One more reason why color management confuses the general user.

[Doug Gray]

Well you can, but it would be nice and useful (I do respect your opinions) about the lab demanding sRGB OR Adobe RGB (1998) yet providing profiles only for soft proofing. Makes no sense to me.

The least the lab could do is state that they print Adobe RGB or sRGB  tagged files using PerCol and that the profiles should be used in PerCol for proofing purposes. A lot of people only print in PerCol and work in Adobe RGB. They would be perfectly happy. At least if the gamut limitations are acceptable and lack of vendor print consistency is ok since there is no standard way of mapping colors for PerCol. Even inside the gamut. Like a box of chocolates.

[digitaldog]

The least the lab could do is state that they print Adobe RGB or sRGB  tagged files using PerCol and that the profiles should be used in PerCol for proofing purposes. A lot of people only print in PerCol and work in Adobe RGB.

IF I had a gun to my head, had to select one RI, it be RelCol. Luckily I don't have to be forced into that RI. We both know that the Perceptual RI differs from manufacturer/color engine and not all are created equally. Even Saturation RI is fair game. It's why we have those three options and should soft proof all images as profiles know nothing about images.


In an effective color management workflow, the working space and RI used in the output profile is not forced upon us.

They would be perfectly happy.

I  hear that all the time in such arguments and I think it's rather silly because:

1. You and other's cannot speak for others.
2. Other's who have no options don't know what they don't have or know how that might be a benefit. 


It's like suggesting those who only work in sRGB are happy. And yes, ignorance is bliss. They have no idea what their output could look like if given the chance so it's kind of silly to speak for them and their happiness based on no other experience. Yes, an sRGB workflow will not by itself, produce poor output. But optimal? No, that's simply not the case.

At least if the gamut limitations are acceptable and lack of vendor print consistency is ok since there is no standard way of mapping colors for PerCol.

That's a load of assumptions. As for mapping to RelCol, again, not all profiles are created equally as we both know:

[Doug Gray]

IF I had a gun to my head, had to select one RI, it be RelCol. Luckily I don't have to be forced into that RI. We both know that the Perceptual RI differs from manufacturer/color engine and not all are created equally. Even Saturation RI is fair game. It's why we have those three options and should soft proof all images as profiles know nothing about images.


In an effective color management workflow, the working space and RI used in the output profile is not forced upon us.  I  hear that all the time in such arguments and I think it's rather silly because:

1. You and other's cannot speak for others.
2. Other's who have no options don't know what they don't have or know how that might be a benefit. 


It's like suggesting those who only work in sRGB are happy. And yes, ignorance is bliss. They have no idea what their output could look like if given the chance so it's kind of silly to speak for them and their happiness based on no other experience. Yes, an sRGB workflow will not by itself, produce poor output. But optimal? No, that's simply not the case. That's a load of assumptions. As for mapping to RelCol, again, not all profiles are created equally as we both know:

Well, I can speak for those that always print in PerCol which includes those that have no idea what a profile is or how to soft proof.  At least they provide the printer profile for proofing and whatever way it happens to map colors, in or out of gamut since it is PerCol, the proof should be an accurate representation of the prints they get back. That's better than 80% of people that send their stuff out and don't use, or know how to use, profiled printers.

It isn't my cup of tea. I also prefer RelCol and like to keep the edited image colors inside the printer gamut. That way I can be certain my prints will come back looking the same no matter what print vendor I use. That pretty much means they have to take images that have been converted using their profile's Rel Col so this place wouldn't work for me.

[digitaldog]

Well, I can speak for those that always print in PerCol which includes those that have no idea what a profile is or how to soft proof.

How can you speak for them when as you admit, they always print RelCol and have no idea how to toggle to other RI's and soft proof?
Do you feel a one size fits all RI for something as complex as images is the best approach? Begging the question, why did the ICC provide other RI's and the ability for each manufacturer to create their own Perceptual Rendering? Should transparency film always have appeared like Ektachrome and should photographers who never had the ability to shoot say Velvia be told they are missing nothing by being forced into an Ektachrome rendering? Seems rather silly to suggest that.
People who don't know what they don't know, are often happy with what they have. As I said, ignorance is bliss.

