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

[Eyeball]

While I do agree with others that there are a number of misstatements in the video I also find that there is a kernel of truth to what Mr. Fong says.  For users whose sole goal is to post pictures on the Internet the Ron Popeil approach should be adopted, "set your camera to sRGB and forget it."  These users shoot JPGs and probably have no clue at all that their camera has a RAW setting.  LuLa participants represent probably less than 0.0001% of the photographic public and I would disagree that we are not elitists.  Many go beyond just the mainstream programs and use a variety of plugins and experiment with other software to get optimal results from the images we capture.  We can all chuckle at the errors in the presentation but at the end of the day, the Internet will reign supreme in terms of allowing anyone the freedom to make a fool of him/herself.

Alan, I just recommend that you don't fall for Gary's strawman.  Few, if any, critics of that video are arguing against a recommendation of sRGB for beginners or people who aren't interested in color management.  If Gary had just stopped with that there would be no controversy here at all.  The problem is that he goes on for another 5 minutes with unnecessary, misleading, and incorrect explanations that are only going to confuse people.  It's not even necessary if, as he says, he is going for a "keep it simple" explanation.

Also be sensitive to his false dichotomy where he presents the situation as "his simplified explanation" vs. "a mind-numbing complex one".  This again, is not the issue.  The issue is that he is giving a "simple but incorrect" explanation when there are any number of ways to explain just as simply but also correctly.

All of these little tricks (strawmen, false dichotomies, mis-representation of what others have wrote, etc.) are the classic signs of a troll.

[digitaldog]

While I do agree with others that there are a number of misstatements in the video I also find that there is a kernel of truth to what Mr. Fong says.

There is, no question. That's not the topic or issue. He completely got the facts of color spaces wrong and hugely diluted that message in the process.
If two wrongs don't make a right, does half a dozen wrongs and one right make a good video? Is that fair to the audience? If other peers point out the wrongs and the presenter absolutely refuses to fix the mistakes and worse, argues and belittles those who are hoping to clarify the presenters understanding, is that OK because the diluted message has a bit of merit? I don't think so Alan.

[Royce Howland]

A couple of years ago, Canon had a promotion for their printers, where they offered a free 8x10 if you send them a file. They said you can send them sRGB or aRGB. Knowing, at the time, that "sRGB is good for web, aRGB better for inkjets," I sent them aRGB. See what I got back:

The fact that somebody downstream from me might make a mistake with my files because they don't know what they're doing, is not an argument for me to throw away my own goals for rich, accurate colour. It's an argument for me to find better service providers to work with.

Others have suggested what Gary Fong should have said was something like "if you don't know what you're doing, use sRGB". But one reply above got it more correct -- "if you don't know what you're doing and nobody else who works with your images knows what they're doing either, you should all use sRGB". My variation on this might be something like "if you don't know what you're doing, use Adobe RGB and find service providers who do know what they're doing". This is what I teach my students, and it works great. Plus at any time they have questions, I can do like Gary Fong suggests and show the evidence why it works, based on real-world examples.

I've known enough about the purpose of capturing good colour at source, vs. dealing with lower fidelity output devices downstream, that from the very beginning of my digital photography I've never seen the rationale for capturing sRGB. That would be like a recording studio saying "we know at some point that low bit-rate budget CD's of this music will be produced, and probably some people will record live performances on something and make really low quality mix tapes available. So let's just trash our high fidelity studio masters and go straight to cassette tape for all our production work." Or the video equivalent.

It's a disservice to provide confusing, misleading and incorrect information to viewers under the otherwise perfectly fine goal of helping them understand how and when to worry about certain technical matters like colour spaces.

Unfortunately, these discussion threads have become very muddled and argumentative, and the likelihood of anything positive happening now seems extremely low. The uninformed viewers are the ones who will lose out because they'll form a poor understanding of getting good colour, and how to make good future decisions about good colour when it matters more to them than it perhaps does now...

[digitaldog]

I am super thankful to Will Crockett for giving me the run through about Rodney Andrew.  
And I am dedicated to bringing out clarity, and truth, especially about the history and background of Rodney Andrew.
English is my first language, and I can read it just fine.

