Luminous Landscape Forum
Raw & Post Processing, Printing => Colour Management => Topic started by: Dave Gurtcheff on January 21, 2013, 10:26:05 am
-
In the past I have used Adobe RGB. My calibrated and profiled monitor produced prints I am happy with. When posing to my web site, I did not change to sRGB, but left them as is and changed them to low Res JPGs. They look fine on my web site. Now my questions:
I am readinf Jeff Schewe 's Digital Negative book, which highly recommends using Prophoto RGB for prints. I am switching to that color space. But he also says this color space is not good for web use, and recommends sRGB. I am not sure how to make a small jpg copy, and change its color space. If I reduce in size a COPY of the Prophotob RGB image the use SAVE FOR WEB, will that change color space and work OK? Next question: I want to have a press printed, bound large coffee table book made. What color space should I use, and how do I change it in Photoshop?
Thanks
Dave......still homeless on NJ coast after Sandy
-
If I reduce in size a COPY of the Prophotob RGB image the use SAVE FOR WEB, will that change color space and work OK? Next question: I want to have a press printed, bound large coffee table book made. What color space should I use, and how do I change it in Photoshop?
Save for Web will convert to sRGB so that's fine. Or you can just use Convert to Profile and select sRGB. Same results.
For the Coffee table book, much depends on the printer and how the images are sent. For example, if you used Lightroom, to Blurb, you'd be sending sRGB through that process automatically. If you were using say Aperture, it uses Adobe RGB (1998). If you have to send the book off yourself, you could ask the printer.
As for why ProPhoto (and why sRGB), this might help:
Everything you thought you wanted to know about color gamut
A pretty exhaustive 37 minute video examining the color gamut of RGB working spaces, images and output color spaces. All plotted in 2D and 3D to illustrate color gamut.
High resolution:http: //digitaldog.net/files/ColorGamut.mov
Low Res (YouTube): http://www.youtube.com/watch?v=n0bxSD-Xx-Q
-
Thank you Andrew. That is very helpful. I will watch your video.
Dave
-
In the past I have used Adobe RGB. My calibrated and profiled monitor produced prints I am happy with. When posing to my web site, I did not change to sRGB, but left them as is and changed them to low Res JPGs. They look fine on my web site. Now my questions:
I am readinf Jeff Schewe 's Digital Negative book, which highly recommends using Prophoto RGB for prints. I am switching to that color space. But he also says this color space is not good for web use, and recommends sRGB. I am not sure how to make a small jpg copy, and change its color space. If I reduce in size a COPY of the Prophotob RGB image the use SAVE FOR WEB, will that change color space and work OK? Next question: I want to have a press printed, bound large coffee table book made. What color space should I use, and how do I change it in Photoshop?
Thanks
Dave......still homeless on NJ coast after Sandy
Dave,
Sorry to hear that you are homeless after Sandy, and hope things work out for you shortly.
If you use Lightroom, you don't have to worry about color space until you export, since the image is not actually rendered into a color space until that time. The internal working space in LR (and ACR) is ProPhoto primaries with a linear tone curve.The readouts and histograms in LR are in Mellisa, which uses the same primaries but an sRGB tone curve. ACR lets you choose among several color spaces and the histograms and readouts are in that space. When one prints from Lightroom it is not necessary to render into ProPhotoRGB and one can send the image directly to the printer rendered into a printer profile from which on can use rendering intents.
Much of the web is not color managed, but if you have a select audience, you can use files tagged with AdobeRGB or ProPhotoRGB and tell your audience to use a color managed browser (such as Firefox or Safari. The latest version of Internet Explorer is also color managed. See this ICC post (http://www.color.org/version4html.xalter). Otherwise, you should use sRGB for the web. If you save to the web in PS or export for the web in LR, out of gamut colors will be clipped, since perceptual rendering is not available (it is listed in Photoshop, but the setting is ignored). If you image has no out of gamut colors, things will be OK.
If you would like to try perceptual rendering, you might check out Ver 4 sRGB. See the ICC site (http://www.color.org/srgbprofiles.xalter#v4pref) and read White Paper 26 in the link. The color gurus do not seem to discuss Ver 4 other than to say that it is not ready for prime time.