That way I can be certain my prints will come back looking the same no matter what print vendor I use.

Really? Then how do you explain this photo of an actual print made using two profiles with RelCol to the same printer and paper (let alone differing print shops):

[Wayne Fox]

Under no circumstances should you convert to our printing profile or embed it in your files.

A customer is provided a profile for soft proofing they can't use to convert. That's silly. It's not nitpicking. Do you believe that all rendering intents are the same, that image A may output better with RI A vs. RI B or even C? Or that an ICC profile know anything about color in context? What RI does the lab force on their customers and which are they supposed to use to soft proof?

While I agree it might be helpful for advanced photographers to use the profiles and convert the image themselves so they have control over the rendering intent, at the volumes these labs are producing, to try and provide an alternative would be detrimental to production, and from a practical point of view be used by a tiny fraction of the photographers submitting work.

The color management workflow is built into the firmware/software of the devices, and the conversion to the paper profile is actually handled by the printer, as the profile is normally attached to the media choice in the printer.   Most devices can handle only 8 bit image files, which means submitting in ppRGB might be problematic.  But most can easily handle submission in either sRGB or aRGB.  To handle files without a color space mosts labs will assume sRGB. 

Using a Chromira printer for example, it is actually pretty difficult to submit an image to the printer unmanaged, which would be what is necessary to use a file converted to the printer profile.  Every time I need to print a target for profiling to my Chromira printer I have to jump through some hoops to keep the printer from applying the current profile because the output conversion is handled by the printer, not the submission software, and the profile is attached to the media choice.  If I don't attach a profile to the media, the printer will error out when it tries to print the file.

so while I'm not disagreeing with the concept, the practicality of it from a lab perspective is difficult, especially considering the volume of work each printer must produce to keep the lab in business, and that only a minuscule percentage of clients are even aware of rendering intents and would take advantage of an alternate workflow which allowed this.

[digitaldog]

While I agree it might be helpful for advanced photographers to use the profiles and convert the image themselves so they have control over the rendering intent, at the volumes these labs are producing, to try and provide an alternative would be detrimental to production, and from a practical point of view be used by a tiny fraction of the photographers submitting work.

I guess that's why the one lab I know provided a sound color management workflow to such devices and allowed any profile to be used went out of business. I'm surprised I wasn't blamed for that workflow 

Most devices can handle only 8 bit image files, which means submitting in ppRGB might be problematic.

Doubt it. Certainly based on the work done by the late Bruce Fraser for Kodak testing 8-bit per color workflows with ProPhoto RGB. Of course, one could covert high bit, ProPhoto RGB to the printer profile and then send 8-bit per color to the lab (ideal for transmission anyway) IF the lab allowed it. They don't so it's moot.

To handle files without a color space mosts labs will assume sRGB. 

That's the bottom line here; it's what they assume and demand. It's intended for their benefit, not that of their customers who understand and wish to implement sound, full color management practices.

Using a Chromira printer for example, it is actually pretty difficult to submit an image to the printer unmanaged, which would be what is necessary to use a file converted to the printer profile.

Difficult (for whom) or impossible? Suppose the lab applied an extra fee? The customer could decide if it's worth the fee or they don't. No such option is provided from these so called "Pro Labs."
Here's where I stand on this: Either demand sRGB or perhaps Adobe RGB and be done OR implement a full color management workflow. The labs are using the later to imply they have their color management act together for their customers, supply profiles for a rather useless practice (soft proofing only) and can't provide proof the profile defines the output conditions. It would be very easy to test this too!

I have no issue with a lab saying one or two working spaces must be supplied. I have an issue with them providing a bogus ICC profile the customer can't use.

Every time I need to print a target for profiling to my Chromira printer I have to jump through some hoops to keep the printer from applying the current profile because the output conversion is handled by the printer, not the submission software, and the profile is attached to the media choice.

