Epson SureColor P800

[BobDavid]

We've discussed this extensively before, most notably when C'tein over on TOP infamously recommended printer manages colors a couple of years back.  MHMG gave the definitive debunking of C'tein's findings and I expect that the same would apply to Bob's:

http://forum.luminous-landscape.com/index.php?topic=104307.msg857415#msg857415

When someone thinks they've discovered the holy grail, like C'tein and Bob, it's pointless to try to convince them otherwise.  The best we can do is to keep posting Mark's critique so that other's reading this thread have a more balanced view, and can see that printer manages colors is not the holy grail.

I never ever made a sweeping generalization about letting printers manage color. I began this thread with one specific observation:  An Epson SureColor printer is able to manage output on Epson Ultra Premium Luster Photo paper. That's all. Holy grails do not exist in the world of color management. There is however a balance between art and science. 

[Wayne Fox]

We've discussed this extensively before, most notably when C'tein over on TOP infamously recommended printer manages colors a couple of years back.  MHMG gave the definitive debunking of C'tein's findings and I expect that the same would apply to Bob's:

http://forum.luminous-landscape.com/index.php?topic=104307.msg857415#msg857415

When someone thinks they've discovered the holy grail, like C'tein and Bob, it's pointless to try to convince them otherwise.  The best we can do is to keep posting Mark's critique so that other's reading this thread have a more balanced view, and can see that printer manages colors is not the holy grail.

very nice link.  Mark did a nice job with this, I totally forgot about it.

[Farmer]

I didn't think it was so hard.

Yes, you can (and do) get excellent results from Printer Manages colour.  That's not new, but it is better with the newer hardware and drivers.  What it doesn't do is provide a workflow which can cater for multiple media, varied images and gamuts, different monitors and viewing conditions and so on.  It's a less flexible pipeline, but within its confines it's very good (even excellent).

The ICC workflow gives you a consistent, single workflow that caters for all the regular (and some irregular) variables and provides a means of controlling those variables in a predictable manner.  And it can achieve the same results.

So, printer manages is 100% OK if you know what you're doing and you like the results.  If you never step outside of the vendor prepared pipeline, it is entirely possible that you will have brilliant results without ever going near an ICC workflow.  If you do ever use or need a different pipeline (certain media, particular images, a stylistic preference, repeatability on different hardware, and so on), then an ICC workflow can give you the same brilliant results across those variables which may not fit the vendor's pipeline.  That's why, after all, the vendors provide access to that pathway, too.

Neither is right or wrong, per se, but the ICC workflow approach, for most photogs provides a better workflow for the same excellent results.

[Doug Gray]

We've discussed this extensively before, most notably when C'tein over on TOP infamously recommended printer manages colors a couple of years back.  MHMG gave the definitive debunking of C'tein's findings and I expect that the same would apply to Bob's:

http://forum.luminous-landscape.com/index.php?topic=104307.msg857415#msg857415

When someone thinks they've discovered the holy grail, like C'tein and Bob, it's pointless to try to convince them otherwise.  The best we can do is to keep posting Mark's critique so that other's reading this thread have a more balanced view, and can see that printer manages colors is not the holy grail.

I've noticed on my Canon 9500 II and, to a lesser extent my Epson 9800, that letting the printer manage color and using the defaults, the rendered print's midrange luminance is boosted and some colors within the printer's gamut are increased in saturation.

If I just print a snapshot from my cellphone w/o any adjustment it often looks better on the Canon print than if I print using Perceptual with a custom I1Profiler generated profile using the Perceptual defaults. The Canon print is brighter and more colorful. Particularly in the greens. The Epson 9800, with the defaults, is also boosted but not nearly as much as the Canon. Both printers boost luminance less when using Photoshop manages color and their provided profiles but Canon still has significant boost as is allowed with Perceptual intent.

Needless to say, the same snapshot prints from the two printers look very different letting the printer manage color and using defaults.

However, using Relative intent and an image that is inside the printer gamuts, both printers produce visually identical prints on the same paper (outside of specular effects like bronzing) with Photoshop managing color and custom profiles.

I discussed this here.
http://forum.luminous-landscape.com/index.php?topic=107556.msg884963#msg884963

[Mark D Segal]

How are you able to say something is wrong unless you've tested it?

It just so happens that I have - extensively. See my article on this website titled "Printing Can be Fun and Easy".

