What about printer managing color? Ctein says "yes."

[robgo2]

Over at The Online Photographer, Ctein has recently been advocating using "printer manages color" rather than a software driver with custom paper profiles. This is contrary to everything that many of us have learned and believed about fine printing. Is Ctein right, or has he gone off the rails?

http://theonlinephotographer.typepad.com/the_online_photographer/2015/09/product-review-epson-surecolor-p800-printer.html

Rob

[Telecaster]

Note that Ctein is discussing his results with one particular printer and a small number of papers. I take his word for it…the guy knows from printing! What works…works. Nevertheless if I were to buy the printer in question—a likely occurence in the mid-future—I'd still experiment for myself, as I always do. In the end beliefs, claims & traditions mean nothing in the face of evidence & experience.

-Dave-

[David Sutton]

Over at The Online Photographer, Ctein has recently been advocating using "printer manages color" rather than a software driver with custom paper profiles.

Rob

Not quite if you read the caveats. The Mac users I know of have for years been putting a extra print through with "printer manages colour" for comparison's sake. Sometimes they like the result better. It's a moot point on Windows

[Ferp]

He made this claim a couple of months ago in relation to his 3880:

http://theonlinephotographer.typepad.com/the_online_photographer/2015/07/are-profiles-obsolete.html

but Andrew Rodney and Dave Polaschek convinced him otherwise:

http://theonlinephotographer.typepad.com/the_online_photographer/2015/07/andrew-rodney-on-the-importance-of-custom-printer-profiles.html

Seems he's reverted to his original position.  I'm not convinced, although I can image that printer manages colour may work well enough for some images. 

The Windows issue with printer manages colour, resulting in an unwanted conversion to sRGB, was discussed in a thread about Printfab.  If you're printing with Printfab and want it to manage the colours, then to stop Photoshop converting the image to sRGB you have to set the printer profile to be the document profile, and ignore what turns out to be a hollow warning from Photoshop:
http://forum.luminous-landscape.com/index.php?topic=103432.0

[robgo2]

He made this claim a couple of months ago in relation to his 3880:

http://theonlinephotographer.typepad.com/the_online_photographer/2015/07/are-profiles-obsolete.html

but Andrew Rodney and Dave Polaschek convinced him otherwise:

http://theonlinephotographer.typepad.com/the_online_photographer/2015/07/andrew-rodney-on-the-importance-of-custom-printer-profiles.html

Seems he's reverted to his original position.  I'm not convinced, although I can image that printer manages colour may work well enough for some images. 

The Windows issue with printer manages colour, resulting in an unwanted conversion to sRGB, was discussed in a thread about Printfab.  If you're printing with Printfab and want it to manage the colours, then to stop Photoshop converting the image to sRGB you have to set the printer profile to be the document profile, and ignore what turns out to be a hollow warning from Photoshop:
http://forum.luminous-landscape.com/index.php?topic=103432.0

Yes, I read that initial post followed by his about-face, but now Ctein has made another 180 degree turn.  It seems counterintuitive to believe that the properties of a given paper have no relevance to the final print, but that seems to be his current position. Soft proofing, he says, is a waste of time. Just let the printer do its thing, and you will get better output.  But any good scientist knows that experimental evidence must be reproducible before it can be accepted as true. Hence, we need other "experts" and ordinary users to test his claim. As for myself, I am putting my money (literally) on ImagePrint and its profiles until I see convincing evidence that I should not.

Rob

[r010159]

Yes, I read that initial post followed by his about-face, but now Ctein has made another 180 degree turn.  It seems counterintuitive to believe that the properties of a given paper have no relevance to the final print, but that seems to be his current position. Soft proofing, he says, is a waste of time. Just let the printer do its thing, and you will get better output.  But any good scientist knows that experimental evidence must be reproducible before it can be accepted as true. Hence, we need other "experts" and ordinary users to test his claim. As for myself, I am putting my money (literally) on ImagePrint and its profiles until I see convincing evidence that I should not.

Rob

This is not one of his defining moments where he is demonstrating his advanced knowledge of real world printing. Switching opinions like he has would IMO underline his ignorance. He may be a great photographer, I do not know, but his expertise in the print is at this point very debatable.

PS I remember on Silicon Investor, back in its heyday, I did something similar to this without thinking it through. As a consequence, I had received allot of negative feedback, and this caused a hit to my reputation. I imagine this is what is going to happen to him.

Bob

[robgo2]

This is not one of his defining moments where he is demonstrating his advanced knowledge of real world printing. Switching opinions like he has would IMO underline his ignorance. He may be a great photographer, I do not know, but his expertise in the print is at this point very debatable.

