Print Service ICC Profiles

[Osprey]

I'm going to try out a new print service bureau that supplies ICC printer paper profiles.  I asked them if I should just soft proof in the profiles that I am going to use and then submit the file in the sRGB, profile and they said that was the case.  Is that normal for a print service bureau?  I ask because I have been using an Epson P800 printer I have access to and have been soft proofing and printing using the printer paper ICC profile and the prints have been coming out well, very close to the soft proofing, so I'm wondering.

Thanks

[Mark D Segal]

I'm going to try out a new print service bureau that supplies ICC printer paper profiles.  I asked them if I should just soft proof in the profiles that I am going to use and then submit the file in the sRGB, profile and they said that was the case.  Is that normal for a print service bureau?  I ask because I have been using an Epson P800 printer I have access to and have been soft proofing and printing using the printer paper ICC profile and the prints have been coming out well, very close to the soft proofing, so I'm wondering.

Thanks

If they are making regular inkjet prints for you, and if you are submitting 16-bit TIFF or PSD files to them, they should be working in a wide space such as Pro Photo to be sure that the full gamut of your file is captured in print to the extent possible. If you are submitting 8-bit JPEGs for printing, then the use of ARGB(98) may be a safer colour space to minimize the risk of banding. For many photos you may find that prints confined to sRGB may come back and disappoint you.

[howardm]

I'm surprised that they dont ask you to simply convert to their profile and then attach a note that basically says 'dont touch this, print as-is'.

[Mark D Segal]

I'm surprised that they dont ask you to simply convert to their profile and then attach a note that basically says 'dont touch this, print as-is'.

I think the OP has uncovered a legitimate issue that should be addressed specifically about the appropriate colour space in which to prepare and print his files using the services of that bureau. At least that is how I interpreted his question.

[digitaldog]

I'm going to try out a new print service bureau that supplies ICC printer paper profiles.  I asked them if I should just soft proof in the profiles that I am going to use and then submit the file in the sRGB, profile and they said that was the case.  Is that normal for a print service bureau? 

For many print 'services' yes and it's a rather worthless process! If you can't fully use the ICC profile, meaning soft proof to pick the rendering intent you desire, convert and get that on output, it's a waste of time. A profile provided and forced to provide sRGB; a clueless color management approach they recommend. They should either fully support an ICC workflow (you convert as you desire, they output the numbers 'as is') or just demand sRGB and provide NO profile. I'm fine with both. I'm not fine with a lab playing color management games like this; supplying an ICC profile that may or may not define the actual output conditions, allowing you to pick a rendering intent and producing that via conversion of an actual ICC profile provided.

[stcstc31]

It also could be the tech they are using, if its mini lap type system lots of them are not really capable of much more

[Mark D Segal]

It also could be the tech they are using, if its mini lap type system lots of them are not really capable of much more

Good thought and also something the OP should be aware of. As he is using an Epson P800 and may be using proper colour management procedures as Andrew proposed above, he could end up being very disappointed by what comes back from the service bureau.

[Doug Gray]

I think the OP has uncovered a legitimate issue that should be addressed specifically about the appropriate colour space in which to prepare and print his files using the services of that bureau. At least that is how I interpreted his question.

As much as Andrew's position that printer services should provide complete icc profiling which would allow selection of rendering intent the fact is many do not and only take sRGB images. Some of these printers will provide an icc profile for soft proofing that accurately models how their printer renders colors including expansion of sRGB colors to printable ones outside of sRGB. Interestingly, this can allow accurate soft proofing within the limits of perceptual only mapping. There are even benefits to using a wide gamut monitor to do a better job of proofing when the profile does such an expansion since a small gamut (sRGB more or less) monitor will  clip the expanded colors. When this occurs one can get soft proofing colors beyond sRGB even when the workspace and image are in sRGB.

Not my cup of tea though. I prefer a full ICC workflow.

[digitaldog]

As much as Andrew's position that printer services should provide complete icc profiling which would allow selection of rendering intent the fact is many do not and only take sRGB images.

