sRGB or Adobe RGB?

<p>Well I converted my original ProPhotoRGB version to sRGB and converted back to ProPhotoRGB and I get the same numbers as jacopo's posted sRGB version as noted previously, but I still get a the same slightly duller (meaning I notice a slight boost to cyan in the yellow) converting to sRGB.</p>

<p>Apple's DigitalColor Meter shows a difference sampling the sRGB version against the original ProPhotoRGB version as well. The numbers I get in the ADM are 255,233,0 in the ProPhotoRGB version and 249,253,52 in the sRGB version.</p>

<blockquote>

<p>Post to the Web in 8 bit sRGB for the widest audience (you need to downsample anyway).</p>

</blockquote>

<p>8 bit sRGB that's tagged as such, that is. The broken color management on the Macintosh just ships the bits over to the monitor* with no conversion at all if there's no color space information in the file. This, despite 99+% of such files on the Web being in the sRGB color space.<br>

* Which, thanks to the non-sRGB color space the Macintosh uses, results in the sRGB file looking all washed out and desaturated.</p>

<p>If viewing sRGB tagged images in Safari, you'll get the same preview you see in Photoshop. Off colors arise when viewing UNTAGGED sRGB and other color space images where Safari ASSIGNS the monitor profile. If on a laptop the untagged colors in Safari will look even worse because of the reduced gamut of some laptops.</p>

<p>I don't view untagged sRGB images or edit color on laptops.</p>

<p>Here's another odd thing about creating color in ProPhotoRGB. Do two color fills in ProPhotoRGB side by side overlapping each other with these RGB yellow numbers...255,233,0 and right next to it 230,230,50. On my display they look identical even when overlapped over one another. The thing is one has clipped data and the other doesn't. See if you get the same preview.</p>

<p>And just to see if my eyes aren't deceiving me I took two Lab measurement readings one off Apple's DigitalColor Meter and the Info palette in CS3 and compared the ProPhotoRGB yellow to the converted sRGB yellow. Here's what I got:</p>

<p>ProPhotoRGB-CS3 info palette>Lab: 92,-2,128...ADM>Lab: 92,-2,98</p>

<p>Converted to sRGB-CS3 info palette>Lab: 91,-4,89...ADM>Lab: 91,-4,89</p>

<p>Tim,<br>

when you go from (230,225,19) in ProPhoto to sRGB, you get (255,230,0). Red and blue channel clipped.<br>

So going back from sRGB to ProPhoto, I get (224,224,75).<br>

You cannot return to the original values.</p>

<p>Tim. Just for your information I was in charge of Research, Development and Acquisition of civil aviation GPS for the FAA when I retired. I think i know good technical writing when I see it; and, I know something about systems design on a fairly large scale. We were particualrly interested in human-machine interface in all of our systems development as that is critical for pilots, air traffic controllers and technicians. If you want to see complicated software go to the software we developed to land and navigate airplanes on GPS. The accuracy for approaches has to be as little as two feet. The software has to fail passively. It has to be reliable .99999 per cent of the time and when its not you have to know within 5 seconds in the cockpit that it's not and you receive that failure notice from a ground station through a geostationary satellite. No failure can feedback into the system. There are several millions of dollars in this type of development. I don't really care about photons no more than the pilots who fly GPS approaches who really only care about getting their airplane on the ground safely. What I really want with my pictures is to get decent prints that match my monitor without delving into the arcane science that went into it. I think that I have to do much too much process to get that kind of print. My monitor does match my prints closely enough to satisfy my requirements. Andrew Rodney explains it a hell of a lot better than most and I don't know why you take issue with him. What I would like to see is software like GPS that does it all for me without doing monitor calibration, soft proofing, matching ICC profiles, 16 bit, Relative Colormetric, etc. No wonder people are confused. Photoshop is a tree based software system rather than function based and as a result there is a hell of a lot steps to go through get outcomes in my humble opinion. Now I understand a lot of this is brought about by a lack of industry standardization between manufacturers, a lack of standards generally and a twenty year old photoshop system that was built system upon system that had as a design parameter "backward compatibility" which has prevented signnificant changes in the human interface with the product. My CS3 still looks about the same in many ways like the first system I bought in the early nineties except it is far more complicated and is not very intuitive. I think PS needs a ground up re-design. The NASA engineers were right in that color managment is way too complicated to effectuate and I bet they all wanted to do some human and software engineering to integrate and simplify getting a decent product for your average user. By the way I have worked with NASA engineers and they are pretty smart people. NASA like the FAA and Navy employs a significant number of human factors engineers to make things easy and safe for the user. The first time I flew the FA-18 simulator I was fully capable with the Heads up Display in twenty minutes. That ease of use came with intensive human factors work with hundreds of subjects. I think Adobe should do some of that to clean up some complicated processes. After all PS is a mass consumer product. I am still learning photoshop after owning at least ten versions of it. There is a disconnect here. All I want to know about color managment is how to effectively use it to make what I am really interested in and that is pictures. Mr. Rodney's tutorial gave me enough basic knowledge to understand where I was trying to go with all the adjusments I have to make to get there. I have the process pretty well memorized. The basic question is what do I use, Adobe RGB or sRGB. I can't really answer that except to say what others have said. AdobeRGB mostly and convert to sRGB for the web. I have used ProPhoto effectively also. </p>