Bill
-
ACR lets you choose among several color spaces and the histograms and readouts are in that space.
ACR shall have melissa for color readout regardless of the choice of the output space, isn't it so ?
-
ACR shall have melissa for color readout regardless of the choice of the output space, isn't it so ?
Until you turn on soft proofing (type S key in Develop).
-
ACR shall have melissa for color readout regardless of the choice of the output space, isn't it so ?
Camera Raw reads out in whatever color space is set in the Workflow Options...only LR uses Melissa RGB for readouts.
-
As for why ProPhoto (and why sRGB), this might help:
Everything you thought you wanted to know about color gamut
A pretty exhaustive 37 minute video examining the color gamut of RGB working spaces, images and output color spaces. All plotted in 2D and 3D to illustrate color gamut.
High resolution:http: //digitaldog.net/files/ColorGamut.mov
Low Res (YouTube): http://www.youtube.com/watch?v=n0bxSD-Xx-Q
That's a terrific video, Andrew. I especially liked the 3D plots of specific photos compared to sRGB, aRGB and ProPhoto. Well Done!
-
That's a terrific video, Andrew. I especially liked the 3D plots of specific photos compared to sRGB, aRGB and ProPhoto. Well Done!
Agreed! A lot of effort went into that, and we should all be damned grateful! :)
Andrew, now you've given me all kinds of high concept notions about better ways to edit imagery for gamut issues. For instance, wouldn't it be nice if Adobe software could give us some hints about exactly in what way those drab gray out-of-gamut scabs were actually suffering.
-
Until you turn on soft proofing (type S key in Develop).
but I am talking about ACR, I do not have LR
-
Camera Raw reads out in whatever color space is set in the Workflow Options...only LR uses Melissa RGB for readouts.
thank you, I think I was confused by some old postings from somebody whom I did not think to doubt and did not notice that it was about LR (and did not check myself - shame on me).
-
Save for Web will convert to sRGB so that's fine. Or you can just use Convert to Profile and select sRGB. Same results.
For the Coffee table book, much depends on the printer and how the images are sent. For example, if you used Lightroom, to Blurb, you'd be sending sRGB through that process automatically. If you were using say Aperture, it uses Adobe RGB (1998). If you have to send the book off yourself, you could ask the printer.
As for why ProPhoto (and why sRGB), this might help:
Everything you thought you wanted to know about color gamut
A pretty exhaustive 37 minute video examining the color gamut of RGB working spaces, images and output color spaces. All plotted in 2D and 3D to illustrate color gamut.
High resolution:http: //digitaldog.net/files/ColorGamut.mov
Low Res (YouTube): http://www.youtube.com/watch?v=n0bxSD-Xx-Q
Andrew, just watched your Youtube video and found it quite concise and professionally presented. I understood everything. Thanks for posting it. Just had some questions...
I'm a bit puzzled by your associating the camera's capture capabilities with images you've chosen that clearly appear to have their saturation cranked up a bit I'm assuming for demonstration purposes. IOW the images are edited to look not according to what the camera actually saw going by their lack of realism and by realism I don't mean scene referred either.
Now I have to wonder how much of the colors mapped within your ProPhotoRGB vs sRGB comparison are actually measuring the gamut of the display's ability in allowing the colors to be pushed that much. That's not proof of what the camera is delivering.
I can crank up saturation on any broadcast scene piped through an old CRT TV hooked to a cable feed way more than what you showed in your LR saturation slider "granularity" demo in illustrating hitting the display's gamut ceiling. Does that mean the broadcast camera and CRT TV has captured and delivered far more gamut than your camera and wide gamut display? I'm having a problem understanding that one.
Were those low rez images used for gamut analysis in ColorThink measuring the Raw data or over cooked jpegs? tiffs?
I realize your video wasn't focusing on camera capture capabilities so I would hope you could create another video that would demonstrate the scene gamut capture capabilities of a camera using real world wide gamut scenes without making them look artificial in determining just exactly how much data is truly captured by a digital camera and what can get thrown away converting to and/or editing in sRGB.
