Pages: 1 [2] 3   Go Down

Author Topic: The end of profiling?  (Read 22694 times)

aaronchan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 617
Re: The end of profiling?
« Reply #20 on: July 27, 2015, 02:20:00 pm »

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: The end of profiling?
« Reply #21 on: July 27, 2015, 02:31:28 pm »

A JPEG, in sRGB, from a 5D (not using my preferred 105 Macro to show more balls) isn't ideal. But here's my profile (top) versus what Ctein recommends (Adobe RGB (1998) using Epson Controls).

To my eye, nothing is superior on the output using his method. It's clear what clipping to Adobe RGB (1998) does here with Greens (ugly banding without profile). Look at how flat the upper blue ball is, no shape like that with the profile. Upper Magenta ball shows banding and more lack of smoothness. Dirty yellow bottom ball.



Of course, actually viewing the prints on a light booth revels the differences much better, but I think you can get the idea.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

David Mantripp

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 826
    • :: snowhenge dot net ::
Re: The end of profiling?
« Reply #22 on: July 27, 2015, 02:49:09 pm »

Sorry, it's in his 2 hour long $80 video, which has about 15 mins content. A fool and his money .... The section on display calibration, for whatever video capture reason, doesn't even show the actual effect of changing brightness/contrast in the calibrator. Actually he's using an Apple Thunderbolt display, maybe it's sufficiently well factory calibrated... But anyway, my impression is that one reason to use hardware calibration is precisely NOT to introduce "eyesight bias". But what do I know.
Logged
--
David Mantripp

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: The end of profiling?
« Reply #23 on: July 27, 2015, 02:51:49 pm »

Actually he's using an Apple Thunderbolt display, maybe it's sufficiently well factory calibrated...
Calibrated to what for what? No, you're correct, it's worthless advise.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Iliah

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 770
Re: The end of profiling?
« Reply #24 on: July 27, 2015, 02:55:26 pm »

> The section on display calibration, for whatever video capture reason, doesn't even show the actual effect of changing brightness/contrast in the calibrator

Possibly shooting with auto exposure and auto white balance ;)

> Apple Thunderbolt display, maybe it's sufficiently well factory calibrated

With 13.8 dE??
« Last Edit: July 27, 2015, 02:57:29 pm by Iliah »
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: The end of profiling?
« Reply #25 on: July 27, 2015, 03:41:09 pm »

So yes, great idea to ping both Dave P and your Epson contacts on this matter - best to get it clarified from the folks who for sure know the internals.
Did you see that Dave P did post on Ctein's page? Note the part about Windows and sRGB, even less ideal IMHO:

Quote
We really are talking about “printer-managed color,” as Photoshop refers to it. Profiling is turned off entirely. No custom profiles, not even the canned profiles that get installed with your printer software.

To clarify, profiling isn't turned off entirely. Whatever profile the document is in (your working profile in Photoshop, or the document profile if you're not converting to a standard working profile, or ProPhoto RGB if you're working in Lightroom) is sent to the printer with the document, so the printer knows what the numbers in the document mean.

When using "Printer Manages Color" on the Mac, Photoshop will do its level best to send the document as-is to the printer. The caveat here is that the color-space the document is in needs to be something that the OS can turn into what the printer wants. In practice, this means any profiled RGB, CMYK or gray color-space, as well as La*b* (CIELAB) go from Photoshop to the OS without color conversion happening inside Photoshop. But all the printer drivers take (only) RGB input, so there's no reason to even think about the fact that they're using CMYK inks unless you want extra color conversions.

The OS then converts that color-data we've handed it into whatever color-space the printer driver wants (no conversion for profiled RGB, but CMYK, gray, and La*b* will probably be converted by the OS), and then the printer driver is in control. Since the OS is possibly doing a color conversion, the version of the OS matters, but since 10.4 or 10.5 things have been pretty stable (and good).