The above text would suggest otherwise.

[Slobodan Blagojevic]

The fact that somebody downstream from me might make a mistake with my files because they don't know what they're doing, is not an argument for me to throw away my own goals for rich, accurate colour. It's an argument for me to find better service providers to work with...

You are missing my (and Mr. Fong's) point entirely. The point being: by sending aRGB you are taking a gamble that the printer will know what they are doing. And if CANON, having a PROMOTION, aimed to demonstrate the quality of their printers, specifically offering to accept aRGB, can make such a mistake, what chances are left for ordinary labs? Apparently, way too many people are experiencing that same mistake, otherwise that would not become a topic of discussion.

I print my 20x30 in a lab that specifically asks for sRGB. Why do I deal with them then? Would I not want to be rather dealing with a pro lab that can send me their printer profiles to softproof first? Sure I would! But guess what: such a lab is about 2-3 time more expensive. And frankly, my printouts are just fine with sRGB.

A question for experts: how many real-world photographs, not theoretical models, actually exceed sRGB gamut? And if they do, by how much? And would you notice? And who would notice? And if you notice, would you care? Would most people care? Would common people (non-photographers) care?

We are all elitists and measurbators here (including myself). We look down on those who do not use RAW, or ProPhoto RGB, display and print profiling devices, wide-gamut monitors, etc., as unwashed masses, worthy of our disdain and contempt. Guess what: those people actually have way more fun and joy with their photography than we do. We can't sleep after discovering a speck of dust bunny we failed to remove,  obsess over CA we didn't notice, fret over a blown highlight, etc. In the meantime, "unwashed masses" are having fun, enjoying special moments they captured, however imperfect.

This is the fate we chose for ourselves. However, like all religions, we insist on spreading our views onto others. We have Jesuits and Conquistadors, Holly Warriors and Inquisition, who will spread by force (e.g., internet mocking) the gospel of "proper" photography technique to the unfaithful.

There is, however, a number of professional and famous, photographers who shoot mostly JPEG (gasp!) and sRGB (gasp!!). It is a choice, people! Their choice. It works for them (and millions of the "unwashed"), so let them be. There is a room for everyone in this wide world of photography. And for different views and choices. No need to shove ours onto others.

[Czornyj]

A question for experts: how many real-world photographs, not theoretical models, actually exceed sRGB gamut? And if they do, by how much? And would you notice? And who would notice? And if you notice, would you care? Would most people care? Would common people (non-photographers) care?

Each one that has a blue sky. And yes, it's really easy to notice the difference, even in case of c-print (the bigger the print, the more obvious difference).

[ripgriffith]

A question for experts: how many real-world photographs, not theoretical models, actually exceed sRGB gamut?

I would say just about every daylight outdoor photograph unless you are shooting an extremely monochromatic scene.

[digitaldog]

You are missing my (and Mr. Fong's) point entirely.

I don't think so. I don't think anyone here is suggesting that people who mishandle images, through color management mishaps or otherwise can produce poor results. Gary's simple point that, if you don't understand anything about color spaces, use sRGB has merit. HOW he provided the message, filled with misinformation is the key point of this thread and what I believe you and most certainly Gary don't accept.

If you provide me with an sRGB image and I view it on my wide gamut display without a color managed app, it looks poor. So with the logic of 'just use sRGB', in this context, sRGB is the problem. That's rather silly since one would expect someone who's gone out of his way to purchase a wide gamut display would understand this could be an issue (and yet, there are posts from people doing just that).

Do we expect that wide gamut displays may or will not get less expensive and more popular in the future? If they do, will the 'just use sRGB' message need to be corrected?

Bottom line isn't which color space is 'better' or what you should use. The issue is the two video's Gary produced are technically incorrect in terms of basic color science. Those issues were reported to Gary in an attempt for corrections and clarification. Gary refuses to expect his mistakes and worse, tells others they are wrong. That's a disservice to his audience and perspective audience. His heart might be in the right place although that's highly questionable considering his refusal to accept the facts let alone correct them. So what is his motivation? I don't know. I don't really care. I do know his message is diluted and filled with inaccuracies. And that's the bottom line in terms of this topic, there isn't any defence for it!

