sRGB vs. Adobe RGB: New color management stand up comic

[supercurio]

I'm going to be using a commercial inkjet printer from a pro lab.  Image shot in AdobeRGB and printed in AdobeRGB, sRGB printed in sRGB.  This is demonstrating how bad my workflow is?  Are you expecting that the average person will vote for the sRGB?  Or maybe it'll be a tie?

Tell me in advance what is bad about the workflow I'm going to use, so I can consider it for the video.

One more thing about this idea, if you're interested making things progress here.

If you know of professional equipment that has bugs in its implementation of color management, I'd say please report that to the manufacturer ASAP.
That's the responsible thing to do: an error in color management is actually something very severe especially for professional print equipment.

Then after a reasonable amount of time if they don't address the error or simply ignore your report, by all means feel free to expose them as a bad manufacturer.
This way, people can avoid buying this flawed product from them, or at least know a workaround the bug.

But yeah otherwise, just shaming them publicly in a video without notice: not so nice.

[ripgriffith]

it's anti color management religion and pro color management science.

This is the best-said sentence in the entire discussion, IMO.

[digitaldog]

I'm going to be using a commercial inkjet printer from a pro lab.  Image shot in AdobeRGB and printed in AdobeRGB, sRGB printed in sRGB.  This is demonstrating how bad my workflow is?  Are you expecting that the average person will vote for the sRGB?  Or maybe it'll be a tie?
Tell me in advance what is bad about the workflow I'm going to use, so I can consider it for the video.

First off, I'd shoot ONE image in raw, then convert to sRGB and Adobe RGB (1998) from the raw converter instead of two images processed in each from the camera. Why? Because it eliminates some variables. Just make the master raw 'look good' then convert without trying to adjust to the two RGB working spaces based on any soft proof. If you must capture without raw, it should be something that is a still life where the two are absolutely identical expect for the camera color settings.

Next keep in mind when you say  AdobeRGB printed in AdobeRGB, sRGB printed in sRGB that makes no sense. Again, there is no such thing as a printer that produces either. Both color spaces will be converted to an output color space. Got a good printer profile? That be darn useful. I'd be very happy to build an excellent quality custom profile for whatever printer you use. Ain't going to happen by tomorrow. If you do build or supply an output profile, we can plot the gamut of the printer. We can also plot the gamut of the two images on top of it. That would be quite useful IF your goal is to really understand what is and isn't out of gamut.

Image content is important. I've illustrated this in my video which has been referenced. It would be very easy to stack the decks in favor of either working space based on image content. Keep in mind Gary that the scene itself has a gamut and while the camera, illuminant and other factors can be static, if you capture a monochromic scene versus one with vivid colors (blue sky was an example), the results can vary tremendously.

The paper surface could play a role, and of course, the way in which the prints are illuminated and viewed by the people voting. You'd be better off getting a group of volunteers of male and female and of different age groups.

I'm sure I'm forgetting a few other items, but if you conducted your tests using all of the above, I know I'd have far less reason to dismiss the results! 

[MarkH2]

Way to retain your equanimity and step up, Andrew.  Hope Gary takes the help, but I wouldn't bet on a digitaldogfong video.

[Slobodan Blagojevic]

... flawed equipment that's incapable of supporting the most basic color management capabilities, and fail to interpret colors correctly...

A significant number of print labs still asks to send them sRGB. You seem to argue that modern equipment are all color managed, i.e., capable of converting on the fly whatever color space you supply, in which case no lab should be asking specifically for sRGB, right?

Oh, by the way, I am still on Windows XP and non-color managed IE- as is a significant percent of users worldwide). Why? Because I am a Mac guy and use a virtual machine for Windows apps only when absolutely necessary.

[digitaldog]

A significant number of print labs still asks to send them sRGB. You seem to argue that modern equipment are all color managed, i.e., capable of converting on the fly whatever color space you supply, in which case no lab should be asking specifically for sRGB, right?

The equipment per se can accept any working space to native output color space, the front end processing path may demand sRGB. Using such a front end would totally defeat the purpose of the testing and results. Gary stated: I'm going to be using a commercial inkjet printer from a pro lab. It would be useful to know the exact printer but I can't think of any off hand that couldn’t accept sRGB or Adobe RGB converted properly to a good output profile whereby that output color space could be sent 'as is' to the printer driver.

Oh, by the way, I am still on Windows XP and non-color managed IE- as is a significant percent of users worldwide). Why? Because I am a Mac guy and use a virtual machine for Windows apps only when absolutely necessary.

IF you're a Mac guy, you have Safari which is color managed.

[supercurio]

A significant number of print labs still asks to send them sRGB. You seem to argue that modern equipment are all color managed, i.e., capable of converting on the fly whatever color space you supply, in which case no lab should be asking specifically for sRGB, right?

Oh, by the way, I am still on Windows XP and non-color managed IE- as is a significant percent of users worldwide). Why? Because I am a Mac guy and use a virtual machine for Windows apps only when absolutely necessary.