  I'm sorry to hear it's such a hassle. Like good process control, it's part of the service business. I can assure you, it's far, far more difficult and expensive to make a room filled with Indigo presses do this and yet, it's done by those who demand superb process control and a color managed workflow.

so while I'm not disagreeing with the concept, the practicality of it from a lab perspective is difficult, especially considering the volume of work each printer must produce to keep the lab in business, and that only a minuscule percentage of clients are even aware of rendering intents and would take advantage of an alternate workflow which allowed this.

You've queried the customer base (yours or the lab in question) and you're certain the so called Pro's wouldn’t be willing to pay an additional fee? I guess in the end, those that really do need this functionality just bypass said labs and produce their own prints. 

[Peter_DL]

WHCC
>>…you can download ICC profiles for soft proofing purposes.
… Under no circumstances should you convert to our printing profile or embed it in your files.<<

For example with a Fuji Frontier Printer -- (its working modes were excellently explained in this earlier post by Czornyj) -- it could potentially be that a lab chooses to ignore the PD mode, and to kept the printer in the so called sRGB mode.  When receiving an image in a different working space, they may have found their way to convert to sRGB first, before feeding the RGB data into the printer driver.

I can't say how likely this scenario is here with WHCC, however, in this case the offered icc profile would refer to the printer in sRGB mode, in the ideal situation describing the de facto printed gamut (deviating from sRGB).
In this broader context, the guideline: "soft-proof but do not convert" was actually stated in a former paper by Fuji, which however does not seem to be around anymore. That's not correct color-management, I'll leave it to others to point this out, but it may not be totally silly. We touched on this topic here.

--

[Doug Gray]

For example with a Fuji Frontier Printer -- (its working modes were excellently explained in this earlier post by Czornyj) -- it could potentially be that a lab chooses to ignore the PD mode, and to kept the printer in the so called sRGB mode.  When receiving an image in a different working space, they may have found their way to convert to sRGB first, before feeding the RGB data into the printer driver.

I can't say how likely this scenario is here with WHCC, however, in this case the offered icc profile would refer to the printer in sRGB mode, in the ideal situation describing the de facto printed gamut (deviating from sRGB).
In this broader context, the guideline: "soft-proof but do not convert" was actually stated in a former paper by Fuji, which however does not seem to be around anymore. That's not correct color-management, I'll leave it to others to point this out, but it may not be totally silly. We touched on this topic here.

--

Fact is in the large majority of photos most of the gamut clipping occurs in the conversion to printer space. Even converting to sRGB, the gamut reduction is small compared to going from sRGB to printer space. Being that printing services are market based, this appears acceptable to most or the market would support converting to the printer's space using the customer's preferred RI. The market supplies what people demand. It's a simple as that.

[digitaldog]

Even converting to sRGB, the gamut reduction is small compared to going from sRGB to printer space.

Converting to sRGB from what working space and how (with limited RI)?

Being that printing services are market based, this appears acceptable to most or the market would support converting to the printer's space using the customer's preferred RI.

How can something be acceptable or unacceptable when those people have neither the options to try something different or observe the differences? I keep asking how such assumptions about the so called market carries weight and as yet, no one has provided any facts based on real evidence.

The market supplies what people demand. It's a simple as that.

People can't demand what they don't know exists.


But look, the bottom is this: If you demand an sRGB workflow to your printer, fine. Do that and be done. If you're going to demand an sRGB document for output, don't try to fool your customers by providing profiles for soft proofing that they can't use and may not have any relationship to the actual output. Be honest. Again, for the last time, I've got zero issue with any lab that tells it's customers they have to submit sRGB for output. I have a big issue with labs that pretend, by providing an output profile the customer can't use, they are somehow color managed. It's just a lie.

[Doug Gray]

Converting to sRGB from what working space and how (with limited RI)?

Sticking with the printer's recommendation, Adobe RGB would be the larger gamut they recommend. Printers use PerCol when rendering RGB spaces by default. Presumably they would render whatever tagged RGB space the customer provides to the printer per PerCol and their printer profiles would accurately soft proof for that RI.