My position (again is), supply an ICC profile and let the customer fully use it as designed OR don't provide an ICC Profile and demand sRGB. Very simple. The fact shops only take sRGB isn't based on providing their customers the best possible output or workflow. It's a limitation they place on their customers for their benefit. But I'm OK with that; just don't pass off an ICC profile that's basically worthless while trying to convince the customer the lab is implementing a full/proper color managed workflow. They are not.

Some of these printers will provide an icc profile for soft proofing that accurately models how their printer renders colors including expansion of sRGB colors to printable ones outside of sRGB.

Please define this accuracy with a metric that's not totally ambiguous; use dE. Otherwise, it's a huge generalization.

Interestingly, this can allow accurate soft proofing within the limits of perceptual only mapping.

Ditto my comment above.


Delta-E and color accuracy

In this 7 minute video I'll cover: What is Delta-E and how we use it to evaluate color differences. Color Accuracy: what it really means, how we measure it using ColorThink Pro and BableColor CT&A. This is an edited subset of a video covering RGB working spaces from raw data (sRGB urban legend Part 1).

[Doug Gray]

My position (again is), supply an ICC profile and let the customer fully use it as designed OR don't provide an ICC Profile and demand sRGB. Very simple. The fact shops only take sRGB isn't based on providing their customers the best possible output or workflow. It's a limitation they place on their customers for their benefit. But I'm OK with that

I agree with all of that. It allows them to work with the LCD at the expense of not meeting the needs of advanced (in the sense of understanding and controlling their color) photographers.

just don't pass off an ICC profile that's basically worthless while trying to convince the customer the lab is implementing a full/proper color managed workflow. They are not.

It certainly isn't a full color managed workflow but worthless? Not if done correctly.

Please define this accuracy with a metric that's not totally ambiguous; use dE. Otherwise, it's a huge generalization.Ditto my comment above.

An "accurate" profile for working with a printer that only produced perceptual rendering and takes sRGB only images is one that provides profiles where the BtoA0->AtoB1 conversions produce the color that will actually be printed. This is, of course, what soft proofing does behind the scenes. Precision, or lack thereof, in a dE statistical sense, is a function of the tools, printer, and paper they use just as it is for a normal, fully color managed workflow profile.

As an example specific to sRGB, take the color Lab=(50,-30,-9), a Cyan near the sRGB gamut edge. if the printer's profile maps this into a more saturated Cyan the profile proofing conversion, BtoA0->AtoB1, might well return Lab(52, -35, -20). This is a more saturated Cyan than can be shown on an sRGB limited monitor but within what can be shown on a wide gamut monitor. The the soft proof will match the print even though both the soft proof color and the print color are outside of sRGB.

Also, even on a large gamut inkjet, about as many printable colors are outside sRGB as there are unprintable colors that are within sRGB. Commercial printing tends to have smaller gamuts and have more unprintable colors that are within sRGB. So even if the print shop doesn't remap sRGB colors to a wider print gamut, there is a significant benefit to a proofing profile as it will properly proof  the sRGB colors that are not printable.

[digitaldog]

It certainly isn't a full color managed workflow but worthless? Not if done correctly.

Customer is provided a profile that may or may not define the output conditions, may or may not be used for actual conversions by the lab, customer can't control RI of output. Customer probably has to guess which RI is forced upon the conversion by the lab (will they tell the customer?). So yeah, pretty much useless. Other approach: customer gets profile that does define output conditions and is used for conversions because customer produces the actual conversion using the RI he/she visually prefers. Customer can edit post conversion if necessary.
So, what does the user do 'correctly' that overcomes these issues? How does the customer even know the profile supplied is ever used for output?

An "accurate" profile for working with a printer that only produced perceptual rendering and takes sRGB only images is one that provides profiles where the BtoA0->AtoB1 conversions produce the color that will actually be printed. This is, of course, what soft proofing does behind the scenes. Precision, or lack thereof, in a dE statistical sense, is a function of the tools, printer, and paper they use just as it is for a normal, fully color managed workflow profile.