PS I remember on Silicon Investor, back in its heyday, I did something similar to this without thinking it through. As a consequence, I had received allot of negative feedback, and this caused a hit to my reputation. I imagine this is what is going to happen to him.

Bob

Ctein says that his definitive article on the subject will be forthcoming soon. I expect there will be much pushback and a lively discussion. Ultimately, he will have to publish samples to prove his point, but those samples may become targets for naysayers who will rebut with their own samples. It should be interesting.

Rob

[robgo2]

Not quite if you read the caveats. The Mac users I know of have for years been putting a extra print through with "printer manages colour" for comparison's sake. Sometimes they like the result better. It's a moot point on Windows

Hmm, I never do that.  Too much wasted time and paper.

Rob

[Ernst Dinkla]

 Soft proofing, he says, is a waste of time.

Rob

I did not read the changes yet there but on this I can agree, at least when reproducing originals I think proof prints are needed, the more when your customer gets the two pieces for control. Preferably with matching viewing and display lights. Soft proofing is nice on subjects you can not compare the print with, like the landscape 5000 miles away. Edit the image as you think it should look like.

There is also the printer profile structure, the parts used for soft proofing or printing can be edited separately and both at input or output. This is not often discussed and many color gurus see profile editing as a rescue of an already broken profile. I agree more if it is on the printing part but wonder about the soft proofing part. Kodak Colorflow Custom Color Tools has a good description of this  in the manual. Not available on the web I am afraid. A reference by Andrew; http://www.ppmag.com/reviews/200503_rodneycm.pdf
The software is not compatible with any recent OS I think.

Edit; After reading the parts on color management in that printer review I think Ctein generalizes too much based on this printer experience. Certainly on the Windows limitation to sRGB when printer CM is selected. Other Windows printer drivers than the P800 driver show a much wider range of RGB spaces and even CMYK color spaces to choose from if the printer color management route is selected.

It would not surprise me if the reviewed Epson printer has improved driver CM LUTs etc for "AdobeRGB" space prints when AdobeRGB images are fed to it, OS-X only then. I wonder whether the next P800 is as consistent for the next user. What surprises me more is that he does not find the media preset choice important, not for third party papers either. Either the papers he used did not differ much and the Epson media presets differ little. Another thought, if Epson gave more choices of color spaces in the OS-X driver and improved the printer color management one may wonder whether it got frustrated with the OS-X interference on CM in the last decade.

We have to wait for the article but I also wonder whether he now advises to use the monitor manufacturer's calibration/profile for AdobeRGB instead of a custom calibration/profiling. There would be some analogy to this printer color management advice.

Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
December 2014 update, 700+ inkjet media white spectral plots

[David Sutton]

Hmm, I never do that.  Too much wasted time and paper.

Rob

That last bit on every roll that usually can't be used I cut into A4 sheets. Very handy for checking things like rendering intent. It's never proved a waste of time, on the contrary.

[PeterAit]

There must be something I don't understand. I thought that whether you used your printing program to manage colors or the printer driver, they would use the same icc profile. Why then would the results differ? Is this not true?

[robgo2]

There must be something I don't understand. I thought that whether you used your printing program to manage colors or the printer driver, they would use the same icc profile. Why then would the results differ? Is this not true?

The printer uses a media/paper type, not a specific profile. However, if Ctein is right in his recommendations, it may mean that there is little practical difference.

Rob

[iCanvas]

About 6 years ago I did tests with "photoshop manages colors" and "printer manages colors" and the results from the "printer manages colors" was superior. There was more "POP" in the results and details in the shadows. Since that time I have always used "printer manages colors". I have a 9900 and print only on Epson Exhibition matte canvas with matte black ink.

Gar

[digitaldog]

Is Ctein right, or has he gone off the rails?

For the prints he tested, he's right (it's his prints, I've never seen them in person). Otherwise, for those that want to soft proof (something he says doesn't work for him so take that as a data point!), those that want to control OOG gamut mapping via soft proof and selection of the Rendering Intent, post edit based on that soft proof, use ACE or BPC, he's in need of a rail alignment IMHO. 

[digitaldog]

Hence, we need other "experts" and ordinary users to test his claim.

That's exactly the advise I advised Ctein to suggest to his readers (did he?). Run a test both ways with a number of difficult, wide gamut and non wide gamut reference images, make your own decisions. See how it affects your workflow. For someone who fails to produce an effective soft proof, the idea of color management becomes less useful begging the question, does soft proofing not work or just not work for some.....

[robgo2]

About 6 years ago I did tests with "photoshop manages colors" and "printer manages colors" and the results from the "printer manages colors" was superior. There was more "POP" in the results and details in the shadows. Since that time I have always used "printer manages colors". I have a 9900 and print only on Epson Exhibition matte canvas with matte black ink.