Again, very helpful and appreciated work on that video. Hopefully it'll reduce the number of threads from people insisting on editing their images in sRGB.
-
Really Great video! Thanks for that.
It does lead me to ask more questions about what happens to those OOG pixels when they get printed. Can anyone suggest a resource that explains the proper strategies for dealing with the out of gamut colors when printing.
Thanks
Stuart
-
It does lead me to ask more questions about what happens to those OOG pixels when they get printed. Can anyone suggest a resource that explains the proper strategies for dealing with the out of gamut colors when printing.
Either Andrew's book or Real World Color Management...or a certain book that is, as yet, unreleased...
Out of gamut colors are handled either by clipping (Relative Colorimetric) or by gamut compression (Perceptual) Rendering Intents.
RelCol is a standard, Perceptual is subject to the profile creator or the CMM (although we've yet to find a smart CMM can can be modified on color handling).
Andrew did a great job on the video but the compression of gamuts is still a "weird science" and proprietary to the color management companies (and how they create profiles). i1 Profiler does a good job of gamut compression with Perceptual Rendering but they don't really document how they do that. I'm personally convinced that the color management companies do a lot of hand work and special tuning to create their own Perceptual rendering intents...with weighting and bias to certain colors while letting other colors kinda do "whatever".
-
It does lead me to ask more questions about what happens to those OOG pixels when they get printed. Can anyone suggest a resource that explains the proper strategies for dealing with the out of gamut colors when printing.
Soft proof each image and pick the rendering intent you visually prefer. Let the profile do the work (if it's a good profile).
(http://digitaldog.net/files/Gamut_Clipping_vs_Compression.jpg)
In the illustration, you can see the differences between gamut clipping (the Relative Colorimetric intent) and gamut compression (Perceptual). Clipping is pretty simply, move all OOG colors to just in gamut. Compression or Perceptual is more complex, takes lots of other colors that may be in gamut into account and attempts, that's the key word, to map the colors such they appear more perceptually affected. Everyone does it differently like every film manufacturer makes their films look different, hoping to please the customer. So you just have to look at the effect of the two intents on the image and pick which you prefer.
You can see that the Perceptual intent is far, far more complex than trying to control OOG using manual controls (desaturation sponge and the like). And with a certain profile, do you like the effect on the image?
-
Thanks, Jeff and Andrew. That helps a lot. I'll certainly be following up w some reading. I feel like I have to revisit color management theory/ best practices fairly regularly. ( I have Andrew's book, but it is out of date, 2005...)
It seems the more I understand, the less I know :)
Stuart
-
Digitaldog,
That is an excellent demonstration of gamut compression, but it does not show the limitations of current perceptual rendering schemes. The color management module has no knowledge of what colors are actually in the image, but applies compression whether or not it is needed. If there are no out of gamut colors, the colors in the image will be compressed and lose saturation. At other times, the compression may not be sufficient and out of gamut colors will remain in the image. The revised graphic shown attempts to illustrate these considerations.
Regards,
Bill
-
A pretty exhaustive 37 minute video examining the color gamut of RGB working spaces, images and output color spaces. All plotted in 2D and 3D to illustrate color gamut.
High resolution:http: //digitaldog.net/files/ColorGamut.mov
Low Res (YouTube): http://www.youtube.com/watch?v=n0bxSD-Xx-Q
a note on RGB Working Spaces...
There is a table based RGB working space called "Photogamut RGB". Technically it's designed like a printer profile - with a neutral Gamma 2.2 gray axis, though. The profile is designed to encompass printable colors (so its Gamut is pretty large in mid tones) but to avoid non-printable colors (Gamut is relatively small in bright tones). So your comments about RGB working space actually only apply to matrix based RGB working spaces ...
http://photogamut.org/E_idea.html
http://photogamut.org/E_ICC_profile.html
-
The color management module has no knowledge of what colors are actually in the image, but applies compression whether or not it is needed.