Without a dE metric, the term accuracy is marketing talk and ambiguous! Not helpful here, very helpful if you're marketing something to people who hear the term 'accuracy' and assume it means something.
Example: Doug is tall. Useless. Doug is 6"3. A measurement, a metric of height; zero ambiguity. We can argue now if we agree that 6"3 is tall (no, 7"1 is tall). But we have a value that defines height and the rest are opinions at least based on each person's understanding specifics we're discussing: is Doug tall? Using accuracy without a similar language is about as worthless as that profile the lab provides but doesn't allow their customers to use!

Also, even on a large gamut inkjet, about as many printable colors are outside sRGB as there are unprintable colors that are within sRGB. Commercial printing tends to have smaller gamuts and have more unprintable colors that are within sRGB. So even if the print shop doesn't remap sRGB colors to a wider print gamut, there is a significant benefit to a proofing profile as it will properly proof  the sRGB colors that are not printable.

So yeah, sRGB is suboptimal for output to anything but the web and mobile devices, and accuracy has zero meaning in the above context. Why use it?

[Doug Gray]

Customer is provided a profile that may or may not define the output conditions, may or may not be used for actual conversions by the lab, customer can't control RI of output. Customer probably has to guess which RI is forced upon the conversion by the lab (will they tell the customer?). So yeah, pretty much useless. Other approach: customer gets profile that does define output conditions and is used for conversions because customer produces the actual conversion using the RI he/she visually prefers. Customer can edit post conversion if necessary.
So, what does the user do 'correctly' that overcomes these issues? How does the customer even know the profile supplied is ever used for output?

Yep, the customers don't have a choice and presumably the profile would render perceptual only. Ideally the print service profile would use the same tables for BtoA1, and BtoA2 as for BtoA0 so the customer couldn't mess up the soft proofing. One size fits all. Very Procrustean.

Without a dE metric, the term accuracy is marketing talk and ambiguous! Not helpful here, very helpful if you're marketing something to people who hear the term 'accuracy' and assume it means something.
Example: Doug is tall. Useless. Doug is 6"3. A measurement, a metric of height; zero ambiguity. We can argue now if we agree that 6"3 is tall (no, 7"1 is tall). But we have a value that defines height and the rest are opinions at least based on each person's understanding specifics we're discussing: is Doug tall? Using accuracy without a similar language is about as worthless as that profile the lab provides but doesn't allow their customers to use!

Huh? The profile is provided for soft proofing. Should be just as good at that as any, full color managed, profile used for soft proofing. On a broader note, I also agree that there is a serious lack of specific color and color error terminology in the areas of soft proofing. Much of it because too few people set up their proofing workflow very well to start with. Lots of guesswork and "looks good enough."  You have a lot of good videos and instructional material that addresses it pretty well but some will not take the time. I too prefer measurements and statistics. To each their own.

One area that I find completely lacking is a standardized approach for producing a printer's "smoothness" metric caused by localized, rapid gradient changes in printed color v inputs. I find this much more problematic than dE76 or dE2k measurements over a limited sample set. You can have a perfectly great printer with in gamut accuracy averaging less than 1 dE2k on a few thousand patches that has serious banding errors on a smooth sunset sky due to localized, large, gradient color changes. Proofing doesn't often reveal these whether it's the limited sRGB proofing or that of a normal workflow.

You get what you measure but good luck on what you don't measure.

So yeah, sRGB is suboptimal for output to anything but the web and mobile devices, and accuracy has zero meaning in the above context. Why use it?

Just because half the printable colors are thrown away because they are outside sRGB to start with there is no reason to not deal with the other half of the sRGB colors that can't be printed. Soft proofing lets you deal with those.

Further, in the event the commercial printer maps sRGB colors to more saturated ones in areas their printer gamut allows, using a profile for soft proofing is the only way to see what the resulting print will look like other than just sending them out and hoping.