This is not about what is "right" or "wrong". Nothing about this is absolutely and systematically right or wrong. It depends on the image, the paper, the printer and how much control you the user wants to be able to exercise over the fine details of the quality of the final output. Judging by what's on your website, this may be less critical for the kind of photos you make, because your colours are very strong, contrasts are stark, and the refinements of tonal gradations and subtle colour variations one can achieve with soft-proofing in a properly colour-managed workflow may not be critical for your kind of photography. But I wouldn't advise it for other people working with photographs that do benefit from refined image editing, and that would be the vast majority. Furthermore, I would only re-iterate the point that Wayne Fox made about the portability/consistency of results between printer models and papers, which is one of the main purposes of application colour management. But if that will never be important to you (hard to imagine, as technology evolves), then perhaps you in particular are not missing much with Printer Manages Color. But again, not advisable for the vast majority of others who may value this systematic consistency between media and equipment.

[Mark D Segal]

Higher S/N ratio: Adobe's algorithms are one step removed from the printer's.

Could you explain precisely what this sentence means? I'm baffled.

Just for beginners, please note that not all the "algorithms" deployed in a colour-managed workflow are either Adobe or Epson. There is something in your computer's operating system called a "CMM". Do you know what that means? Just in case you don't, let me help: it is the "Color Management Module". And within that CMM there is another processing module called the "PCS". Do you know what that means? Again, in case you don't, let me help. It is the Profile Connection Space. Do you now what that does? This time I'll let you look it up just in case you aren't fully tuned-in to the nature of these processes. A CMS (Color Management System) operates some functions in the application, some in the computer operating system and some in the printer. So yes, they are all "removed" from each other, but they are highly inter-dependent, and it would be wrong to assume that when you set your printer driver to Printer Manages Color that nothing is going on with the other two. Some processes still need to get the image file colour values from the file to the printer, and those processes matter to what the printer then does with them.

[Mark D Segal]

I didn't think it was so hard.

Yes, you can (and do) get excellent results from Printer Manages colour.  That's not new, but it is better with the newer hardware and drivers.  What it doesn't do is provide a workflow which can cater for multiple media, varied images and gamuts, different monitors and viewing conditions and so on.  It's a less flexible pipeline, but within its confines it's very good (even excellent).

The ICC workflow gives you a consistent, single workflow that caters for all the regular (and some irregular) variables and provides a means of controlling those variables in a predictable manner.  And it can achieve the same results.

So, printer manages is 100% OK if you know what you're doing and you like the results.  If you never step outside of the vendor prepared pipeline, it is entirely possible that you will have brilliant results without ever going near an ICC workflow.  If you do ever use or need a different pipeline (certain media, particular images, a stylistic preference, repeatability on different hardware, and so on), then an ICC workflow can give you the same brilliant results across those variables which may not fit the vendor's pipeline.  That's why, after all, the vendors provide access to that pathway, too.

Neither is right or wrong, per se, but the ICC workflow approach, for most photogs provides a better workflow for the same excellent results.

The results of my research confirm the general thrust of this comment. But I wouldn't go as far as to say "100% OK". May or may not be. This is where judgment enters the picture. In terms of colour accuracy, high quality custom profiles in an ICC-managed workflow are hard to beat, but the last degree of that accuracy may not be visible to the average viewer. One must also be aware of possible printer drift. As you know Phil, when these machines leave the factory they are calibrated to a common standard and they can be pretty stable, but not 100% over an indefinite time period. I would surmise that as the printers drift from the initial standard (depending on what ways), the comparative results of Printer Manages Color could vary.

[Mark D Segal]

I've noticed on my Canon 9500 II and, to a lesser extent my Epson 9800, that letting the printer manage color and using the defaults, the rendered print's midrange luminance is boosted and some colors within the printer's gamut are increased in saturation.

If I just print a snapshot from my cellphone w/o any adjustment it often looks better on the Canon print than if I print using Perceptual with a custom I1Profiler generated profile using the Perceptual defaults. The Canon print is brighter and more colorful. Particularly in the greens. The Epson 9800, with the defaults, is also boosted but not nearly as much as the Canon. Both printers boost luminance less when using Photoshop manages color and their provided profiles but Canon still has significant boost as is allowed with Perceptual intent.

Needless to say, the same snapshot prints from the two printers look very different letting the printer manage color and using defaults.

However, using Relative intent and an image that is inside the printer gamuts, both printers produce visually identical prints on the same paper (outside of specular effects like bronzing) with Photoshop managing color and custom profiles.