[digitaldog]

A question for experts: how many real-world photographs, not theoretical models, actually exceed sRGB gamut?

Lots (I know that's not a statically acceptable answer). Did you not see my video on gamut where I show a number of such examples?
Let's say it's only 25% which I think is probably very low. Would you be willing to clip data you could capture and could output when you don't have to?
If you understand what sRGB is based upon, you'll soon see many real world images don't fall into that simple shape for a number of reasons.
Why do you suppose photographers capture raw and why many raw converters use a vastly larger processing color space than sRGB?

[digitaldog]

We are all elitists and measurbators here (including myself). We look down on those who do not use RAW, or ProPhoto RGB, display and print profiling devices, wide-gamut monitors, etc., as unwashed masses, worthy of our disdain and contempt.

We all don't and shouldn't. There is absolutely nothing wrong shooting JPEG instead of raw IF you know why you want to go that route. News and sports photographers have very legitimate reasons for not shooting raw. But if someone produced a video and said: "Shooting JPEG gives you more data to work with and a better image than raw", I'll bet a lot of other's would have issues with that technically incorrect statement. Yes, if you gave someone who had never processed a raw file such data and a converter they didn't have a clue how to use, the JPEG would very likely be a better appearing image. Is their lack of experience using the converter a fair way to convey that JPEG is better? And the statement about more data would be blatantly wrong!  This is exactly what Gary has done in his video. Don't separate the correct message with all the technically flawed delivery Gary provided in two videos. Wrong is wrong, even if the underlying message has merit. Worse, to ignore and argue with dozens of people who correctly pointed out the errors of the delivery is kind of inexcusable.

[Czornyj]

Do we expect that wide gamut displays may or will not get less expensive and more popular in the future? If they do, will the 'just use sRGB' message need to be corrected?

I'm sure there will be virtually no "narrow gamut / sRGB" displays in near future. ITU-R Recommendation BT.2020 assumes a gamut much wider than AdobeRGB as TV broadcast standard, the amount of consumer mobile devices and TV with wide gamut (GBr LED, quantum dots, OLED, laser projectors) is constantly increasing...

[supercurio]

Hi everybody.  I saw a direct link to this forum, and saw your comments, and thought I would address them.

Hi Gary, thanks for the good attitude of coming here to talk politely.
First, congrats on your long career, I have no doubt your success are results of passion and hard work.

Reading your message explain how you are experienced with color and color management.
You are experienced as a user, tuning parameters, sharing recommendations based on your perception.

A mistake made here is that you confuse experience with knowledge.
Unfortunately, your perception of how color management work and and what color space are is inconsistent with the science, maths and algorithms powering the standards and its numerous implementations.

Also, reading your replies on YouTube made me sad.
My fiancée is a professional and successful YouTube reviewer and I learned through her how to interact with an audience and also how not to, and I'm genuinely concerned with your initial reactions that over time this YouTube thing will become just miserable to you, especially if you read and moderate comments directly.
Believe me I'm speaking from experience here.

My recommendation would be to simply put offline the Adobe RGB vs sRGB videos, they're factually wrong.
It's no big deal, professional YouTubers put down videos all the time, and re-upload new ones that check out with facts and current state of knowledge.
Unlike Vimeo, you can't replace the content of a video on YouTube, the only choice is to put offline and redo. In this context, it really is the only thing to do and it will help you grow as an educator as well.

Then please continue interacting with experts in a field, that might actually be good at what they do and not so self-proclaimed.

In order to build trust with your audience, I'd also recommend as you have ties to Sony to publish an ethic statement, and avoid naming product presentations "reviews"
A pretty good example to follow here IMHO is Philip Bloom.
He has been working with camera manufacturers, on paid jobs and even promotional shoots. But he discloses that clearly and actually keeps his complete freedom of speech as he reviews stuff.

