   Go Down

Author Topic: Relative or Absolute Colorimetric?  (Read 21514 times)

Tango_01

  • Newbie
  • *
  • Offline Offline
  • Posts: 20
Relative or Absolute Colorimetric?
« on: June 04, 2007, 08:07:56 pm »

Or Perceptual...

I noticed that Absolute gives more punch and brightness to the photos, and the whites are more white. When I use Relative or Perceptual there is like a grayish cast that dulls the whole picture but it's better for skin tones.

I also notice that if I use Absolute the colors are punchier but the white parts don't get any ink so they get what I think is called "gloss differential"?

What do you guys use and when? And how do you lessen the Gloss diff effect?


Thanks
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Relative or Absolute Colorimetric?
« Reply #1 on: June 04, 2007, 08:43:04 pm »

Quote
Or Perceptual...

I noticed that Absolute gives more punch and brightness to the photos, and the whites are more white. When I use Relative or Perceptual there is like a grayish cast that dulls the whole picture but it's better for skin tones.

I also notice that if I use Absolute the colors are punchier but the white parts don't get any ink so they get what I think is called "gloss differential"?

What do you guys use and when? And how do you lessen the Gloss diff effect?
Thanks
[a href=\"index.php?act=findpost&pid=121137\"][{POST_SNAPBACK}][/a]

Which ever looks better! There's no right answer (well, unless you're doing proofing, Absolute isn't going to fly). Profile don't know anything about images, only devices. It pays to look at the soft proof and pick the intent you prefer visually. Some images will look better with one or the other, however, it's not an extreme difference.

There's no standard in how a perceptually mapping can be conducted. The ICC leaves this up tot he vendor making the profiles. So YMMV.

Absolute simulates the paper white from one device onto another hence its usually only useful for proofing. IOW, you want to see the paper white of a press simulated on an Epson printer that you're using to proof that first device. Anywhere there's a white, some ink might be laid down since the press sheet paper is different. As such, you'll forgo this expect in proofing. Also note, that other than the paper white simulation, Absolute and Relative are the same! They share the same profile tables.

Don't necessarily dismiss Saturation either. Its usually designed for 'business graphics' (pie charts) but might look OK for images. But unless you try it via a soft proof, you'll never know.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Leping

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 87
    • http://www.lepingzha.com
Relative or Absolute Colorimetric?
« Reply #2 on: June 05, 2007, 02:10:55 pm »

In theory when there is no out-of-gamut color the absolute and the relative give you the same print.  In practical, at least on my Epson 3800, it is totally otherwise, just the original post described, much more contrast as well as color saturation with the absolute, either Epson profiles or the ones I made with an Eye-One.

I have been trying to match my color prints to Lightjet/Chromira/LaserLab prints on the same kind of surface (gloss and matte), from the same high contrast files with bright saturated colors.  Only the absolute intend would match the "punch" and "vibrancy" looking of the chemical paper.
« Last Edit: June 05, 2007, 02:11:38 pm by LEPING »
Logged
Leping Zha
www.lepingzha.com leping@att.

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Relative or Absolute Colorimetric?
« Reply #3 on: June 05, 2007, 02:22:42 pm »

Quote
In theory when there is no out-of-gamut color the absolute and the relative give you the same print. 

Well don't forget the paper white simulation with Abs. Otherwise its identical (the same table).

The paper white is also being considered for the perceptual rendering intent depending on how a GMB profile is made.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Relative or Absolute Colorimetric?
« Reply #4 on: June 05, 2007, 08:02:14 pm »

There's no doubt that Absolute intent with proof colors and 'simulate paper color' alters the appearance of the image the least (using Bill Atlinson profiles on my 7600), but unfortunately has a tendency to throw areas of the image out of gamut.