Plenty of questions:

 

Any comparison of rendering intents (as implemented, since algorithms may vary and are probably not standardized) in

different software/browsers? Additionally, even while editing in ProPhoto, one only gets to see the colors displayed on

the monitor (which might not cover sRGB exactly). Also, does soft proofing in sRGB help on a wide gamut monitor

(while editing in ProPhoto)? Finally, has anyone been able to get a good wide gamut monitor profile using a spyder2

(not spyder3 which is advertised as wide gamut)?

<p>Dick,</p>

<p>I never said I take issue with Andrew Rodney. I said he doesn't know everything about this subject or at least the issue and point of view I've presented. Andrew and I are mainly in agreement on a lot of things concerning digital imaging.</p>

<p>I've read through his book among a few others from other authors and throughout the ten years researching digital imaging 6 of which has been in web discussions with Andrew I've always been thorough in my questioning trying to nail down exactly what is being discussed. If I'm wrong, I'm wrong and I'll admit it. But I've found there IS a plumb mark in knowing all the things going on under the hood in all of this.</p>

<p>Usually folks who are in the know and develop a reputation as such don't like or ignore my specific questions and requests for proof or a demonstration in what they say. Or they answer my question in a way as if I'ld just fell off the turnip truck by redirecting the discussion or providing an answer in a way that implies my question is unimportant or of little concern and to just take their word for it. Sometimes they even reply that they want to be paid for such answers. When that happens I know I found the plumb mark and I move on.</p>

<p>All I do is use these products and notice all the little things that happen and see if it matches up with what the experts profess. Sometimes it doesn't and when it doesn't, I holler loud and clear. That's the free exchange of ideas that is the web.</p>

<p>The thing is on the web when someone starts to come up with questions, answers and issues through their own research and experimentation and present those findings as I've done here, they too get perceived as being an expert but don't get taken seriously because of the lack of credentials.</p>

<p>I never said NASA engineers weren't smart. I'm sure they're smart. I don't know what your point was in mentioning that. It sounds like you were being defensive or taking it personal. If not then it's just a miscommunication that often happens in discussions on the web.</p>

<p>Hell, what am I? I'm just an ex-cartoonist/illustrator/prepress technician/art director without a degree trying to learn as much as I can and maybe show others something they haven't considered about this technology. My philosophy on becoming a so called expert is you never stop learning and asking questions. Problem is I never get to see or know on the web where these experts ask questions to expand their knowledge. That troubles me.</p>

<p>So I'm not trying to confuse anyone. If they want to edit their raw files in the color space of their choice then so be it, but I still get to have my say just like everyone else here and if that confuses them, that's not my problem.</p>

<p>Geez! Not again. I just shot a macro of some intense orange flowers in the shade. They look gorgeous in ProPhotoRGB, no clipping of any data in the histogram. The orange is mainly made up of 195,120,20 in ProPhotoRGB.</p>

<p>Converted to sRGB and the preview of the orange flowers dulled to a kind of orange brown rust. What the hell is going on?</p>

<p>my car in srgb and adobergb. Both converted from prophotorgb. the srgb histogram clips in the shadows (so I guess the image is lighter).</p><div></div>