[Rand47]

just don't pass off an ICC profile that's basically worthless while trying to convince the customer the lab is implementing a full/proper color managed workflow. They are not.

It's actually worse than what has been stated so far.  In doing research for a Color Management seminar I held for a local photo club, I did research on a bunch of online labs, because my questionnaire for the students indicated a high interest in preparing/soft-proofing files to be sent to labs for printing.

After emailing, live chatting, and phone calling - the range of answers from Costco to some very high priced labs resulted in this range of results:

There are a bunch of labs who will tell you to send them any file format, any color space, etc.  They infer that they are fully color managed.  When it got down to it, the "front end" of their process converts everything they receive into sRGB jpegs for printing.

One lab proudly offers a single ICC profile for use with any of their paper selections.

Some want the file with the provided ICC profile used for soft proofing.

Other labs clearly tell you to soft proof to the ICC profile but to send them the soft proofed file in ProPhoto or Adobe RGB.

What I ended up telling the class is to do their homework with any lab, not to believe what the web site says, and that like John Arbuckle, "You get what you pay for."

Interestingly enough, Costco provides ICC profiles for each of their paper/printer combinations, and it is "by store location" specific and they are updated as needed.

There certainly wasn't "one answer" to give the class on how to prepare files for outside printing!

Rand

[digitaldog]

Yep, the customers don't have a choice and presumably the profile would render perceptual only.

How's that accomplished? Why force that RI if even possible? Color management kludge!

Ideally the print service profile would use the same tables for BtoA1, and BtoA2 as for BtoA0 so the customer couldn't mess up the soft proofing. One size fits all. Very Procrustean.

So what labs are doing this?

Huh? The profile is provided for soft proofing. Should be just as good at that as any, full color managed, profile used for soft proofing.

A soft proof from a profile not used to actually output the data is worthless. How does the customer know the lab's is doing this? Or that the output device conforms to the profile's prediction? They can't.

You get what you measure but good luck on what you don't measure.

We can't nor should talk about accuracy without those measurements!

Further, in the event the commercial printer maps sRGB colors to more saturated ones in areas their printer gamut allows, using a profile for soft proofing is the only way to see what the resulting print will look like other than just sending them out and hoping.

You're defining why the 'just send is sRGB' is such a silly workflow for the customer (not the printer)!
The bottom line is, in a sound and fully implemented color management workflow, we don't restrict ourselves to sRGB for conversions to print. We have good ICC Profiles that actually define output conditions. We have all the options for selecting a rendering intent based on how our images appear, via a soft proof with the various tables. The just send us sRGB and here's a profile for soft proofing doesn't provide that. So again, just send sRGB is fine for that audience and price point. Forget providing an ICC profile in such a case, it's only marketing to those who falsely believe the lab is implementing color management. Or better, send an actual ICC profile to the customer and let them use it as it was designed for. A half baked approach is as I've said, worthless and kind of a big fat lie.

[digitaldog]

Interestingly enough, Costco provides ICC profiles for each of their paper/printer combinations, and it is "by store location" specific and they are updated as needed.

Thanks in part to Drycreek. That's the way it should be done. Shocking that Costco provides a full color managed workflow while so many others, those that do charge more, don't. Plus, Costco proves that it is possible to implement a full, sound color managed workflow for customers all over the country; proof of concept. Yet one shop 'catering' to pro's can't. Very sad.

[Doug Gray]

A soft proof from a profile not used to actually output the data is worthless. How does the customer know the lab's is doing this? Or that the output device conforms to the profile's prediction? They can't.

You are making two assumptions. First that the profile they supply is not used for making their prints. Second, that if it isn't the profile it is worthless. The first assumption may or may not be the case but the second assumption is simply wrong.