On Windows, GDI (and ICM2.0) uses sRGB as its interchange space, so if you use "Printer Manages Color" on Windows, all color information will get converted to sRGB at some point, possibly compressing your color gamut. If you can see a difference, use "Photoshop Manages Colors," but make a print first to see if you can see a difference.
Posted by: Dave Polaschek | Monday, 27 July 2015 at 09:33 AM
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: The end of profiling?
« Reply #26 on: July 27, 2015, 04:58:10 pm »

  Mark,

  I apologize, I failed to understand your first post properly.

 Rich

No problem - happens to the best of us. :-)
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Iliah

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 770
Re: The end of profiling?
« Reply #27 on: July 27, 2015, 05:19:50 pm »

> possibly compressing your color gamut

Clipping, in my experience, not compressing, as the conversions between matrix profiles are handled in the system using relative colorimetric intent.
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: The end of profiling?
« Reply #28 on: July 27, 2015, 05:21:24 pm »

Did you see that Dave P did post on Ctein's page? Note the part about Windows and sRGB, even less ideal IMHO:


Thanks for reproducing that Andrew - very interesting. From what Dave says, if the file is in an RGB working space, the OS does not need to do any conversions, so it hands the RGB data to the printer driver. Then it would appear from what you explained further above, "Printer Manages Color" actually has two options - this is where it gets subtle: "Colorsync" or "Epson Color Controls". If I understand correctly, with Colorsync selected, the printer uses the same profiles we can use in an application controlled workflow, so if the results are similar whether the route is the application or the driver controlling, and poorer than what you can get from a custom profile, it simply means the canned profile is inferior to a good custom profile and therefore to avoid. Then you mentioned that if one rather selected "Epson Color Controls", the results are different again, but you say also inferior to a good custom profile. It makes me wonder about the secret sauce by which the printer driver is interpreting the numbers in this option, because there has to be something akin to a profile for doing so. Regardless of the option however, the bottom line emerging from your tests is that both Printer Managed options are inferior to a custom profile with the application managing, but in different ways. Now if you could determine that so readily, one wonders about how Ctein came to believe what he wrote. Perhaps he is using photographs that don't reveal the extent of differences in the way that bespoke test images do, and therefore he's not seeing the whole story. And there remains the question of 3rd party papers which he didn't mention, but clearly can't be left to Printer Colour Management, except in ABW mode where there is no choice.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: The end of profiling?
« Reply #29 on: July 27, 2015, 06:08:38 pm »

It makes me wonder about the secret sauce by which the printer driver is interpreting the numbers in this option, because there has to be something akin to a profile for doing so.
I don't believe it's a secret interpretation, the driver either gets sRGB or Adobe RGB (1998) at some point based on which of the two settings the user selects. Then it does what it does to produce CcMmYKk or whatever.
Quote
Regardless of the option however, the bottom line emerging from your tests is that both Printer Managed options are inferior to a custom profile with the application managing, but in different ways. Now if you could determine that so readily, one wonders about how Ctein came to believe what he wrote.
I wonder if he does believe it (he asks If you do this test, please report back here and let us know what you find. It would be nice to have an assemblage of information on which printers work better with printer-managed color and which don't.). He also replied to my post there, asking about soft proofing (seems to have escaped him?).

IF his technique indeed produced a better result than a custom profile, which it doesn't, that be one thing. That it doesn't and you funnel the data into Adobe RGB and can't soft proof, is another. Two major strikes against this idea. But to be fair to him, the article is titled Are Profiles Obsolete? Hopefully he doesn't believe that.

One thing he writes in his article I'm kind of shocked to read is:
Quote
Then there's printing. For printers, you have a custom profile for every paper you use, although you can often get away with one profile for each category of paper (matte, canvas, glossy, etc.).
Say what? Seems far from ideal and again, no possibility to properly soft proof.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: The end of profiling?
« Reply #30 on: July 27, 2015, 06:23:05 pm »


One thing he writes in his article I'm kind of shocked to read is:  <<For printers, you have a custom profile for every paper you use, although you can often get away with one profile for each category of paper (matte, canvas, glossy, etc.)>>

Say what? Seems far from ideal and again, no possibility to properly soft proof.