Nope as I say earlier I know very little about printing and printer.
Paper or is not really part of my environment.

But as valid a color space conversion routine can be written in just a few lines of code (and with essentially no effort if using libraries), I would have hard time to consider professional quality any printing equipment that doesn't support color management, today, in 2014 as digital printing is not any kind of new technology anymore.
Tho I have no doubt it exists.

In my domain, in mobile (smartphones, tablets) there's pretty much no color management whatsoever today despite the incredible variety of displays out here and popularity of wide gamut panels.
I'm working on that.

[Bart_van_der_Wolf]

A significant number of print labs still asks to send them sRGB.

Hi Slobodan,

Converting to sRGB will solve that. What Gary seems to do is assigning sRGB to an image in Adobe RGB colorspace, which will make the image look dull and desaturated. The latter would be operator error, not a problem with profiles.

Converting to sRGB would avoid that, but may still be followed by another operator error if the actual output profile is not sRGB and the correct output profile is then assigned to the sRGB image instead of converted from sRGB to output colorspace.

Cheers,
Bart

[Schewe]

A significant number of print labs still asks to send them sRGB.

Yeah, that's the sad thing about photo labs now, however, lowly Costco has adopted a properly color managed workflow. You can go to Dry Creek Photo and download custom profiles for Costco stores all over the world. Oh, BTW, they make nice prints for really cheap.

If I was going to outsource chem or inkjet prints (they offer both in some stores) that's where I would go. In fact I know some volume wedding/portrait shooters who do just that–you can request the printed by Costco be removed on the back of the print (you don't want your clients to know you are using a cheap, properly color managed photo lab :~).

[digitaldog]

Converting to sRGB will solve that. What Gary seems to do is assigning sRGB to an image in Adobe RGB colorspace, which will make the image look dull and desaturated. The latter would be operator error, not a problem with profiles.

Not just Gary but good old Ken Rockwell, another color management religious zealot who I suspect Gary and Will have read and accepted as fact:
http://www.kenrockwell.com/tech/adobe-rgb.htm

Adobe RGB is irrelevant for real photography. sRGB gives better (more consistent) results and the same, or brighter, colors.
1.) Adobe RGB requires special software and painstaking workflow not to screw it up. Make one mistake anyplace and you get dull colors, or worse. You cannot use Adobe RGB on the internet or for email or conventional photo lab printing. If you do, the colors are duller.
Adobe RGB squeezes colors into a smaller range (makes them duller) before recording them to your file. Special smart software is then needed to expand the colors back to where they should be when opening the file.
If you use Adobe RGB you will have to remember to convert back to sRGB for sending your prints out or sharing them on the Internet. Otherwise they look duller than sRGB!

Pretty absurd but that's the kind of color management voodoo that remains on the web that other's read, accept and preach.

[Slobodan Blagojevic]

Converting to sRGB will solve that...

Indeed, Bart. This is what you and I, and everybody else on this forum would do. But Garry is apparently addressing an audience that doesn't do that, doesn't know how to to do that, or doesn't want to know that. And apparently, those print labs do not do that either. Not because they do not know how, but because (I am guessing here) that is an extra step they are not willing to waste time on.

[garyfong]

DigitalDog:

Depends on the paper surface?  Male vs Female?  Different ages for the test group?  Wait for you to build a custom profile for the commercial printer we are going to use?  But it won't happen by tomorrow?  Then you wouldn't be critical of the test?

It'll be images of an X-rite color checker, same lighting, custom white balanced, within seconds of each other.  Results shown to the general population, who will vote.  And I will publish far and wide the results.  If AdobeRGB is that much better, then that would be a huge reason for people to maintain that workflow.  If not, then it won't be.  I have a hunch how it's going to go, but I can't say for sure until I test it in a controlled environment with no variables between samples.

I'm not going to take it in RAW then polish it up.  They will be identical.  SOTC.

All of this heated discussion going on, isn't AdobeRGB going to be hugely the winner for all of the work you do to hone this in?

Are you confident that AdobeRGB is so much better than sRGB, or not?

[garyfong]

Andrew - following the results of this test, would you be up for a Skype interview for my YouTube channel?  I'll limit the questions to your interpretation of the results, so you can speak your mind. 

[digitaldog]

But Garry is apparently addressing an audience that doesn't do that, doesn't know how to to do that, or doesn't want to know that.

And neither of his video's address this while spewing technically incorrect data points about color spaces. Is the new mantra "don't do stupid stuff"?
How many video's are out there telling people that JPEG files set at quality level 1 don't look as good or print as well as quality 9? How many video's or articles are there that tell new users not to take their full resolution captures, resize using Nearest Neighbor to 100x100 pixels and save over the original high rez file? How many video's are there telling people it's a bad idea to put their DSLR's in the dishwasher to clean the lens?