How can something be acceptable or unacceptable when those people have neither the options to try something different or observe the differences? I keep asking how such assumptions about the so called market carries weight and as yet, no one has provided any facts based on real evidence. People can't demand what they don't know exists.

The market is the evidence. There is no shortage of people talking about how small the sRGB gamut is and it is small. That doesn't mean it is optimum. Witness Betamax. Cheap and good enough often beats better.

But look, the bottom is this: If you demand an sRGB workflow to your printer, fine. Do that and be done. If you're going to demand an sRGB document for output, don't try to fool your customers by providing profiles for soft proofing that they can't use and may not have any relationship to the actual output. Be honest. Again, for the last time, I've got zero issue with any lab that tells it's customers they have to submit sRGB for output. I have a big issue with labs that pretend, by providing an output profile the customer can't use, they are somehow color managed. It's just a lie.

The printer suggests Adobe RGB or sRGB for those that don't know what Adobe RGB means.  Good advice. I think they hit a sweet spot in the larger market. Some of us artisans and color picky, knowledgeable folk go a different path.

I don't see much hope for change until high gamut video becomes a "must have."  In many ways it's more compelling and makes more of a difference than 4K over regular HD. Oled tech will drive that because backlit, high gamut LCD screens use much more power to get the large gamut.  Batteries are king in mobile devices.

[digitaldog]

The market is the evidence. There is no shortage of people talking about how small the sRGB gamut is and it is small. That doesn't mean it is optimum. Witness Betamax. Cheap and good enough often beats better.

But consumers had the options to actually try Betamax. The labs don't provide the same to consumers who want even try anything but sRGB!

The printer suggests Adobe RGB or sRGB for those that don't know what Adobe RGB means.

Some, few do. Most don't even give the choice of Adobe RGB (1998)! So do you suppose there might have been users who tried both sRGB and Adobe RGB at lab A and decided to stick with them because lab B only offered sRGB. And further, they actually tested both approaches and found they preferred Adobe RGB (1998)?

I don't see much hope for change until high gamut video becomes a "must have."

It's not a must have, it's an option. An option not provided. But again, if a lab demands one working space, with no options, and the price point and quality of output, and as importantly consistency, is good, I'm fine with that. I'm not fine with a lab pretending to practice modern color management providing an output profile that can't be used. AGAIN, that's just a marketing lie.

[Doug Gray]

But consumers had the options to actually try Betamax. The labs don't provide the same to consumers who want even try anything but sRGB.

Unlike video tape, consumers have long been able to use large gamut printers and high gamut workflows. And professionals often work in Adobe RGB today.

It's not a must have, it's an option. An option not provided. But again, if a lab demands one working space, with no options, and the price point and quality of output, and as importantly consistency, is good, I'm fine with that. I'm not fine with a lab pretending to practice modern color management providing an output profile that can't be used. AGAIN, that's just a marketing lie.

Consumers measure distance with a tape measure, angles with a protractor, or walk a property line and count steps.  Professionals use a total station.

Depends on what you need. There are tons of tools that skilled people use that others choose not to bother with. There is only so much time in one's life.

[digitaldog]

Unlike video tape, consumers have long been able to use large gamut printers and high gamut workflows.

To the lab's under discussion, just the opposite! Consumers did have access to differing video tape options. So what you're saying here appears to me to be just the opposite of reality. Sorry.

And professionals often work in Adobe RGB today.

Of course they might. What's your point? How did you gauge what a professional is and what color spaces they use? You can suggest professionals often work in Adobe RGB today and I can suggest they often work in ProPhoto RGB, especially those who use an Adobe raw converter. Again, so what? We can both speculate about the numbers. This has nothing to do with anything discussed here.

Consumers measure distance with a tape measure, angles with a protractor, or walk a property line and count steps.  Professionals use a total station.

More assumptions and speculations that are just generalizations.

"All generalizations are false, including this one".
-Mark Twain

The facts are pretty clear if you don't ignore them. Many so called 'pro labs' (whatever that is supposed to mean or who it's supposed to include) demand sRGB data for output. Fine. The facts are, when a lab demands sRGB for output, the customers have no other options. They have no way to see what the lab could produce with a different RGB working space. Again, it doesn't mean the output will suck. The facts are as clear as the nose on your face however; these customers don't have other options for an RGB working space out to the printers. So without such options, they are unqualified to state what they might have seen using another option.


