Options
Mis tagged color spaces ARGB vs sRGB or Hey, am I in a pickle?
NeilL
Registered Users Posts: 4,201 Major grins
Most recently I have been doing RAW conversion in CaptureOne5Pro and exporting to CS3. Just tonight I noticed that CO5P was converting into Adobe RGB (I missed changing the default). The PS colorspace is sRGB! So, all my recent stuff has been edited in PS as Adobe RGB in a sRGB colorspace.
What are the terrible consequences of THAT???, could someone please tell me the bad news and get it over and done with!
Thanks (NOT)!:D
Neil
What are the terrible consequences of THAT???, could someone please tell me the bad news and get it over and done with!
Thanks (NOT)!:D
Neil
"Snow. Ice. Slow!" "Half-winter. Half-moon. Half-asleep!"
http://www.behance.net/brosepix
http://www.behance.net/brosepix
0
Comments
Probably nothing to horrendous, as it sounds as if you just moved into a smaller editing space. Things would have been worse if PS was assigning an erroneous color space, but I'm guessing it just converted you to the smaller space.
However, you might want to go into PS and choose Edit -->Color Settings and make sure that "Profile Mismatch/ Ask when opening" and "Missing profile/ Ask when opening" are checked. This will alert you when a file comes in that has a different profile than your working one and give you some options for correcting the mismatch.
Yes, understand. Thanks, Mark.
In fact, I do have those options checked, but didn't get any alert when working on my own images. Nothing in the look of the editing I was doing gave me a clue that there might be anything wrong, either. However when I tried to blend into one of my images a stock background, I *did* get just that alert. That's how I discovered that my images had been coming out of CO5P in Adobe RGB, because the stock background wasn't Adobe RGB.
I know that it's not good for the quality of an image to be moving in and out of different color spaces, so part of my worry was what happens when these Adobe RGB images go to the printer? And how are they going to display on the web on displays using sRGB?
First time I've faced this situation of 'mixed' color spaces.
Neil
http://www.behance.net/brosepix
Neill, why not bring your tiffs from C15P to CS3 as 16 bit tiffs in the largest space available to you - ProPhoto if C15P supports it, and I would suspect it does. CS3 certainly does.
I do not reduce my files to 8 bit sRGB until I am ready to upload them to the web. Otherwise they remain in 16 bit larger color spaces. All it costs is some disc space which gets cheaper every day.
Converting many files in AdobeRGB to sRGB will result in the loss of some hues, but frequently not enough to notice when the shots are of people and do not involve strong color tones. May be more noticeable in blues and greens, but these are not tones strongly present in portraits for example. Landscapes the change will be more discernable, depending on what printer renders the print.
This link concerning the AdobeRGB versus sRGB I found quite helpful, and in line with my thinking as well - http://www.cambridgeincolour.com/tutorials/sRGB-AdobeRGB1998.htm
Here is more discussion of the ways AdobeRGB and sRGB differ - http://www.earthboundlight.com/phototips/srgb-versus-adobe-rgb-debate.html
Bear in mind that most monitors only display sRGB size gamuts; if you own a wide gamut monitor, you may be able to see AdobeRGB images IF your software supports ICC tags, but some software does not. PS and LR certainly do, as well as many, but not all, browsers.
Moderator of the Technique Forum and Finishing School on Dgrin
Thanks, Jim, this sheds some new light.
I always do convert to 16bit tiff, but when I went a-checking yesterday what color space was set in CO5P I found out it was Adobe RGB.
So since I've been converting with CO5P the files I've sent to CS3 have been in Adobe RGB while CS3 is set to use sRGB. The effects and repercussions of that are what concern me, for editing itself, but also eg I have saved edited versions of some of those files as 8bit tiff, and jpegs. What color space information do they contain for displays and printers, and will viewing and printing as intended be buggered (Australian term for unworkable)?
CS3 didn't warn me those files were coming in in Adobe RGB, and nothing happened in the editing that tipped me off that there was the mismatch, until I started to blend in a stock photo DL from web as BG to one of my edited 16bit tiffs.