It doesn't matter whether the profile is used for making prints. Customers have no interest in what the profile outputs to the services' printers actually are. They are not used by the customer for anything. The customer only cares about how well the profile renders the actual colors requested. Over time printers change and profiling them regularly allows the printers to maintain consistent color. It's important for commercial printers to profile their equipment on a regular basis. And these changes can result on significant shifts in the BtoA and AtoB conversions. But guess what? When you go roundtrip from BtoA0 then to AtoB1, as is done behind the scenes for soft proofing, the changes from age and wear almost totally cancel and the residuals are tiny. With one caveat. You can get big shifts when changing profiling software or settings but if you keep these the same there is no need to soft proof with current profiles. There is only a need for the printer to use current profiles.

[Doug Gray]

Thanks in part to Drycreek. That's the way it should be done. Shocking that Costco provides a full color managed workflow while so many others, those that do charge more, don't. Plus, Costco proves that it is possible to implement a full, sound color managed workflow for customers all over the country; proof of concept. Yet one shop 'catering' to pro's can't. Very sad.

Costco does it best. But do we know how accurately they do it? Are the profiles current? Same questions for commercial printers like WHCC. Does anyone ever check their accuracy? The proofing profiles they offer are from last decade and use PM5. It's still a perfectly good s/w package but if I used any of them I would sure include a set of patches in an image run from time to time and measure them.

[digitaldog]

You are making two assumptions.

Actually no, I'm asking questions (that's what that character that looks like this:? at the end of many of my sentences mean!
The person making all the assumptions is you Doug.

First that the profile they supply is not used for making their prints.

It might be, it might not be. How do we know? IF NOT, that profile is useless.

Second, that if it isn't the profile it is worthless.

IF the profile doesn't define the process, it's worthless. Try making a print on an Epson using a profile for a Canon. Good luck.

It doesn't matter whether the profile is used for making prints.

Sure it does! Or to put it this way, the profile needs to define the process.

Customers have no interest in what the profile outputs to the services' printers actually are.

Speak for yourself.

Over time printers change and profiling them regularly allows the printers to maintain consistent color.

And that change needs to be reflected in the profile the customer is provided. Or it's worthless.

It's important for commercial printers to profile their equipment on a regular basis.

Nonsense! As someone who's built digital press profiles for some very big companies over the years, what commercial printers need to do is produce consistent output which the profile defines.
But lots of printers can't handle sufficient process control so they let it go all over the map and reprofile. OK, did they supply that new profile to the customer each and every time?

There is only a need for the printer to use current profiles.

Make up your minds!

[digitaldog]

Costco does it best. But do we know how accurately they do it?

Yes! Do you need Ethan Hansen's email address?

[Mark D Segal]

You are making two assumptions. First that the profile they supply is not used for making their prints. Second, that if it isn't the profile it is worthless. The first assumption may or may not be the case but the second assumption is simply wrong.

It doesn't matter whether the profile is used for making prints. Customers have no interest in what the profile outputs to the services' printers actually are. They are not used by the customer for anything. The customer only cares about how well the profile renders the actual colors requested. Over time printers change and profiling them regularly allows the printers to maintain consistent color. It's important for commercial printers to profile their equipment on a regular basis. And these changes can result on significant shifts in the BtoA and AtoB conversions. But guess what? When you go roundtrip from BtoA0 then to AtoB1, as is done behind the scenes for soft proofing, the changes from age and wear almost totally cancel and the residuals are tiny. With one caveat. You can get big shifts when changing profiling software or settings but if you keep these the same there is no need to soft proof with current profiles. There is only a need for the printer to use current profiles.

This makes no sense to me. If printers need to reprofile periodically because of printer drift, each profile will have somewhat different transforms to preserve colour accuracy. That being the cause, the softproof may also look different to the discerning eye from one profile to the next, depending on how radical the drift. Anyone hoping to receive a print with expected colours back from the printer is best advised to prepare the files under soft-proof using the profile that the printer will use for making the prints, otherwise all bets are off. Being consistent in the use of the printing profile for file preparation under soft-proof is the best assurance to avoid unwanted surprises.