Print labs – color management (or not)

[bjanes]

Bill,

The gamut plot below compares the Fuji-generic profiles for the Frontier printer in so-called sRGB-mode vs. PD-mode. The sRGB-mode gamut is somewhat smaller for green/cyan colors. The difference is quite noticeable with Czornyj's test image. The sRGB-mode gamut is somewhat clipped, not as a direct consequence of the limited input space: sRGB, in sRGB-mode the printer ignores embedded profiles anyway, it is simply that the printer does not use its full potential when it is set to sRGB-mode.

The two printermodes: sRGB-mode and PD-mode, may correspond to the difference shown in your post between the Dry Creek Photo Costco profile and the gamut plot from Andrew for another Frontier printer. My understanding from the Dry Creek Photo website is that their profiles refer to the sRGB-mode, which is why "convert-to" works (the embedded profile is ignored, no ICC-type conversion by the printer).

Peter,

Thanks for your comments and the link to the Fuji Europe website. I downloaded two of the Frontier profiles posted there and compared the gamuts to that produced by Drycreek for my local Costco. Plotting the gamuts in Colorthink, I see that the gamut for the profile of my local Costco is indeed smaller than the one posted by Andrew and those on the Fuji web site. The gamut projections at the bottom of the plots give a good indication of the 3D gamuts. I will try to talk with the manager of the photo department at the store and find out what Frontier mode they are using and why thier gamut is so small.

Regards,

Bill

[Peter_DL]

Re: 2.) Is it worthwhile at all to bother oneself with this sort of color management ?

2. Probably not.

One reason which supports your No, and which speaks against trying to color-manage a Print lab by means of their ICC profiles is indirectly related to the relatively small print gamut (compared to inkjet prints), which leaves a lot of room, maybe too much, for out-of-gamut colors and their arbitrary handling by the RelCol or Perceptual intent of the profile.

I can understand that different profiles come to a different balance regarding lightness and saturation when moving out-of-gamut colors, but large hue shifts should be avoided, as I believe, which is unfortunately not what I find (even with RelCol).

The larger the print gamut the less nonsense can the profile do in this regard.

--

[digitaldog]

Re: 2.) Is it worthwhile at all to bother oneself with this sort of color management ?

One reason which supports your No, and which speaks against trying to color-manage a Print lab by means of their ICC profiles is indirectly related to the relatively small print gamut (compared to inkjet prints), which leaves a lot of room, maybe too much, for out-of-gamut colors and their arbitrary handling by the RelCol or Perceptual intent of the profile.

Perhaps. I don't think the gamut size per se is a reason to throw the baby out with the bath water with respect to a color managed path with a lab because we'd like to:

• Soft proof to achieve WYSIWYG (a main goal of color management).
• Decide what rendering intent visually produces a more desirable result. We are forced to use one.
• Having a profile that correctly defines the output conditions so we can work by the numbers reported by the profile and produce output specific edits.

I say No, because the lab isn't color managed fully despite providing a profile you can't really use. Just send sRGB (or if they allow it, fewer do, Adobe RGB (1998)), hope for the best.

[Peter_DL]

[/list]I say No, because the lab isn't color managed fully despite providing a profile you can't really use. Just send sRGB (or if they allow it, fewer do, Adobe RGB (1998)), hope for the best.

The print gamut size & shape may provide some indication what the lab does.

The 3D plot below compares the ICC profile of the ProAm lab which was mentioned earlier in this thread vs. the Fuji-generic profile for CA-DPII paper and for the printer in PD mode which makes use of the printer's full gamut. Both spaces are quite close together, most likely too close for being just a coincidence.

The RelCol rendering of out-of-gamut colors is however quite differently implemented with both profiles as shown in the image below:
/> left:  original test image in source space eciRGB (its presentation below just suffers a bit from the final screenshot / monitor profile to sRGB conversion).
/> middle:  source space soft-proofed, RelCol+bpc, to the ProAm Profile
/> right:  source space soft-proofed, RelCol+bpc, to the Fuji generic profile for CA-DPII paper and for the printer in PD mode

The ProAm profile likes to shift red hues towards orange, whereas the Fuji-generic profile is more hue preserving but comes out darker and is more prone to posterize.

Yes, sure, we can edit the image under softproof with either profile
but I can't help myself to think that such differences between the profiles and their rendering intents are a mess.

[Stephen Ray]

A couple of things about the post above from Peter_DL,

Are you sure you've selected the correct profiles to compare gamut shells? If I've selected correctly, I show significant differences in shape. The profile from ProAm is for an Agfa D-Lab printer, not a Fuji Frontier, but uses Fuji paper.

Also, it appears they've used different CMS. One is Argyll and the other is Xrite. Further, who knows with what reading instrument?

Devil in the details?

[Peter_DL]

Thanks a lot for checking into the details.

I had used the Fuji Frontier profile as shown again below.

The Agfa-dlab profile from Fuji is somewhat differently shaped (second plot), however, I’d still claim that the main difference is in the profiling software and the implementation of the rendering intents as shown above.

Are you perfectly sure regarding the Agfa printer ?

--