Gar

Even so, doesn't this just beg the question as to which print version more closely resembles your on-screen version? Presumably, if you soft proof an image, you could dial in more pop and shadow detail, but it would be under your control, not the printer's.  In any event, the printer cannot exceed its inherent gamut limits.  I think that the great unknown in soft proofing may be the quality of the profile.  I would welcome comments on this point from others.

Rob

[Ernst Dinkla]

There must be something I don't understand. I thought that whether you used your printing program to manage colors or the printer driver, they would use the same icc profile. Why then would the results differ? Is this not true?

I doubt that it is the case for all drivers' CM systems. What I understand is that printer driver's color management relies on driver or firmware baked in Look Up Tables for the conversion of assumed color spaces (sRGB, AdobeRGB and sometimes more spaces) with a fixed rendering type. The user has to specify the color space to be expected by the driver and make sure the image has that color space assigned. So simple color management aboard that does not even read the assigned color space. No more LUTs than the number media presets aboard either, maybe even less and so aimed at OEM media. I can not interpret the driver menu selections otherwise and the driver does not differ in choices when it is approached from an application without any color management aboard. I actually think that printer driver manages color is intended for that use. It could be that some drivers check the OS profile archive but that still does not give rendering choices or the use of custom profiles. One of things that make Ctein's advice so dangerous is that we do not know how sophisticated a driver CM is. Printer manufacturers do not inform users on the details either as I guess they think it is there for the great unwashed, another approach to fool proof color management.

Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
December 2014 update, 700+ inkjet media white spectral plots

[AFairley]

That's exactly the advise I advised Ctein to suggest to his readers (did he?). Run a test both ways with a number of difficult, wide gamut and non wide gamut reference images, make your own decisions. See how it affects your workflow. For someone who fails to produce an effective soft proof, the idea of color management becomes less useful begging the question, does soft proofing not work or just not work for some.....

I have a color managed workflow, and I can't get good enough soft proofing to put me anywhere but merely in the ballpark as to the final print appearance with my Epson 3880, using any paper. (I am perfectly willing to admit that my equipment, Dell U3011 and Spdyer Pro 3 are not the best.). I rely on a series of work prints to get me where I want to end up (it usually takes me 3 to get to final)  and I have to believe that is the case for most of us here who are going for more that just "good enough" (which I can usually hit on the first try, applying the adjustments I know a particular paper will require based on my experience.

And I think that is Ctein's point; the difference between using printer manages color or not comes down to a decision of what looks better for the final look you are heading towards in the print, no  different than choosing between choosing between a canned or custom printer/paper profile, or deciding with rendering intent you are going to use, and no matter what you do, or how good your profiling is, what you see on the screen will never match the print exactly because they are two very different media.

  I think the guy has been printing long enough and at a high enough level so that he can make an informed decision about the optimal workflow for his needs.

[robgo2]

I have a color managed workflow, and I can't get good enough soft proofing to put me anywhere but merely in the ballpark as to the final print appearance with my Epson 3880, using any paper. (I am perfectly willing to admit that my equipment, Dell U3011 and Spdyer Pro 3 are not the best.). I rely on a series of work prints to get me where I want to end up (it usually takes me 3 to get to final)  and I have to believe that is the case for most of us here who are going for more that just "good enough" (which I can usually hit on the first try, applying the adjustments I know a particular paper will require based on my experience.

I'm curious as to what papers and profiles you use.  Also, if you were to switch to letting the printer manage color, how many work prints do you estimate would be required to get the desired result.   The only way to answer the question is to run your own trials.  At least with a color managed workflow, you usually get in the ballpark on the first try.

Rob

[MHMG]

The printer uses a media/paper type, not a specific profile. However, if Ctein is right in his recommendations, it may mean that there is little practical difference.

Rob

I don't know about windows, but on the Mac, when you select "printer manages color"  you also have to go to the "color matching" menu in the driver interface. The default is "Colorsync" which then calls the corresponding Epson supplied ICC profile based on what you also chose as your media setting. However, Ctein is choosing the other option. Under the color matching menu, he is selecting "Epson color matching", often simply called "vendor matching" if one is using a printer other than an Epson.