Gary isn't addressing the audience you suggest he is, at least he isn't teaching them what not to do. I'm unclear how you see defending his video is in any way justified other than what some have already stated in a fraction of the time it takes to view his video with all the misinformation: If you don't understand anything abould color management, set the camera to sRGB. Anyone could create a video what that statement in mere seconds, no one would or should dispute it. 

[garyfong]

Quote: Indeed, Bart. This is what you and I, and everybody else on this forum would do. But Garry is apparently addressing an audience that doesn't do that, doesn't know how to to do that, or doesn't want to know that. And apparently, those print labs do not do that either. Not because they do not know how, but because (I am guessing here) that is an extra step they are not willing to waste time on.

Correct.  I'm addressing the topic of, "why do my colors look duller when I switched to aRGB setting?".  This is a very, very common complaint from the average photographer with average experience.  I am excited to do my comparison test and put it to the public, and show the test results.  I have no predictions, but just a hunch, as to how the results will be.

[digitaldog]

Depends on the paper surface?  Male vs Female?  Different ages for the test group?  Wait for you to build a custom profile for the commercial printer we are going to use?  But it won't happen by tomorrow?  Then you wouldn't be critical of the test?

What isn't clear about the testing process I propose?

It'll be images of an X-rite color checker, same lighting, custom white balanced, within seconds of each other.

Why just a Color Checker? Not a good idea as the sole object in the test for a number of reasons.

I'm not going to take it in RAW then polish it up

I didn't suggest you polish. IF you capture raw, we can (you could) also see the results of ProPhoto RGB on the data. Again, IF you viewed my video, you'd see that Adobe RGB is too small a working space for many scene's and output devices. You are limiting the data you can gain from the test by allowing the camera to funnel the color even into Adobe RGB (1998)!

All of this heated discussion going on, isn't AdobeRGB going to be hugely the winner for all of the work you do to hone this in?

I'm somewhat concerned that you feel there has to be a winner and thus a loser. Based on your push back in just your last post about how to conduct such a test, after you yourself asked us for comments, I suspect this is just another waste of time.
Why don't you tell us the full aspect of the test: what capture device, of what subject under what illuminant. Printed on what device using what kind of color management. What print size and viewed how? If you really DO want to get to the bottom of this, what is the testing process and what is it you hope to gain?

Lastly, how will any of this fix the mistakes on your other two video's?

[fdisilvestro]

It seems to me this case is hopeless. The issue with muted colors is usually related to wrong encoding (not using color management) rather than out of gamut colors.

Encoding issues could led people to believe there are differences in the reds. If you check, the R and B coordinates are the same in sRGB and Adobe RGB, the differences are in the G coordinate, so the out of gamut colors are in the cyan-green-yellow/green areas.

The other difference is the TRC, which causes some out of gamur colors at high L values.

[digitaldog]

Correct.  I'm addressing the topic of, "why do my colors look duller when I switched to aRGB setting?".

Because you did something wrong. Simple as that. Want to see the use of sRGB look grossly over saturated? It's as easy to do as what you just proposed, take Adobe RGB and treat it as sRGB. What have you gained other than show someone they treated the data incorrectly? Take sRGB and use Photoshop's desaturate command or set Hue/Sat (saturation) to -100. All the color disappeared. You didn't want the color to appear to desaturate? Don't set the controls that way dummy.

[bjanes]

A significant number of print labs still asks to send them sRGB. You seem to argue that modern equipment are all color managed, i.e., capable of converting on the fly whatever color space you supply, in which case no lab should be asking specifically for sRGB, right?

Oh, by the way, I am still on Windows XP and non-color managed IE- as is a significant percent of users worldwide). Why? Because I am a Mac guy and use a virtual machine for Windows apps only when absolutely necessary.

If you are Ken Rockwell sending your prints to Costco, you wouldn't really lose much gamut by sending your images in sRGB for printing, since the gamut of the printer is hardly larger than sRGB. Shown below is a 3D gamut plot showing the gamut of the Fuji Frontier printer at my local Costco which prints on glossy Fuji Crystal Archive paper (in color) and that of sRGB (in white). This is the image on the bottom. There are only minor differences as shown. Costco does not offer glossy paper for their Epson inkjet, so I plotted the gamut of the Epson 3880 using Premium Glossy Epson paper against sRGB, shown on the top. The Epson has a much wider gamut, and one could not obtain the benefit of the wide gamut of this printer using sRGB as a source.

Bill

[fdisilvestro]

It'll be images of an X-rite color checker, same lighting, custom white balanced, within seconds of each other.  Results shown to the general population, who will vote.  And I will publish far and wide the results.  If AdobeRGB is that much better, then that would be a huge reason for people to maintain that workflow.  If not, then it won't be.  I have a hunch how it's going to go, but I can't say for sure until I test it in a controlled environment with no variables between samples.

Out of the 24 patches of the color checker, there is only one cyan patch that is out of sRGB. If your results have any other differences, then your process is wrong