[MarkM]

This:

So I intimately know color management…

and this (from here):

You have absolutely no idea what you are talking about. Cyan green getting clipped? those colors are in the center of the spectrum.

are simply incompatible statements. The second quote can't even be called a misunderstanding of color management — it is a TOTAL absence of understanding the most basic tenants of color management.

Gary, you need to swallow your pride, stop considering yourself such an expert that you don't need to listen to or learn from people who know what they're talking about, stop being so defensive, stop confusing success in one area with expertise in another, and go back to the basics and learn the foundations of color science. If you don't want to do that, you need to stop "educating" people on this subject. You do them a huge disservice. If they want to go from your teaching on to further understanding of the subject, they will need to unlearn you explanations before they can move on.

I'm not sure you are sincerely interested in what anyone has to say, but just in case you are, here are a couple sincerely offered suggestions:

Get rid of the wide rainbow explanation of color gamuts.
This doesn't accurately describe how color gamuts work and leads to HUGE misunderstanding. AdobeRGB doesn't include a wider range of wavelengths on a one dimensional spectrum. It is a larger volume in a three dimensional space. The xy chromaticity diagram, while not a perfect representation of color gamuts, is a much better way to compare gamuts. It's not hard for even a beginner to understand — one triangle is bigger that another — not hard. It has the added benefit of being the way most of the world compares color gamuts. When your student looks at other sources they will encounter this diagram often. Looking at a comparison between AdobeRGB and sRGB on this diagram you will see that actually, yes, cyans and greens can easily be clipped.

Don't explain color gamuts by switching monitor profiles to working spaces like sRGB
You and I may know that you shouldn't set your monitor to sRGB, but it is a commonly occurring misunderstanding among people starting out. I frequently see people who want to work in a color space like sRGB and think the way to do that is by setting their monitor to sRGB. There's never a good reason to set your monitor's space to working space. If you want to show what happens when the color numbers from wide space like AdobeRGB are interpreted as though they were in a small space like sRGB, just assign the sRGB profile to an image.  This is a much simpler and more accurate way to show the phenomenon of duller colors coming from a wide space to small space. It would also be an opportunity to show what happens in a color managed environment by comparing the results of assigning a space vs converting to a space. When you convert from profile to profile, which is what will happen in a color managed workflow, you will see that using AdobeRGB doesn't cause dull colors in the way it does when you work in a non-color managed workflow and go from large to small spaces.

Just avoid talking about the number of colors in a gamut
I left a comment about this on the Youtube page. You didn't have a coherent argument for why you thought I was wrong but offered this to your audience:
"Folks, this is a prime example of someone who knows absolutely nothing about this topic, yet spouting off like he knows.  He doesn't know the topic."

I was (and still am) right about this. But it doesn't matter — for the point you are trying to make and only muddies the waters. Just avoid it altogether—if you want to compare color spaces, just use terms like 'wider' and 'greater volume', 'larger ranger', etc. Talking about either the number of colors, or the amount of information doesn't clarify anything and will likely lead to confusing inaccuracies.

Be honest
It's a wonderful ability to be able to explain technical subjects to a general audience. People who are great at this, like Richard Feynman, are able to simplify subjects without intruding inaccuracies and offer simplified explanations that make it easy to move on to more complete understanding. It's more expedient to say AdobeRGB causes dull colors than it is to say using a non-color managed workflow can cause dull colors in certain situations, but the later is true and the former is only half-true. In a non-managed workflow going from a large gamut to small gamut, such as going from AdobeRGB to a iPad without conversion does result in duller colors. But going from a small space to sRGB to a large one like a wide-monitor has the opposite effect in a non managed workflow. If you are honestly trying to help your audience build a foundation of knowledge, you need to find a way to explain both these phenomena. If you want to do this, you need to be honest about your own understanding of the fundamentals, and it's not clear that you are.

Based on my previous encounter with you, I suspect you will not take any of this to heart, but it is offered sincerely, not necessarily for your benefit, but for the benefit of your audience for whom I have a lot of sympathy.

[Slobodan Blagojevic]

... Did you not see my video on gamut where I show a number of such examples?...

