When to apply wider color space

[Photoartist]

I have read your excellent tutorial on the benefits of switching from adobe rgb color space to prophoto rgb.  My 'new' workflow takes files input from my cf cards and they are initially processed in acr and opened as a psd smart object in 16 bit, prophoto rgb.  This seems to be working out very well.

My point of inquiry is, when I open an older image that is already in adobe color space it has to be converted to the new prophoto space.  When I open a camera raw image right off the camera and it gets made into a prophoto the software has the full sensor data to process with so I get this much wider gamut or space.  If I open an image that is already a psd in adobe rgb it is the smaller size of the adobe space and so it would seem that this would be a useless conversion since the color space is already shrunk down.  It would be like converting an 8 bit file to 16 bit.  The file labels change but the functional gains are already lost.

To shorten this down, in case I have done a poor job of laying the groundwork.

If I have a pre-existing adobe rgb, psd file can I convert it to prophoto rgb with full benefits other than a metadata change or is it better for me to go back to the cr2 file and do the conversion with the full raw data.  The down side of course would be the loss of all work done to the psd files which makes this an unacceptable choice.  I guess it is possible to process all older work in adobe rgb and all new stuff in prophoto but this has workflow issues as well.

Like many of you, I have thousands of images (in my case now as adobe rgb and psd) and need to know if I should only convert newly shot work to the wider space or if converting my whole image collection to prophoto ( and possibly dng ) is good or at least not harmful.

[Tim Gray]

My 2 cents: Prophoto is recommended because it gives you the largest color space from which to begin - it maximizes flexibility and options.  However, your final display choice will not be prophoto.  For the web (ignoring the color awareness of Firefox for the moment) sRBG is what you get.  If you're going to print an image then the gamut is driven by your paper and ink, and a soft proof of the file will be necessary to see what is being given up in terms of gamut to get the image on paper.  You would need to go back to the original raw file if you wanted to re-print under your new work flow.  If the original raw file was initially converted into a color space other than prophoto, converting the resulting tif or jpg or psd into prophoto isn't going to re-capture the gamut lost on the original conversion.

[marcmccalmont]

A good overview
Marc
http://www.josephholmes.com/propages/AboutRGBSpaces.html

[Brian Gilkes]

A good overview
Marc
http://www.josephholmes.com/propages/AboutRGBSpaces.html

Excellent advice.
You will note from Joe's notes that he is selling Adobe RGB Variant spaces. If you are stuck with old files in Adobe RGB and if you have not clipped channels somewhere along the way, then you can increase colourfulness by assigning a variant space., which is possible up to the stage where clipping is observed on the media soft proof.
Cheers,
Brian
www.pharoseditions.com.au

[Serge Cashman]

Well, short of buying J. Holmes' working spaces and "assigning" them there's no reason to "convert" from a narrower to a wider color space. If you just convert to a wider space you probably would just lose some insignificant amount of image info.

From what I understand Mr Holmes is trying to get more colors in an image than what was actually captured, in which case you'd need assigning wider spaces that are not radically different from eachother.

"Assigning" and "converting" are two completely different things (see the article for more details but for most people "assigning" spaces is irrelevant). You definitely wouldn't  want to "assign" ProPhoto to anything.

[Brian Gilkes]

I don't want to get too far into the assign vs convert discussion, save to say the difference is important and photographers should know the importance of assigning and when to do it.

Assigning is non destructive- the colours retain their numbers, but are remapped into another space , which may be preferable . Converting retains the appearance of the colours by changing their numbers to fit a new space. This is
  imperative when images in different colour spaces need to be combined, however it may not be easy to go back home.
 There are other situations where assigning or converting may be appropriate.
Both are equally useful. A bit like f numbers and shutter speeds. You need both but they are quite different.
It takes a bit of mental effort to get ones head around it. Joe's stuff is spot on , but must be read slowly and carefully.
Good luck
Brian
www.pharoseditions.com.au

[Photoartist]

Thanks for all the feedback.

After thinking about the advice given and reading the referenced sites I ran an experiment that I wan to run by you all.

I opened a problem image in adobe rgb that when soft proofed shows a tremendous amount of gray, out of gamut indicator (shift+cmd+y) in the soft proof option.

The image was closed and opened again this time converting it to prophoto rgb.  The soft proof screen was updated to reflect the new color space and I again (shift+cmd+y) to show what was being clipped and found that the huge amount of clipped/gray colors had gone down to almost none.

Does this indicate that even though I did not get full benefit from expanding the adobe to the prophoto space (from a psd) like I would if I went all the way back to the raw image and converted it, that I still got my colors mapped into the image and will not have the rendering intent chop or adjust as much which should give better colors in my print.

If this assumption is correct then I can continue to process all new images into photoshop as prophoto and can keep all my hard work on the older images stored with layers and smart objects intact by just converting to prophoto without having to go back to cr2 and recreate all the work I have done.

I know this sounds lazy, but I have gone back (a couple of times) and changed my image files to reflect changes in software and hardware.  If we keep on going back and modifying complex, multi-layered image files by the thousand to reflect new bit depths, smart objects, plug-ins and color spaces we will never have time to shoot a new image.  Do you guys have a point where you say that older images that are printing just fine and selling good need to be re-worked yet again?  Or do you hit a point where you say that certain images are going to be left alone and then focus on current shoots to use new hardware and software or processing techniques?  I am trying to balance the growth from new technology and the growth that comes only from shooting new work and not just performing maintainence on the old.

[digitaldog]

Assigning is non destructive- the colours retain their numbers, but are remapped into another space , which may be preferable .

Assigning is indeed non destructive as the numbers don't change as you point out. But the conversion is and here lies the rub. You will get different results in terms of color (duh) and destructive editing based on the conversion which uses the assignment to complete the color space conversion. There's no free lunch here assuming there's another color space conversion.

[snickgrr]

I have read your excellent tutorial on the benefits of switching from adobe rgb color space to prophoto rgb.  My 'new' workflow takes files input from my cf cards and they are initially processed in acr and opened as a psd smart object in 16 bit, prophoto rgb.  This seems to be working out very well.

Which excellent tutorial are you referring to?