There's a long backstory on this issue so I will be keen to read Ctein's upcoming "Part II" of the story once he posts it on T.O.P.  But, to bring you up to speed you should go to this Ctein article first:

http://theonlinephotographer.typepad.com/the_online_photographer/2011/09/kinda-interesting.html

That said, I will share some personal correspondence, as objectively as I can, that I had very recently with Ctein, and let you begin to draw your own conclusions.  Ctein kindly sent me the source file that produced the print he likes using Epson color controls rather than ICC profiles. I replicated his result very nicely using printer manages color/Epson color controls on my Epson P600. So, printer-to-printer color consistency across Epson's P series of printers seems to be quite good.  I also replicated the fact that a custom built profile applied via "PS manages color" made with PM5 using its "logo Colorful" rendering recipe indeed exacerbated the famous "blue-turns-purple" color reproduction problem in the image in question, and it produced an inferior "out of the box" print compared to the "printer managed color" version.  The source file was also very difficult to fix with further edits in a soft proofing mode unless one was wiling to take heroic steps in Photoshop.

However, I then switched to a custom built PM5 profile using its original "logo classic" rendering recipe, and it gave me a much better starting point for the necessary additional edits ("classic" favors accuracy in tone over accuracy in colorfulness whereas logo colorful favors colorfulness at the expense of tonal accuracy). I didn't go much further with my experimentation than that, but I did convince myself this file can be successfully edited with an ICC workflow if you know what idiosyncracies to avoid in the perceptual and relative rendering flavors of various ICC profiles (yes, even relcol should be considered another vendor biased flavor of rendering), and you have competent PS editing skills. I applied a selective color layer followed by an additional hue/sat layer to get it looking good in a soft proof without the obvious  cyan-blue magenta-blue hue shifting.  In our subsequent discussions, Ctein did tell me he made numerous attempts to get a good print from an ICC profiled workflow, and was unable to do so. I wasn't really seeing that problem once I elected to use the Logo classic recipe, but this does beg the question:  If you don't roll your own custom profiles and have a few profile making options in your bag of tricks, how good are generic profiles for the majority of printmakers who have to rely on vendor supplied canned profiles?  Hence, I can't really disagree with Ctein doubling down on the virtues of his printer manages color workflow for the typical photographer who wants to bring printing in house. For many photographers just starting out in printmaking, it might be a great way to go on an Epson printer, but at the same time I do worry that painting ICC profiled workflows with such a broad brush is doing a disservice to Ctein's larger audience.  Anyway, different strokes for different folks.

In further discussions, Ctein confirmed that he had made other specific color moves on the original scan in order to achieve what he desired in his "printer managed" workflow. Fair enough, but I will simply state this fact: On my calibrated wide gamut NEC Spectraview II monitor,  his edited source file clearly already exhibits the classic "blue turns purple" reproduction problem baked right into it. Ctein appears to be using a factory calibrated IMAC 4K retina machine, so I'm not sure whether he sees this color error. The file was aRGB (the Spectraview shows 99% of that colorspace, whereas the IMAC is much closer to sRGB and may mask this color error for all I know). Anyway, the color error didn't just mysteriously appear in the final ICC profiled print. It is in the aRGB tagged image file Ctein sent to me.  I also verified the hue shifting in the light beams of the image in question (see the article cited above) using the PS info tool set to LAB readouts. An absolutely perfect source-to-destination image reproduction pathway would dutifully reproduce the unwanted hue shift which exists in the source file, thus much of Ctein's empirical findings on printer manages colors should be understood in the context of a carefully edited image being expressly edited on an sRGB-like monitor for printing within the confines of a closed loop Epson printer system.  Attempting a printer manages color approach for this file on any other printer brand would very likely lead to entirely different conclusions.

I hope Ctein wouldn't strongly disagree with my remarks I have made here, but he did say to me my findings about the source file color error are not germane to his particular printmaking environment.  According to him, the fact that epson color controls are fixing his edited image's color errors in the night sky and that he can make easier edits to the file his way than with ICC profiles is indeed a fortunate outcome for him and for this challenging image. Hard to argue him, I guess. Still, I'm am very curious to know why from a technical perspective.  My best guess is that the differences in the two methodologies are probably related to Epson's own hardwired internal rgb to CMYK mapping of the color space which probably bypasses a conventional PCS (profile connection space) used in the ICC color management approach. The PCS does indeed get weird in the blue sector because it's an inherent weakness in the CIE color model, hence one really does need to rely heavily on soft proofing when dealing with these rich deep blue gradients in today's digital imaging world or be prepared to make a lot of iterative prints to get what you want.

To summarize, Ctein's recommendation to use printer manages color on your Epson printer is unquestionably an easy workflow to master if you own an Epson printer and stick with select papers and media settings.  However, I still personally prefer the outstanding practical merits of soft proofing and in open rather than closed loop color management approach, because I routinely print on different printer brands and with many different paper choices. There is no Epson color control setting on my Canon iPF8300. 

cheers,
Mark
http://www.aardenburg-imaging.com