I discussed this here.
http://forum.luminous-landscape.com/index.php?topic=107556.msg884963#msg884963

I agree with your observation that printer colour management *can* produce a denser outcome showing as more saturated colour, as I've observed this in some of my testing. Not sure how far to generalize that however, and obviously it is not always the preferred result. And yes, colour management does work - one should/does get very similar results using the same paper in different printers of roughly similar colour reproduction capability provided the profiling and colour management set-up are good for both. I'm working between three different printer models just now (Canon Pro-1000, Epson P800 and Epson P5000) and I've printed the same photos in all of them with my own custom profiles for the papers being tested. Honestly, unless I labelled them, the next morning at first glance, and even the second, I wouldn't know which piece of paper came from which printer. I agree this should not be expected to hold with Printer Color Management. But perhaps for BobDavid this is not a relevant consideration.

[Mark D Segal]

.... I have considerable academic and professional credentials. I designed and worked in a special effects studio, owned and operated a fine art reproduction business, taught college courses on color theory, worked on machine vision projects, have an MS in Visual Studies from MIT, and had a stint as a research fellow at MIT related to visual sensation and perception.

Yes, I saw your credentials on your website - very impressive and respectable. But there remains room for honest and open debate between qualified people. The world would be a boring place if it were full of PhD's who agree with each other about everything. Let us not personalize the discussion.

[BobDavid]

Thank you Farmer... '... What it doesn't do is provide a workflow which can cater for multiple media, varied images and gamuts, different monitors and viewing conditions and so on.  It's a less flexible pipeline, but within its confines it's very good (even excellent).'

As I clearly stated at the start of this thread, an Epson SureColor SP800 printer does an excellent job of managing output on Epson Ultra Premium Luster media. I never dismissed best practices re color management. I am however mighty impressed that a prosumer printer is able to manage output on one particular paper. One.

I get what it takes to optimize files for SWOP and fussy fine art media.  Color management is not trivial, especially when a good part of one's living depends on running a tight ship.

I am amazed this thread has gone on as long as it has. If my initial statement came off as naive, I apologize.

[Alan Goldhammer]

One must also be aware of possible printer drift. As you know Phil, when these machines leave the factory they are calibrated to a common standard and they can be pretty stable, but not 100% over an indefinite time period. I would surmise that as the printers drift from the initial standard (depending on what ways), the comparative results of Printer Manages Color could vary.

I was going to mention this as well but how many of us do regular printer calibrations to measure drift?  I think the newer Canon and Epson prosumer printers have senors that allow one to do this (there was another thread on the Epson printers several weeks ago).  For my 3880 I have only two solutions, Epson ColorBase software which really only works with Epson papers and is very old software (first reviewed by Michael Reichman about a decade ago) or ArgyllCMS which does have tools to do calibrations. 

[digitaldog]

I never ever made a sweeping generalization about letting printers manage color.

No, just this wrong, made up, theory about color management:

• No need to soft proof, the monitor is setup for aRGB.
• Try "absolute
• Rendering intents are disabled when the printer manages output.
• Higher S/N ratio: Adobe's algorithms are one step removed from the printer's
• Truth is, I think the Epson algorithms are superior to Photoshop's. (Truth: You Bob of course didn't examine (understand?) the source code from those two companies).
• I'm curious to hear about results rather than theory.

Do you realize that the text above is simply WRONG yet and that's why you've been called out?

[digitaldog]

I have considerable academic and professional credentials.

None of which appear to aid you in backing up anything you've written here about color management and illustrate you are incapable of proving what you've written after being called out to do so!

Look, I have a degree too, but it isn't in medicine and I find it a massive mistake to argue with doctors on forums dedicated to medicine. You might want to examine your text as even particularly factual before posting.

[Ferp]

I never ever made a sweeping generalization about letting printers manage color. I began this thread with one specific observation:  An Epson SureColor printer is able to manage output on Epson Ultra Premium Luster Photo paper. That's all. Holy grails do not exist in the world of color management. There is however a balance between art and science.

That's not what your original post said:  "The P800 has exceeded my expectations: It makes stunning prints. The coolest thing about the printer is that I let it manage output rather than Photoshop. Who would have ever thunk that that would work. Truth is, I think the Epson algorithms are superior to Photoshop's."  No qualifications about paper choice.

I didn't think it was so hard.

Yes, you can (and do) get excellent results from Printer Manages colour.  That's not new, but it is better with the newer hardware and drivers.  What it doesn't do is provide a workflow which can cater for multiple media, varied images and gamuts, different monitors and viewing conditions and so on.  It's a less flexible pipeline, but within its confines it's very good (even excellent).

The ICC workflow gives you a consistent, single workflow that caters for all the regular (and some irregular) variables and provides a means of controlling those variables in a predictable manner.  And it can achieve the same results.