How do I see what color space a jpeg is in if I open it in PS?
I have set things straight for the future as regards the color space I am converting into, but is the stuff that has got through at any risk because of the two different color spaces that got mixed?
Sorry to be so longwinded. Would be grateful for anything else you might like to add to what you've said.
And thanks very much for the links!
Neil
http://www.behance.net/brosepix
As Mark suggested, hit Edit>Color Settings and you will see the following dialogue box, and how I have mine set. As Mark suggested, you want to have Profile Mismatches, and Missing Profile boxes checked so that CS will call the mis-matches to your attention. The dis-advantage of checking the boxes like this, is the nanny state will harrass you about each image you open unless it is in your working color space, which as you can see for me, is ProPhoto 16 bit.
You can always see what tag is associated with an image in PS, by hitting Edit>Convert to Profile and seeing what the source document space is. You can then just cancel to exit that dialogue box. Converting to a smaller space, via that command, will not usually display much change at all, since the date is remapped in that newer space. The Intent box offers different modes of mapping the out of gamut data to the newer space. I always use the Adobe conversion engine, not Apple's.
As for what your edited files are tagged as, that depends on how you save your files from Photoshop. If you use "Save for Web" they will not have a color space attached, as Save for Web removes the tag.
'Save As' allows you to choose the file type gif, jpg, jpg 2000, tiff psd etc, and include color space tag as well - this is a screen shot ( shot with a 50D, not a screen capture, because I could not capture the open dialogue box portion offering jpgs, jpg 2000, px, tiff, psd, png, etc with screen capture ) and you can see the box the offers - Embed Color Profile sRGB. The offerings of the check box alter as you select jpg, png, pcx, tiff etc.
As for you images that were converted from AdobeRGB in your camera to sRGB in CS3, they are a bit smaller gamut, but may not be seriously damaged. Personally, if they looked fine to me, I would ignore it. If you think they do not look right, you have the opportunity to go back and re-edit them.
There is no way to convert an image from sRGB BACK to Adobe RGB and recover the original exact Adobe RGB file you started with. The trip from ARGB -> sRGB is a one way trip - you cannot take the converted sRGBs -> Adobe RGB and have the original file. You can do this round trip, but the file you end up with is not the one you started with. But the sRGB file may well be close enough to your original Adobe RGB to not be a serious issue. Only the viewer/owner can determine this.
If you have any further questions, or if my explanations are as clear as mud, I will be happy to try to explain this further. Hopefully this information will help you in the future.
Moderator of the Technique Forum and Finishing School on Dgrin
Great reply, Jim, thanks very much. Very helpful and comprehensive. I'll use it to go back and go through my settings, and also check the specs of the edits I've saved.
Neil
http://www.behance.net/brosepix
Moderator of the Technique Forum and Finishing School on Dgrin
A former sports shooter
Follow me at: https://www.flickr.com/photos/bjurasz/
My Etsy store: https://www.etsy.com/shop/mercphoto?ref=hdr_shop_menu
Bill, I have a confession to make - I went back and checked my color spaces for CS4, and they were NOT like I have them in CS3. I wonder if we upgrade sometimes, and forget to reconfigure our color settings. It seems that I did, exactly that. My CS3 was correct, but not my CS4....... I'll bet I am not alone, too!
Moderator of the Technique Forum and Finishing School on Dgrin
IF you didn’t do any work yet in Photoshop, just re-render out the raws to TIFF in a larger color space. Everything should be as you had it prior to the sRGB export.
That said, having sRGB isn’t a deal breaker but its far from ideal considering the source data (raw). Now if all the images were shot of a scene with a low gamut (bride in white wedding dress), its likely a larger encoding space than sRGB will buy you nothing. Sun sets at the beach, colorful flowers, well yes, you did toss some gamut away but again, not a huge big deal. But now, set it for ProPhoto RGB 16-bit and get out all the possible color the scene and capture device could handle.
Author "Color Management for Photographers"
http://www.digitaldog.net/
Hi Andrew, understand, and thanks.