Andrew, which video? Link?

[digitaldog]

Andrew, which video? Link?

High resolution: http://digitaldog.net/files/ColorGamut.mov
Low Res (YouTube): http://www.youtube.com/watch?v=n0bxSD-Xx-Q

You may want to fast forward to 17:44 into the presentation to see the examples of images and color gamut, the first part of the show I'm sure you fully undertand.

[digitaldog]

But do not capture in A and go to S.  Simple message.

Wrong again. In fact if you’re telling non raw shooters how to setup the camera, capture Adobe RGB (1998), convert to sRGB when one needs sRGB. Can’t go the other way (it’s pointless).

Yes, this bristles with the people who spend a lot of time teaching about color management (seems like this is a forum for instructors, if I'm not wrong)…

That is indeed wrong. What bristles people here is blatantly incorrect information. And not just here, there are prime examples on your two color management videos.

Photographers en masse express in sRGB down the chain and those would benefit by not switching their camera to AdobeRGB - which is what this video addresses in a simple way using metaphors like "muffin top" and "wide rainbows" as well as changing a monitor display profile to show color clipping.

The metaphors and examples used are wrong, simple as that. Many others besides just me have told you this.

It was meant to be simple, and at the beginning of the video I said this is one of the most hotly debated topics, so I knew to leave the comments in for people to sample what happens here, in DPReview, at conventions, etc.

It is only hotly debated by people who refuse to understand and discuss basic color management like yourself. Those like the vast majority in this post understand the topic and disagree with you.

The elite are unhappy that their refined techniques are not adopted by all to produce the greatest results.

Elites, really? This is part of the stand up comic act right?

And, by the way, most of the younger photographers or newer ones wind up buying Instagram-looking photoshop actions and torch the midtones anyway.  So much for your enhanced color detail.

Comment has absolutely nothing to do with color management or the misunderstandings expressed in your two color management videos! Why go there except to misdirect?

I will be making a series of videos that we plan to roll out to especially educate on this topic.  I'll be doing Skype interviews for my YouTube channel about where all of the confusion is coming from, and I'll be making sample prints in both sRGB and AdobeRGB workflows, and we will put them on display on the sidewalk outside my office, and let people vote which is better.

OK so you are going to keep the original incorrect misinformed videos up, and you are going to make yet another comic attempt without peer review? You think that’s a good idea Gary?
Tell you what, you build the video, keep it private but post it here first for peer review. I'll even promise to keep out of it if that’s what it takes to get you to submit it for review. IF you really do care about accuracy on the topic, you have nothing to lose and everything to gain by letting a few folks here give you an opinion of the technical correctness of the topic. I can't see you doing this but I'd love to be proven wrong (something you've struggled to do for days now).

We plan on mentioning this thread, and some of your names and opinions as topics of discussion.

Great, can’t wait.

[Eric Myrvaagnes]

I finally watched the video to see what all the fuss was about. Now I'm certainly no expert on color management, but I have picked up a bit from the likes of Andrew and Jeff and Eric Chan over the years.

The essence of Gary's argument seems to me to be this:

1.   sRGB has a much shorter spectrum ("rainbow") than does AdobeRGB.
2.   Gary seems to think the way one converts from sRGB to AdobeRGB is by chopping off the ends of the AdobeRGB rainbow (so I wonder why doesn't that just leave you with a few yellows and a bit of green??)
3.   His comparison "shows" that AdobeRGB gives muddy colors in an image converted from bright colors in sRGB.
4.   His conclusion seems to be that sRGB has brighter colors than AdobeRGB.

If Gary's "syllogism" is correct, it follows that the shorter the spectrum, the better the color rendition. So obviously sRGB is still much too wide. Just think how bright your colors would be if you used a color space that was only, say, 10% as wide as sRGB!!!

Fantastic!!!

So I guess the ideal would be a color space that only has two wavelengths in it (I suppose a single wavelength wouldn't let you distinguish between colors at all).

Have I understood this correctly, guys?   