There is such a range of papers within each of these major categories in principle it can't possibly be true, even before considering the soft-proofing question. But again, what one sees - or not - partly depends on what one is looking at.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: The end of profiling?
« Reply #31 on: July 27, 2015, 08:15:41 pm »

But again, what one sees - or not - partly depends on what one is looking at.
It will be interesting to see what Ctein has to say to a post I made on his blog (still awaiting moderation):
Quote
Ctein, there are at least three conceptual problems I see using Printer Manages Color and one actual problem:

1. You're clipping all your image data into Adobe RGB (1998) . As I'm sure you're aware, both our printers (P600, 3880) exceed Adobe RGB (1998) by quite a bit (I can provide gamut maps or gamut volume values if you wish). And yes, it's clipping, there's no option for gamut compression as there's no perceptual conversion to Adobe RGB (1998), one less option. All my original data is raw, processed in the ACR color engine which does it's processing in ProPhoto RGB gamut. If you're interested, you or others can test how such a workflow, albeit with very saturated  imagery found form raw, in my Gamut Test File, how and why anything but ProPhoto RGB is suboptimal for output to the devices we're talking about:
-----
AR: I pasted info on video I did about benefits of wide gamut working space.
----
2. You can't soft proof. The profile used to produce the output values equally has to be used for soft proofing. So Printer Color Management restricts soft proofing. Maybe you and your readers don't soft proof? I find it critcally important, especially to produce output specific edits.

3. The Epson driver doesn't know anything about 3rd party papers, how can it? Same with Canon and other drivers. When you select Epson Luster, the driver 'assumes' you're using Epson Luster paper. So while Printer Manages Color might work kind of OK with an Epson paper, how can it know or be optimal for other papers? Do you believe that all papers considered "Luster" or Matt behave the same way? I can't see how using Printer Manages Color with a host of papers from 3rd party companies could work equally well or the same, can you?

Lastly, the actual problem I've seen today. The output using Printer Color Management doesn't get close to producing output that's as good as my custom profile. Hence the photo's of the prints which don't really show all the warts of this process.

Your readers should do their own testing. They should use good color reference images to do so, images that are designed to show issues with color output. You have two, one being the Granger Rainbow and I use that as well. But when testing using my wide gamut file, Bill (Atkinsion's) 14 balls are excellent for uncovering color issues. The saturated images (the blanket image for example) is very useful. Of course this is outlined in the above video.

I'm happy to see that at the end of the title of this article is question mark. Based on the above, I don't think that's necessary as profile are not obsolete. I suppose for those photographers who need a KISS approach and don't mind suboptimal output, using Printer Manages Color is akin to shooting a JPEG in sRGB than shooting raw. I'm OK with that approach as long as the photographer understand the practical limitations of those choices.

Just before, someone posted this which is also interesting:

Quote
Ctein,
Since you are reviewing the new Epson printer, do you know if they are planning to add a windows print feature for RGB? I am currently using Moonphoto in Seattle for custom printing and they do their work in RGB, not sRGB. Regarding the new ink sets, can they handle the pro-photo color space? I am interested in this color space but have not used it.

[ Mathew — Sorry, I don't do Windows, so I wouldn't know about that. There is no printer out there that can fully handle the ProPhoto color space. But, all modern quality printers can produce colors that fall a little outside of Adobe RGB. While Adobe RGB is a roughly good match for printer gamuts, it doesn't catch all the colors a printer can reproduce. Working in ProPhoto RGB does add usable hues to your workspace. — Ctein]
Posted by: Mathew Hargreaves | Monday, 27 July 2015 at 05:52 PM
comment-form-atp

Quote
Interesting article. Thank you.

What profile are you using for soft proofing? Or does Printer Managing Colours render soft proofing unnecessary too?