This is intriguing. I noticed that Jim uses the ProPhoto RGB as his working color space. Would you mind just running through how I would set it all up - CaptureOnePro5 has the option of converting into ProPhotoRGB, and CS3 has the option of ProPhoto as a working space, so I set them both to that? What about displays set to sRGB, and printers, how is the ProPhoto profile's compatibility with the most common working space people use? I wouldn't want, after editing, to get into converting into different color spaces for different purposes.
Neil
http://www.behance.net/brosepix
I don’t use C1 but suspect somewhere there is a preference to define the encoding color space you wish from the rendered raw data. Then in Photoshop, set the RGB working space to ProPhoto, that way you will not get any kind of warnings (profile mismatches) when you open those documents into Photoshop. Any other document that has a differing color space will pop a dialog to remind you that you prefer ProPhoto RGB but the current document is in another color space. If you keep the polices set to Preserve, just hit the OK button and the document will open honoring whatever color space that data is in. The screen capture above looks good to me (although you’ll probably want that 3rd check box, paste mismatch on as well).
The display profile is totally separate from how you edit your images. Once you calibrate and profile your display, and assuming the system is accessing it, there is nothing more you have to do.
As for the differences in ProPhoto versus other spaces, this might help:http://www.adobe.com/digitalimag/pdfs/phscs2ip_colspace.pdf
Author "Color Management for Photographers"
http://www.digitaldog.net/
You know, I just went into PS and had a look at those warnings - they were all unchecked. Now I didn't do that, so somehow something is causing PS to drop those settings. I'll keep a sharp eye on that for a while.
Neil
http://www.behance.net/brosepix
Yes, good, thanks. I'll have a look at that link. I might have misled you, I meant once all the editing is done and saved, the images will be in the ProPhoto color space. Will that need to be converted into another color space eg sRGB to display properly on most monitors, or to print on most (home) printers, and the like?
Neil
http://www.behance.net/brosepix
It would be great if someone could post an example of the differences.
One could argue that at some other point the image may be output to a space that is larger than sRGB or that a perceptual rendering intent could be used - so it is better to keep the larger gamut than to clip it from the start.
Another advantage of working in a wide gamut space such as ProPhoto RGB when the output is sRGB is that it provides more "room" for saturation expansion, even more so when the saturation is performed using poor technique (normal blend mode). Using good technique (color blend mode), it matters less. Another good technique is to use a saturation mask to protect higher saturated areas from increased saturation. Using a method such as "vibrance" has built-in protection against pushing near saturated colours beyond the limit of the working colour space.
Richard, you wish to see an example...here you go (yellow and saturated yellow hues are quite tricky, 100% yellow in good CMYK print conditions is often out of gamut for many RGB spaces!):
http://members.ozemail.com.au/~binaryfx/howto_safesaturation.html
The choice of editing space is a tough issue. Some folk simply pick the largest space available, even if the image content does not need such a large space and the inefficiency that goes along with it. Other folk pick a working space that is large enough to contain the image gamut, this space will vary from image to image.
The working/editing space is *independent* of the output space, so it does not matter if different images have different source profiles.
One should keep in mind that when one converts from a larger space to a smaller space, using common RGB matrix based working space profiles - the conversion is *relative colorimetric*. What does this mean? No gamut compression, the wider gamut colours are *clipped*. Even if one selects the perceptual intent when converting, if the profile does not contain a perceptual intent it will use relative colorimetric!
Note that there is a "beta" ICC v4 sRGB profile that does contain a perceptual table for gamut compression, however it appears to suffer hue shifts.
Another option is to use "Photo Gamut RGB" as an intermediate space, when converting from a larger space to a smaller space. ProPhoto RGB > Photo Gamut RGB (perceptual intent) > sRGB. Once again, the hue shifts from using Photo Gamut RGB may mean that the user does not like the result, despite this profile offering a true perceptual rendering intent.