<p>and adobe rgb</p><div></div>

<p>Tim. I was the FAA Laison with NASA for a period of time. I believe, if you gave them the project of recommending standards and technical solutions including ergonomics for simplifying color managment throughout the process they could do it. I think it needs to be done. But there needs to be industry consensus to move in that direction. What struck me that they, the engineers, found color managment too complicated. As I said above I certainly agree. <br>

As far as presenting scientific argument, IMO after several years of experience in FAA research programs, I believe it has to be structured, assess variables, defend assumptions, and be logical. The structure should clearly define the problem, the factors bearing on the problem(assumptions, variables etc.), offer proof, discuss alternatives, come to findings and make recommendations. The writing needs to be lucid, tightly edited and not overly complicated so as to convince those decision makers who may not have the same level of familiarity with the subject as the presenter. <br>

As far as your recommendation, I do keep my RAW files pristine. Any modifications I make are saved in another format. </p>

<p>

<p>For most of the photos I take it would make no difference whether I use sRGB or Adobe 1998, this is because most of my photos don’t have colors outside the gamut of sRGB.<br>

When I do have am image that goes outside the sRGB gamut I get a warning that this is going to have in the form of clipped shadows for one or more colors.<br>

As an example the boat in this raw photo has a lot of green and blue in it but very little red, so little that the reds will clip to zero if I convert to sRGB, not so in Adobe RGB 1998.<br>

Sewcon.com/IMG_1468.CR2</p>

<p>Since my monitor is pretty much limited to the sRGB color space I can’t really see the color of the boat as it really is, when viewing on my monitor. If I print using a printer that has a wider gamut, particularly into the green, then I will be able to make use of the wider gamut that sRGB provides.<br>

Since this image was going on the web and not being printed I am pretty much stuck with sRGB, in which case I can live with the reds clipping to zero in parts of the image or de-saturate the blue-greens in the photo.</p>

<p>Since I have my raw files I can of course go back to any photos I which and convert them into any number of color spaces. If I get a monitor that has a larger gamut sometime in the future I might well go back and convert relatively small number of photos that I have that don’t fit inside the sRGB gamut.</p>

</p>

<blockquote>

<p><em>Converted to sRGB and the preview of the orange flowers dulled to a kind of orange brown rust. What the hell is going on?</em></p>

</blockquote>

<p>hmm, I don't know: clipping? Are you Assigning sRGB or Converting to sRGB?</p>

<p>Dick,</p>

<p>That what sound advice about the need for organization toward a standard. I couldn't agree more. Maybe you should talk to the folks at ICC organization, Adobe, Apple and Microsoft to work together as methodically as you just lined out. Sounds like you do boardroom/administrative and general top down speak quite well way beyond my capabilities.</p>

<p>Indraneel,</p>

<p>Good grief that's worse than I get converting to a matrix/math based color space. Yours looks like you converted to a look up table printer based color space. Frankly they both look good, but I those bushes in the background look kind of odd. Maybe it has something to do with downsampling to the web and what it does to the appearance of sharpness which looks kind of crispy only in the background.</p>

<p>Ellis,</p>

<p>I'm just opening in ACR 4.6 in ProPhotoRGB, hit Auto which pulls the ends in to eliminate clipping in the histogram and then I convert to sRGB. I may have found a workaround though.</p>

<p>What I do is convert to sRGB then go into HSL and correct hue and saturation on yellows and oranges and adjust color temp to get pretty close to what I had in ProPhotoRGB and then convert back to ProPhotoRGB while still in ACR. It's kind of a cludgy editing process for controlling gamut clipping Soft Proofing, but I never thought I'ld ever see this happen in a matrix/math based color space but then I don't shoot a lot of flowers. I just started doing this to see if I can push the gamut with processing.</p>

<p>Another way I do to fix this is to switch to the DNG PE Color Chart Wizard profile I made which makes images noticeably richer and saturated with oranges taking on a deeper orange and yellows unfortunately taking on a slight cyan cast.</p>

<p>Here's the YouSendit link to the original orange flower raw PEF off my Pentax K100D:</p>