{ Jason — I don't find soft proofing especially useful for my work; it doesn't tell me well enough what a print will actually look like. So, I couldn't really say the best way to make use of it with this approach. — Ctein ]
« Last Edit: July 27, 2015, 08:17:13 pm by digitaldog »
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Iliah

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 770
Re: The end of profiling?
« Reply #32 on: July 27, 2015, 08:38:58 pm »

I think as most of the shooters have sRGB JPEG files from cameras Epson did an excellent job for this particular audience. However I'm afraid this audience does not print much.

For anything else, including, pardon me, hard proofing, which is lost somehow in the argument, more quality and workflow flexibility, more user control over the workflow are needed.

The good thing is that while working on improvements to "Printer-based colour management" Epson did a grand job polishing and stabilizing ink control. Still, I'm not printing at 20 Celsius all the year around (not to mention humidity changing from 40% to 90% during the year), and it is pretty much the same as with film processing - 2 degrees difference makes for a marked difference in colour and density. Neither are papers perfectly same between batches. IMHO the idea of printer managed colour is not adequate for mid- and high-end printing.

And Retina not needing calibration and profiling is quite odd.
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: The end of profiling?
« Reply #33 on: July 27, 2015, 08:44:57 pm »

I think as most of the shooters have sRGB JPEG files from cameras Epson did an excellent job for this particular audience. However I'm afraid this audience does not print much.
I have to admit, while I've heard the name Ctein over the years, I didn't know much about him. Then I read this just now after reading the blog post Are Profiles Obsolete?:
http://ctein.com/whoami.htm
Quote
I do have an international reputation as one of the premier photographic printers alive; in fact, experts at Kodak at one time declared me to be the finest color printer, bar none. For many years my printing skills have been in great demand by others. If you need custom printing done, I'm your guy.
Maybe he is addressing the sRGB JPEG group at this site. But some of what he states based on his bkgnd has me scratching my head.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Iliah

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 770
Re: The end of profiling?
« Reply #34 on: July 27, 2015, 08:56:02 pm »

I like Ctein, actually. But I'm afraid he got carried away here, just a little :) Point is, simplicity sells. You can witness "over my head" being a very strong argument.
Logged

David Sutton

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1345
    • David Sutton Photography
Re: The end of profiling?
« Reply #35 on: July 27, 2015, 09:49:39 pm »

It will be interesting to see what Ctein has to say to a post I made on his blog (still awaiting moderation):

Wait no longer.  :)
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: The end of profiling?
« Reply #36 on: July 27, 2015, 10:04:36 pm »

Wait no longer.  :)
Ctein contacted me via email, I'm going to build him a couple profiles so he can further test on his end. So, a fellow with very open mind about all this. Stay tuned.
Someone on the blog asked about differences in printing between Lightroom and Photoshop; is one better? They should be identical, they have been in the past, the times I've tested this. So I'm also going to do this test again with LR6 and Photoshop CC 2015.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

amolitor

  • Guest
Re: The end of profiling?
« Reply #37 on: July 28, 2015, 06:17:46 pm »

It is, I feel, interesting to compare the way digitaldog talks about Ctein with the way Ctein talks about digitaldog. It may not say much about color management, but it says a lot about character.
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: The end of profiling?
« Reply #38 on: July 28, 2015, 07:35:29 pm »

Someone on the blog asked about differences in printing between Lightroom and Photoshop; is one better? They should be identical, they have been in the past, the times I've tested this. So I'm also going to do this test again with LR6 and Photoshop CC 2015.
The two behave the same, doesn't matter which you use, you'll get the same color appearance.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Iliah

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 770
Re: The end of profiling?
« Reply #39 on: July 28, 2015, 08:06:10 pm »

> The most important book on photography written in the last 123 years

That is an amazing achievement, Andrew! Can I have a pdf copy for a review please?
Logged
Pages: 1 [2] 3   Go Up