By this time, it doesn't even know the gamut or color space of the original! It gets Lab data. Now if the PRMG is ever fully developed, along with V4 profiles (unlike those we have today V4 profiles in V2 clothing), that could improve. Then we'd have to see Adobe and other major companies update and supply a family of V4 ICC profiles for this to take off. I'm not holding my breath.
-
In the past I have used Adobe RGB. My calibrated and profiled monitor produced prints I am happy with. When posing to my web site, I did not change to sRGB, but left them as is and changed them to low Res JPGs. They look fine on my web site. Now my questions:
I am readinf Jeff Schewe 's Digital Negative book, which highly recommends using Prophoto RGB for prints. I am switching to that color space. But he also says this color space is not good for web use, and recommends sRGB. I am not sure how to make a small jpg copy, and change its color space. If I reduce in size a COPY of the Prophotob RGB image the use SAVE FOR WEB, will that change color space and work OK? Next question: I want to have a press printed, bound large coffee table book made. What color space should I use, and how do I change it in Photoshop?
Thanks
Dave......still homeless on NJ coast after Sandy
Off topic.
Dave, sorry to hear about Sandy affecting you. Something that is very important for people in that situation is to remove any wet insulation. It will hold water breeding fungus. Besides the health issues of that alone, that fungus will then go to work on the structure of the home. If the wood can dry it will be fine. If its kept wet the whole house will need to be torn down sooner. A house with a 100 yr life expectancy might become a 5-20 yr life. That is a lot more money than the cost to replace the insulation.
-
please what is "PRMG"?
-
please what is "PRMG"?
PRMG is Perceptual Reference Medium Gamut. Since in the past, conversions using a perceptual intent were not well defined with respect to the source gamut prior to making its way into the PCS, this is supposed to be a less ambiguous color space (is that the correct term?, I would think it is) rather than some vendor specific gamut assumption and one that is unknown or defined as far as the user is concerned.
-
please what is "PRMG"?
Probably the Perceptual Reference Medium gamut. Info from the International Color Consortium here http://www.color.org/v4_prmg.xalter
Edit: oops, typing while you replied, Andrew. BTW, great video, thanks. The Colorthink plots of images vs color spaces are very instructive.
-
Thanks to the both of you! - Hening.
-
A pretty exhaustive 37 minute video examining the color gamut of RGB working spaces, images and output color spaces. All plotted in 2D and 3D to illustrate color gamut.
Very interesting. Some was "over my head", but it did clear up some questions. Thanks.
I found myself wondering whether your ColorThink 3d images could show how Rendering-Intents like Perceptual and Relative-Colorimetric would adjust colors to fit within gamut. I'm fuzzy on how they work, and speculate that my understanding may very well be flawed. Do you have a video (or videos) related to Rendering-Intent? Black-Point-Compensation? Or video(s) on some other web-site you could recommend?
I did wonder about the down-sampling to get the original images to 300 pixels for ColorThink. It seems like there would be a LOT of averaging, and extreme values would get averaged and lost. I understand that ColorThink would really slow down with less (or no) down-sampling, but perhaps it could run overnight and then your video would just show the results. Or does the 3d spinning on axis slow way down?
High resolution:http: //digitaldog.net/files/ColorGamut.mov
Low Res (YouTube): http://www.youtube.com/watch?v=n0bxSD-Xx-Q
Had trouble viewing the first ... even after removing the space between the colon and slash. YouTube worked ok.
-
I found myself wondering whether your ColorThink 3d images could show how Rendering-Intents like Perceptual and Relative-Colorimetric would adjust colors to fit within gamut.
ColorThink plots gamut using Absolute Colorimetric intent which is what you want for viewing 3D gamuts (not necessarily for conversion of data to make a print, you then toggle between RelCol or Perceptual and pick the one that looks best).
I did wonder about the down-sampling to get the original images to 300 pixels for ColorThink. It seems like there would be a LOT of averaging, and extreme values would get averaged and lost.
Shouldn't be a factor in terms of seeing the gamut. And there's no way you're feeding one let along multiple 5DMII rendered images at full rez in CTP!