[Czornyj]

So the pursuit for the perfect editing space has no end, as you can always choose two wavelengths that are closer and closer to each other.

I just can't wait to see that Crockett & Fong "Spectral explorers and lightwave surfers" video!

[Tibor O]

I think you are all missing the point here 

I think that we all, including Mr. Fong, agree that:

1. sRGB is a narrow gamut color space
2. Adobe RGB is a wide gamut color space

However we see a color space as a 3D space, where Adobe RGB expands above sRGB mostly in the greens and cyans as seen in the pic below. Please compare the smaller sRGB triangle (blue) to the bigger Adobe RGB triangle (yellow). If capturing and post-processing in the wider Adobe RGB color space and then posting for web or printing in e-labs etc. one needs only to convert the file from Adobe RGB to sRGB by doing this simple Photoshop command Edit - Convert to profile ... No colors from sRGB are lost, of course, just the colors that exist in Adobe RGB but not also in sRGB are translated into sRGB (more color information is squeezed into less color information). Please note that for the purpose of this simple illustration a 2D picture is used, having 2 axis, x for width and y for height. The "dimensions" of course are Red, Green and Blue.

But, it is Mr. Fong's theory that color space is not 3D but 2D instead, where Adobe RGB expands beyond sRGB because there are more color nuances in Adobe RGB than in sRGB which are equally distributed (therefore, more reds, yellows, greens, cyans, blues and magentas). Please see picture from his video below, which only has one axis, the x for width. So, if capturing and post-processing in the wider Adobe RGB color space and then posting for web or printing in e-labs etc. one will loose the reds and the magentas because the web browser, e-lab printer etc. will clip these colors, because they fall out of the sRGB gamut.

However, and please listen very carefully, because this is a groundbreaking discovery in color theory, with this procedure one will get in his or her picture (on the web or in print) the colors that don't exist in sRGB color space! Wow!

So, according to Mr. Fong's new color theory if you want more color nuances and are willing to sacrifice some reds and magentas capture in Adobe RGB and print in sRGB. You will gain more greens, more yellows, more cyans, more blues, more of everything, except for reds and magentas.

However, from the 5 minute video from Mr. Fong I did not get the info how this is possible. So, I would like to use this opportunity and ask Mr. Fong to elaborate on this topic. My question is "How are the colors that do not exist in sRGB able to go from Adobe RGB to sRGB if you shoot in Adobe RGB and print in sRGB?" I have marked this in the attached picture from Mr. Fong's video (not scientifically, but by naked eye).

Remember, we are all missing the point here

[garyfong]

Wrong.  I never said that sRGB had more data.  In fact, in the beginning, I say AdobeRGB is better provided you have the equipment that exploits it.  The balance of the video shows video clipping in a simple way, using metaphors like "wide vs narrow" rainbows, muffin tops, etc.  I address the photography enthusiast who wants to appreciate and enjoy photography.

What happens is the neophyte reads information that doesn't apply to them, "AdobeRGB is better" (which I said in the beginning of the video, it is) but the vast - vast majority of photographers JPG and sRGB.  The wide rainbow is a metaphor for a wide gamut being spilled onto equipment that can only express a limited width, thus the cropped colors.  Would it have been more techy had I split the rainbow up and move the colors around so the clipped colors would be at the ends?  Or make a bump in the parts of AdobeRGB that has more color information?

The challenge is this.  In five minutes - give a majority of the people working information that they can use right away. 

The video I am doing with Will Crockett will address just this.  How such confusion spills out into the search engines, how sRGB vs AdobeRGB is such a hotly debated topic (as is JPG vs RAW) and, additionally (this is inspired by Rodney) how the elitist 'measurbators' (I loved that word!) spout out anger like this which results in people who want to enjoy photography and be more discouraged.

There can be no debate that if you shoot in AdobeRGB and share via web browser, or make wet process lab prints, you would have to convert your aRGB file to sRGB first or suffer muted colors.

Nowhere in the file did I say sRGB had more data.  The illustration shows how sRGB compresses all of the colors available to the naked eye into a space with less space between colors (narrow gamut).