Stephen Marsh
http://members.ozemail.com.au/~binaryfx/
http://prepression.blogspot.com/
http://members.ozemail.com.au/~binaryfx/
http://prepression.blogspot.com/
I see that my dialogue box does not have the paste mismatch box checked; yet, whenever I try to move a ProPhoto selection to a layer on top of a layer of sRGB, Photoshop does warn me about the mismatch profiles.
I will check that box also Andrew. Thank you for pointing that out.
Moderator of the Technique Forum and Finishing School on Dgrin
If I understand your question, Richard, you are suggesting that a ProPhoto file cannot really be displayed on my Apple Cinema display, even with it being properly profiled by an i1D2, because the gamut of the Cinema Display CANNOT display ProPhoto gamuts. That is true, yet the image I will finally see will not be in ProPhoto, but will be an 8 bit file in AdobeRGB or sRGB. My monitor can only really display sRGB, so why bother with 16 bits at all.
One reason for me, is that computer generated gradients, like a mask created with the gradient tool, can be posterized in some 8 bit spaces, whereas I never see that in 16 bit ProPhoto.
Since most of my global editing is done in Lightroom2 ( which incidentally is also ProPhoto color space ), editing I perform in Photoshop, usually involves fairly extensive selections, gradients, and/or pastings. Hence my choice of 16 bit ProPhoto as my working space.
I am aware of my working space, and make sure I always go from a larger space to a smaller space, and never go backwards - well, almost never. I rarely venture into CMYK these days. I know someone will chastise me for that, but what can I say......
Moderator of the Technique Forum and Finishing School on Dgrin
Thanks for the illustration, Stephen.
I never use the Saturation tool in CS3. I have tried it a few times, but always end up discarding it because of its ugliness even using saturation/color blending modes.
Yes, of course working space is independent of output space, but the worrying area for me is the transition between the two. My philosophy has always been to keep them the same, ie sRGB. But maybe there is some confusion in my thinking, so let me ask: if an image is in ProPhotoRGB on my HDD and I show it on my sRGB display, does a conversion of the color space data take place (from ProPhotoRGB to sRGB) to accomplish that, equal to an on-purpose conversion of the file from ProPhotRGB to sRGB color space in PS?
As I said before, I really want to avoid exchanging color space currencies under any circumstances if at all possible.
Neil
http://www.behance.net/brosepix
That is the same experience I had with the stock BG I was attempting to blend: alert was unchecked (unknown to me) but appeared nonetheless.
Neil
http://www.behance.net/brosepix
Just another point, Stephen. How can one set perceptual intent in a profile (do you mean a calibration color profile here, or color space?)?
Neil
http://www.behance.net/brosepix
That argument only works if the content of the container (the container being ProPhoto RGB) is larger than the gamut of the small Cinema display. You can capture a gray card and encode it into ProPhoto and its going to fully fall within sRGB. A utility like ColorThink from Chromix is indispensable because it allows you to plot actual image gamuts over color spaces defined by profiles.
Now we see why wide gamut displays, for those that work in wider gamut scene capture, working space and output spaces are useful. But unless your only output is a display, it doesn’t matter that much. The proof is in the print. If you funnel everything to the lowest common dominator (your sRGB display), you’ll use far less of the data available everywhere else.
The wider the gamut color space, the more bits become useful. Spreading the same encoding values over a larger distance means the spaces between them is thus larger.
And there are plenty of true CMYK output devices (devices that require you send them a CMYK document) that exceed sRGB.
Because it might go somewhere else. In fact, unless you are sending documents to the internet or other displays, well thats the only sRGB output device you’ll find. There is no such thing as an sRGB printer.
Author "Color Management for Photographers"
http://www.digitaldog.net/
Nearly all the RGB working spaces we find installed by our software products are called Matrix profiles. That’s why they are so small (about 4k) which is a good thing considering they should be embedded in our documents. They only have one table for a rendering intent called Colorimetric. Relative and Absolute share this table. More complex LUT based profiles, those that can easily be 1mb or more, have 3 tables including the Perceptual table. When you use a Matrix profile in say Photoshop, it offers you Perceptual but no such table exists so you always get Colorimetric.