I don’t know if you are purposely trying not to understand or accept these facts, or if you are really struggling with it but believe it's the former.

[Doug Gray]

The facts are pretty clear if you don't ignore them. Many so called 'pro labs' (whatever that is supposed to mean or who it's supposed to include) demand sRGB data for output. Fine. The facts are, when a lab demands sRGB for output, the customers have no other options. They have no way to see what the lab could produce with a different RGB working space. Again, it doesn't mean the output will suck. The facts are as clear as the nose on your face however; these customers don't have other options for an RGB working space out to the printers. So without such options, they are unqualified to state what they might have seen using another option.


I don’t know if you are purposely trying not to understand or accept these facts, or if you are really struggling with it but believe it's the former.

I have no idea why you believe I disagree with any of that. Obviously people that use these printers don't know (most of them, I suspect) or care (those that know) but are sufficiently happy with what they get that they don't bother doing their own printing or raising hell with the printers. Likely most of their customers don't have a clue.

[Doug Gray]

Then how do you explain this photo of an actual print made using two profiles with RelCol to the same printer and paper (let alone differing print shops):

I'm a little late answering this. I missed the blue ball question earlier. Obviously the balls are outside of the printer's gamut. By a rather huge amount (100 dEs or more?) given the rendering difference. Since there is no defined, or even agreed upon, way that out of gamut colors are mapped in RC, or any intent for that matter, one can't conclude from that alone that the profile is defective. The one that renders the ball black might be more accurate at rendering printable colors. Or not. No way to tell from that.

[digitaldog]

I'm a little late answering this. I missed the blue ball question earlier. Obviously the balls are outside of the printer's gamut. By a rather huge amount (100 dEs or more?) given the rendering difference. Since there is no defined, or even agreed upon, way that out of gamut colors are mapped in RC, or any intent for that matter, one can't conclude from that alone that the profile is defective. The one that renders the ball black might be more accurate at rendering printable colors. Or not. No way to tell from that.

Yet both prints (yes, actual prints photographed seen here) are made using RelCol. So how does this statement hold up:
AtoB1 tables are supposed to accurately colorimetrically represent the printed color. Always.

[Doug Gray]

Yet both prints (yes, actual prints photographed seen here) are made using RelCol. So how does this statement hold up:
AtoB1 tables are supposed to accurately colorimetrically represent the printed color. Always.

Pay attention. The printer profile's AtoB1 tables have nothing to do with the way the actual prints look.

Sure the blue ball renderings look very different. They look different because the two profiles rendered the prints using BtoA1. BtoA1 LUTs are only required to render colors that are outside the printer gamut. How they render these obviously varies with vendor.

EtoA: For clarification purposes because someone may interpret the statement above to mean that only the colors outside the printer's gamut are the colors rendered. The statement better reads:

BtoA1 LUTs are only required to render all colors. They are not required to map colors outside the printer's gamut in any defined way. How they render these obviously varies a lot between vendors and but doesn't affect the accuracy of the profiles within the meaning of ICC. For colors that are within the printer's gamut, BtoA1 LUTs should render the colors unchanged to the extent possible given the constraints imposed by less than perfect interpolation processes and data unevenness caused either by measurement noise or actual large printer response gradients.

[digitaldog]

They look different because the two profiles rendered the prints using BtoA1. BtoA1 LUTs are only required to render colors that are outside the printer gamut. How they render these obviously varies with vendor.

I've been trying to drum that into your head, in that other post for days!

[Doug Gray]

I've been trying to drum that into your head, in that other post for days!

I've never maintained that colors outside a printer's gamut are treated consistently by different profile vendors in RC (or any other intent for that matter). I presume you aren't suggesting that they should be mapped the same way and therefore are interpreting my posts differently from my intended meaning. If that's the case I'll try to be clearer.