<p>http://www.yousendit.com/download/dVlwd0VFNXY1R05MWEE9PQ</p>

<p>See if you get shifts converting to sRGB after hitting Auto or make the image look good. It's a little soft because I was testing out the diopter on my camera trying to focus in low light shooting macro and I didn't quite get absolute sharp focus.</p>

<p>Yesterday I read that color management slows down Firefox by 40%.  Is it worth it?  Not for me.  But if you're interested, here's how:<br>

Type <b>about:config</b> in the address bar of Firefox 3. Click thru confirmation page and find <b>gfx.color_management.enabled</b>. Double click that until it says "true". Restart Firefox.<br>

Indraneel's red Toyota images are a case in point. The sRGB image is better because it is more saturated. However the JPEG artifacting is so bad, probably another case of Photoshop abuse, that I cannot believe that you guys are arguing over stupid things instead of what essentially more important things such as JPEG encoding.</p>

<blockquote>

<p>Yesterday I read that color management slows down Firefox by 40%.</p>

</blockquote>

<p>I'm not seeing that at all. Feels the same as always.</p>

<p>I created <a href="00T4S0">a new thread over here</a> to move this discussion a little more towards actionable advice instead of color theory. Please check it out and if you have a color-managed workflow you like, post about it.</p>

<p>Thanks.</p>

<blockquote>

<p>The orange is mainly made up of 195,120,20 in ProPhotoRGB.</p>

</blockquote>

<p>Tim,<br>

More than orange the color is a pink.<br>

sRGB values are 254,116,138. So the color is internal to sRGB.<br /> <br /> There is no way to have any shift.<br /> <br /> I suspect, another time, that the ProPhoto to sRGB conversion is wrong.</p>

<p>Jacopo,</p>

<p>When I convert ProPhotoRGB orange=195,120,20 to sRGB I get 255,116,0 as read in CS3 info palette. The preview of this color fill slightly darkens and dulls after the conversion.</p>

<p>I don't get pink or the sRGB numbers you indicated.</p>

<p>Tim,<br>

sorry, I converted ProPhoto (195,120,120) = sRGB(254,116,138).<br>

And the color is pink.<br>

ProPhoto(195,120,20)=sRGB(255,116,0). Right!<br>

And is orange.<br>

But I don't see any color shift.</p>

 

<p>Tim,<br>

some numeric values using your monitor profile:<br>

- transforming ProPhoto(195,120,20) into sRGB, you get (255,116,0)<br>

- transforming ProPhoto(195,120,20) into your monitor profile, you get (255,115,0)<br>

(same as the to sRGB transform)<br>

But:<br>

- transforming sRGB(255,116,0) into your monitor profile, you get (249,114,39)<br>

That color is the "the orange flowers dulled"</p>

<p>I suppose that putting sRGB as your monitor profile you can get a better rendition.</p>

<p>jacopo,</p>

<p>Can you and anyone else see a color difference between the ProPhotoRGB preview vs the converted sRGB preview in the image below? If so, this is the shift I'm getting. This image needs to be viewed in a color managed app.</p>

<p>And if you want to see differences editing in wider color spaces other than sRGB, convert copies of this image to sRGB, ProPhotoRGB and AdobeRGB and apply a saturation boost in Photoshop. The image is currently in my i1 Display profile so you'll have to convert to each color space to do this. On my system there's no change to the orange flowers boosting saturation in sRGB but there is in ProPhoto and AdobeRGB. See if you get the same results.</p><div></div>

<p>Tim,<br>

yes I see the difference.<br>

But I explained from where you have it: you have a monitor profile problem, specially in the blue channel for that colors.<br>

If you increase the saturation, or reduce the blue color, you can get in sRGB an image that looks like the ProPhoto image. But this is not the way.<br>

Have you tried the sRGB profile as your monitor profile?</p>

 

<p>No I haven't but I'll give it a try. I'll also probably go ahead and recalibrate and see if that changes anything. </p>

<p>When I increase saturation in sRGB nothing is affected. It's like the Hue/Sat slider is broken or something because nothing happens. I thought the orange color was out of gamut for sRGB but since you mention you can change the orange in sRGB then something must be amiss with my profile.</p>

<p>Great! Something else I have to deal with.</p>

<p>Thanks for the responses and help you've given to me. Much appreciated.</p>