So, printer manages is 100% OK if you know what you're doing and you like the results.  If you never step outside of the vendor prepared pipeline, it is entirely possible that you will have brilliant results without ever going near an ICC workflow.  If you do ever use or need a different pipeline (certain media, particular images, a stylistic preference, repeatability on different hardware, and so on), then an ICC workflow can give you the same brilliant results across those variables which may not fit the vendor's pipeline.  That's why, after all, the vendors provide access to that pathway, too.

Neither is right or wrong, per se, but the ICC workflow approach, for most photogs provides a better workflow for the same excellent results.

Well said.  Pretty much a summary of what MHMG said in response to C'tein (see above link). 

[digitaldog]

And please refrain from condescending remarks such as "this silliness of your inability..."

You're using the wrong tool for the wrong job, better?

[Doug Gray]

I want to thank BobDavid for starting this thread with his observations. It's a most interesting thread, for it is consistent with things I've noticed from my own experience as well as digesting comments from others here and elsewhere.

Having developed tools for analyzing ICC profile and printer performance I have some ideas on how to characterize more exactly how "printer manages color" performs. I strongly believe that, for obvious commercial reasons, there are material color shifts made to accommodate expectations of the larger market. These shifts are, I suspect, typically much stronger than those built in to the Perceptual table generation of I1Profiler. The ICC profile Perceptual Intent is designed for viewing under 500 Lux with a specific print media dynamic range, But the larger market is likely to compare printers under lower illumination levels and printer manufacturers are more driven to satisfying the larger consumer base, at whatever level of sophistication, than adhering to ICC requirements, let alone mere ICC suggestions. And Perceptual Intent mapping is a mere suggestion.

I am working on a set of tools to quantitatively characterize "printer manages color."  Ideally, printer reviews would explore this untouched but significant area. Expect a specific thread on this topic in the future.

[digitaldog]

I want to thank BobDavid for starting this thread with his observations.

Including these observations?

No need to soft proof, the monitor is setup for aRGB.
Rendering intents are disabled when the printer manages output. And then: Try "absolute" (begging the question about disabled RIs).
Higher S/N ratio: Adobe's algorithms are one step removed from the printer's. (still a total mystery of a comment despite requests for explanation).
Truth is, I think the Epson algorithms are superior to Photoshop's. (Where do you suppose Bob got the source code for these two products let alone as Mark suggests, the CMM and other OS attributes).
I'm curious to hear about results rather than theory. (So far, all theory, no results and no facts). I see the irony of that last comment, does Bob? More importantly his audience!

[digitaldog]

Yet what you state about soft proofing makes no written or more importantly, any visual sense. Can you explain that or not?
I HAVE provided examples of that what you've written not being the case; do tell us what I did wrong.

He can't. But I'll still wait but without holding my breath. I'm curious to hear about results rather than (made up?) theory 


Verify, THEN trust.

[Doug Gray]

Including these observations?

No need to soft proof, the monitor is setup for aRGB.
Rendering intents are disabled when the printer manages output. And then: Try "absolute" (begging the question about disabled RIs).
Higher S/N ratio: Adobe's algorithms are one step removed from the printer's. (still a total mystery of a comment despite requests for explanation).
Truth is, I think the Epson algorithms are superior to Photoshop's. (Where do you suppose Bob got the source code for these two products let alone as Mark suggests, the CMM and other OS attributes).
I'm curious to hear about results rather than theory. (So far, all theory, no results and no facts). I see the irony of that last comment, does Bob? More importantly his audience!

They are quite odd statements from someone with the OP's stated background. And it is particularly odd given that he has done repro work professionally which is highly quantitative. One doesn't do repro and ask "how attractive it is" but how close it matches the original side by side under the same lighting and physical size/surround.

However, that said, his observation of a "stunning" looking print is clearly perceptual. It's due to whatever mapping or "secret sauce" the printer driver is doing. But it isn't magic. An ICC perceptual table that matched the device driver would produce equally "stunning" results on most printers since physical gamut limitation variation amongst printers are relatively insignificant even if hyped in marketing.

So it is this variation in device driver color mapping vendors are doing I find intriguing. So I'm pleased he brought up the topic. It is one that needs thorough exploring and understanding. I think the markets my be bifurcating between the graphic design folks where consistency and accurate color are baseline requirements and photography/art where pleasing appearance, even at the cost of vendor and model variance, is paramount.

[digitaldog]

They are quite odd statements from someone with the OP's stated background.

You are much more kind than I am 
Odd yes. Mostly wrong too.