I've seen the sRGB vs. Adobe RGB debate since I first got into digital photography. Adobe RGB has a larger color gamut than sRGB, but photo labs want sRGB. ProPhoto RGB has an even larger color gamut than Adobe RGB.
Adobe Camera RAW will let me process my RAW files to any of these 3 color spaces, but my camera (Canon) only lets me choose between sRGB and Adobe RGB.
What color space setting do you use in your camera(s) and for your editing in Photoshop, and why?
Part two. In short since I realize this is a huge subject, but a brief overview of how you use this to get a print to match the screen?
In camera I use sRGB since the only thing that will be used for is the jpg proofs that I will never open myself but the clients will, and the review screen on the camera which is not colorspace aware.
In raw conversions I use ProPhoto or adobe RGB, always 16 bit and more often if its an all inhouse production and post work we use ProPhoto, if it may be outsourced prior to final retouching we use Adobe RGB since thats what many graphic houses and retouchers are familiar with. ProPhoto offers the larger space and since we are working on a master file and it makes very little difference which we choose we work on what offers us the most latitude Pro 16bit, before things leave they are often converted always flattened. Most often for post work things will go in Adobe RGB, for graphics we ask a design house what they prefer, for prepress we do our own CMYK conversions and approvals are included.
All of the printers have smaller colorspaces than Adobe or ProPhoto, but they may have a specific color that hits outside the smaller spaces so its best to do all work on the largest file and colorspace and than convert down to smaller spaces. If you do not do much offset prepress work its better to let the prepress house do it. Sharpening as well! They do it for output and its always better. Capture sharpening is not needed and does nothing but make the viewer of the process feel better in between, and in fact may result in degradation of captures detail through the artificial induction of sharpening that served no purpose! If you can do it at the beginning you can still do it at the end, but if you change your mind or went to far its too late unless you want to start all over, what was the point? none. Sharpen last after conversion to colorspace and all resizing is complete and you are going to a specific output if possible.
Clients and anyone outside of prepress or retouching/graphics/design only get sRGB as most often they do not know nor care about the difference and I have seen them watch a disc of proofs not even on a uncalibrated monitor which is normal, not on one that is not colorspace aware, but on a television dvd player at home that reads and plays jpgs!!!!!! God I don't think they can look any worse if I let my cat screw with them before I sent them out. But thats not that unusual I have had many calls from people who were doing or mentioned just that. They are only looking at proofs for a selection so it really does not matter, but I hate the idea of what they are seeing and thinking it looks like. Thats life, I get over it
So I make sure its in a space that most devices, the internet and browsers
are most comparable too and that seems close to sRGB. But hey thats me, sometimes I think too much
Now how do you match a print to a screen?
You need to have both ends calibrated, so you have a monitor profile than a printer profile which generates an Icc. profile for that printer/paper combination. they drift and change over time or with chemical/ink changes but should be close for a while.
Now the adobe or prorgb or even srgb being a nice and relatively big colorspace is a good choice for most people to work and retouch and do all corrections to as a master file. Save that file as that larger space as a master file.
Once your done with the file and ready to go off to a printer of any sort, that printer should have an icc profile associated with it, ideally made from that specific printer and paper not a generic one for all epsons with lustre so to speak, but that generic is likely better than none at all.
You get that icc profile and in PS you can go to view proof setup and choose that colorspace/profile (View >Proof Setup> custom >Select profile here)
You are now seeing what that printer can and should print from the file as it is interpretated into what your monitor can see based on your monitor conversion.
The ideal way (as best it can be considering the difference between reflective and transmissive media) Is when selecting softproof options, (View >Proof Setup> custom >) than select your icc profile, and select simulate paper white. Do not look at the screen when doing that, it will only show you how bad it is to not be backlit, try it without looking and it wont be as dramatic a difference to you. This basically cuts the dynamic range down and tries to simulate what the image woudl look like on actual paper. Try it, see if you like it. I do not usually do it that way, but I am use to this for a long time and know things will not be as punchy when not lit from behind
You adjust what that looks like in the color space of the printer until it is optimized for print. Than you need to save that file as a copy but first you need to actually convert the file to that colorspace as so far it was just showing you a proof of it. You go to convert to profile and convert to profile. Once converted you have two options, you can save with no embedded profile attached, which works for many of the printers who ignore profiles anyway, or you can save with the embedded profile so that anyone who opens it will know what it should look like.