There are a few newer (v4) working space profiles that contain and utilize a perceptual table and do so with a more “robust system” called PRMG which as yet, isn’t well supported in many applications. Some would suggest this will produce “better” conversions with more control when dealing with working space conversions but until we see this supported in mainstream apps, its probably not worth worrying about.
Author "Color Management for Photographers"
http://www.digitaldog.net/
Precisely - Even though the gamut of ProPhoto is vastly larger than my Display's gamut, images that will end up on paper, will be limited to the ink and illumination limits. Paper is not ever going to offer the brightness levels of monitors. A grey card can be rendered in a smaller color space, because it never was near the gamut's edge in the larger, let alone, the smaller color space.
You stated precisely the point I was trying to make above. Thank you.
This is the argument between Adobe RGB and sRGB again - larger gamut with the same number of steps of luminosity, means bigger steps between each step. Also do we really need the colors offered by AdobeRGB vs sRGB for most 'typical' images - say a portrait versus a landscape. See here - http://www.cambridgeincolour.com/tutorials/sRGB-AdobeRGB1998.htm
I am sure there are, but my comment was a joke, not a serious statement for or against CMYK. Just trying to keep the tone light here.
MY comment was just by way of trying to explain to Richard, why I choose to use ProPhoto as my working color space in CS4, even though I do not have a monitor, or an output system that can really render that extreme gamut on paper or a typical monitor.
I know that you do not suggest using only 8 bit color spaces for most image editing either.
Moderator of the Technique Forum and Finishing School on Dgrin
I host my pics on SmugMug and print them through SM as well. And that's it. So I think what I'm hearing is that there really is no advantage for me in what I actually do, but only in what I might possibly do in the future. Right? I understand that other people's mileage may vary, but that doesn't matter to me.
It depends on the scene of course. To know for sure, you’d have to shoot the scene, process it (ideally from raw), render as you prefer in several color spaces then plot them in ColorThink. Obviously except as a science experiment, no one is going to shoot that way. If the scene is low gamut, you would be fine in an sRGB encoding but what if it exceeds sRGB and (not hard to find), Adobe RGB (1998)? Might as well just work in the wider gamut working space and move on. Its like having a 21mp capture device but half the time, you use the images for the web. No harm done but if you shoot with a 1mp camera, you’re out of luck if you want an 8x10 print, let alone something much bigger.
Author "Color Management for Photographers"
http://www.digitaldog.net/
Well if you’re sure for the rest of your life, that’s the only output device OK. And no, they don’t have an sRGB printer either, no such product exists.
The question would be, what would be the advantage of throwing away data from the get go? Seems like you might be painting yourself into a corner. An sRGB only workflow will not produce poor results! Its not like a wider gamut space will be hugely better (depending on who’s making the assessment here). The fact remains you started with more colors and toss them away and the question would be, depending on your “workflow”, why do this if you can have the best of both worlds.
IF you shot JPEG in sRGB on your camera, fine. If you shoot raw, there’s no colorspace there until you specify one. Why funnel that into sRGB? In a product like Lightroom, there is zero extra work in sending data in sRGB to the web while retaining all the additional data for other uses. Its not heavy lifting.
Author "Color Management for Photographers"
http://www.digitaldog.net/
Well, I always keep my raw files, so on the off-chance I want to reprocess something in the future because of new needs or possibilities, I can do so. I guess what I am trying to avoid is spending time in editing only to find that my pics look different than I expect once they are online. Right now, I only need to soft-proof when I am going to print.
That they demand sRGB has zero bearing on the gamut of the output color space. They could just as easily ask for ProPhoto or Adobe RGB (1998).
They shouldn’t look any different. In Lightroom, you’re not viewing the data in any of the big three spaces. You export to a web gallery and view that in an ICC aware browser, it looks 99% like what you saw in LR.
Take your raw and encode it into ProPhoto RGB, sRGB and view the two in say Photoshop. You see a big difference?
Author "Color Management for Photographers"
http://www.digitaldog.net/
Just bumping this in the hopes of clarification.:D
Thanks.
Neil
http://www.behance.net/brosepix