Black point compensation does nothing in perceptual rendering intent?

[octagon]

I’ve been assessing a few profiles I’ve built in i1Profiler by converting a test image to the paper profile in PS. That way I can see the RGB values that get sent to the printer, and compare two different profiles and just assess whether one has better shadow detail than the other directly from the numbers.

When doing the conversion I have the option of using BPC or not, and generally it’s doing what I expected, and lifting the very dark areas of the image away from the black point of the paper. It does this in quite a predictable manner when using relative colorimetric rendering intent.

However, when using perceptual rendering intent it seems to do very little to the RGB numbers at all. So my conclusion is that BPC is beneficial to getting shadow detail when using relative colorimetric, but of very little use when using perceptual.

Of course the perceptual RI already seems to have a BPC-like dark value ramp built in.

Can anyone make sense of this? For years I’ve been using BPC with perceptual RI when printing, but I seem to have been wasting my time, or at least under a false apprehension.

[gromit]

Some reading for you:

http://www.color.org/AdobeBPC.pdf
http://cias.rit.edu/~gravure/tt/pdf/cm/TT5_Jorge01.pdf

[tony22]

octogon, I'm actually a little surprised as the implication of how far off the profiles are. If I look at the 24x3 and 125x3 test patches relative to what you got (with BPC in both cases) for RC and Perceptual, the deltaE values are (for 24x3) 5.5 and 6.6, and (for 125x3) 16.1 and 16.5. How are you doing this test print conversion in PS? Are you looking to get a representation of how it will turn out on your printer with this paper? Unless I'm missing something I'd be as worried about the large deltaE in both these profiles. I must be looking at this wrong.

[octagon]

octogon, I'm actually a little surprised as the implication of how far off the profiles are. If I look at the 24x3 and 125x3 test patches relative to what you got (with BPC in both cases) for RC and Perceptual, the deltaE values are (for 24x3) 5.5 and 6.6, and (for 125x3) 16.1 and 16.5. How are you doing this test print conversion in PS? Are you looking to get a representation of how it will turn out on your printer with this paper? Unless I'm missing something I'd be as worried about the large deltaE in both these profiles. I must be looking at this wrong.

I’ve simply converted the test image from Adobe RGB into the printer/paper profile color space, and measured the RGB values of the patches with the eyedropper in PS. Nothing very revolutionary! Once the file is converted to the printer device color space, you are seeing the exact rgb values that PS delivers to the printer driver in a color managed workflow. So it is then possible to make some guesses about how dark or light a rendering will appear in print. Of course you can also make a test print and use your eyes as well.

The difference you see in the RGB values between the RC and Perceptual renderings is indeed large, and is borne out in the printed results. i1Profiler RC+BPC rendering is lighter in the darks than the Perceptual rendering as you have stated.

What I’m surprised about is that BPC has no effect in Perceptual rendering, when I’ve always understood that BPC should be used religiously when printing.

[octagon]

Some reading for you:

http://www.color.org/AdobeBPC.pdf
http://cias.rit.edu/~gravure/tt/pdf/cm/TT5_Jorge01.pdf

Thanks very much for those links. I hadn’t seen either of those pdfs. I think the answer to my question is in these two extracts, one from each link respectively:

Color conversion using Perceptual intent already maps source white to destination white and source black to desti¬nation black. Because this mapping preserves the relationships of the shades, it is unlikely that a whole shadow section will be mapped to the same black value. Therefore, BPC should not be necessary. BPC is available, however, for this rendering intent, to be used with malformed profiles. For a given picture, the user can decide whether using BPC improves the color conversion and can select it or deselect it accordingly.

The perceptual rendering intent was compared to the
media-relative rendering intent with BPC. The results show
very close agreement between both color conversions
over all the near neutral colors, and just slight differences
(less than 3 ΔE00) for higher chroma colors. In practical
terms, the perceptual intent can be interpreted as making
its own BPC and the differences between both conversions
is almost not perceptible. To avoid the confusion among
the users, the perceptual rendering intent should be
suggested instead of the use of BPC.