The choice might then boil down to 'which is easier?'; brightening, increasing saturation and local contrast etc in rel col, or selecting out of gamut areas in abs col to bring them back into gamut.
Logged

madmanchan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2115
    • Web
Relative or Absolute Colorimetric?
« Reply #5 on: June 06, 2007, 09:42:12 am »

Quote
In theory when there is no out-of-gamut color the absolute and the relative give you the same print.  In practical, at least on my Epson 3800, it is totally otherwise, just the original post described, much more contrast as well as color saturation with the absolute, either Epson profiles or the ones I made with an Eye-One.
[a href=\"index.php?act=findpost&pid=121275\"][{POST_SNAPBACK}][/a]

That is not quite right. With Relative Colorimetric (RC), the white point of the RGB working space is shifted to match the media white of the paper (which usually isn't neutral). All other colors (including the gray axis) are shifted accordingly. This is exactly what "media relative" means. With Absolute Colorimetric (AC) the white point of the RGB working space isn't shifted; consequently highlight detail will/may be somewhat clipped because the white point of the RGB space (e.g., LAB = (100,0,0)) usually can't be reproduced on the paper. So the neutral highlights with AC will be a little dimmer (darker gray) and you might lose some detail.

Since colors are shifted with RC due to the white point remapping, the definition of gamut boundaries can also change. Thus, what is out of gamut with RC may not be the same as what is out of gamut with AC.

Again, as Andrew says, the practical advice boils down to trying the various intents and picking the one that works best for your image, on an image-by-image basis. Profiles know nothing about the image content, so that's where you have do make the judgment.
Logged
Eric Chan

Tango_01

  • Newbie
  • *
  • Offline Offline
  • Posts: 20
Relative or Absolute Colorimetric?
« Reply #6 on: June 06, 2007, 05:31:45 pm »

Thanks for the replies, I have a better idea now.

I see that choosing a rendering intent is more flexible than I had thought. I was afraid that AC was hardly ever used and I was doing something wrong somewhere along my color managment line.

One thing I know for experience is that Absolute is not good for skin tones in headshots, but I had very nice results on landscapes and city shots on those days when the sky is overcast and the photo lacks contrast and vibrance of its own.
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Relative or Absolute Colorimetric?
« Reply #7 on: June 06, 2007, 05:34:20 pm »

Quote
One thing I know for experience is that Absolute is not good for skin tones in headshots, but I had very nice results on landscapes and city shots on those days when the sky is overcast and the photo lacks contrast and vibrance of its own.
[a href=\"index.php?act=findpost&pid=121467\"][{POST_SNAPBACK}][/a]

IF you're not proofing you shouldn't be using Absolute. Relative is identical expect for the paper white simulation.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Leping

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 87
    • http://www.lepingzha.com
Relative or Absolute Colorimetric?
« Reply #8 on: June 06, 2007, 07:45:19 pm »

Thanks to all -- I have much better idea now also.

And, as Tango_01's experiences, it works out wonders on my color landscape works.  Dislike LCD monitors I still work with two 22" CRTs, and even I calibrate them bi-weekly with my Monaco Optix XR Pro they are weak differenciate highlight values (density range from 0 to ~10) so in order to match my inkjet prints to the chemical prints with perceptual or relative colorimetric I have to make my images look rediculously bright and blown out on my monitor and then softproof with paper white simulation on to decide if the amount of brightness boost is proper.  So far, with the images I tried the absolute intend provided a shortcut.

With B&W the QTR RIP with the curves I made myself has no such problems.

Quote
Thanks for the replies, I have a better idea now.

I see that choosing a rendering intent is more flexible than I had thought. I was afraid that AC was hardly ever used and I was doing something wrong somewhere along my color managment line.

One thing I know for experience is that Absolute is not good for skin tones in headshots, but I had very nice results on landscapes and city shots on those days when the sky is overcast and the photo lacks contrast and vibrance of its own.
[a href=\"index.php?act=findpost&pid=121467\"][{POST_SNAPBACK}][/a]
« Last Edit: June 06, 2007, 07:46:49 pm by LEPING »
Logged
Leping Zha
www.lepingzha.com leping@att.
Pages: [1]   Go Up