The subject of camera profiles has come up many times here and I have on occasion, mentioned my understanding of the answer Thomas Knoll gave me at an Adobe seminar. Since that point, certain other questions have come up, so I emailed Ian Lyons - from the popular computer-darkroom site (and a 'regular' at the Photoshop forum at Adobe.com). The following is my email and after that his response. I hope it is helpful

Ian,

I'm a regular poster at dpreview 'retouch'. I also archive many of the threads there that have tips, tutorials etc. Color management is one of the common threads that can be hard for people to understand. Your computer-darkroom site is often cited as a good reference - thank you for creating that.

One question that continues to come up is on 'camera profiles' where a camera manufacturer has settings of sRGB, Adobe RGB I, Adobe RGB II, etc. They have different 'names' for the settings. And people want to know if there is 'lost data' if they bring a file from the camera with an 'Adobe RGB' setting into 'sRGB' workspace in Photoshop. (or visa versa - can you 'gain' info by bringing an sRGB file from the camera into Adobe RGB workspace in Photoshop).

There is little info on this on the net. At Adobe, there is this, (under Additional information):

tags vs. profiles
http://www.adobe.com/support/techdocs/327705.html

"The EXIF format is limited to two color space tags--sRGB and Uncalibrated RGB. Most digital cameras attach the sRGB EXIF color space tag to JPEG files, but they don't embed a color profile in the JPEG file. Because the camera doesn't embed a profile in the JPEG file, Photoshop reads the EXIF data and opens the JPEG file into the sRGB color space."

Some insist that there are actual camera profiles because their camera has these 'settings'.

I had the opportunity to take a seminar this summer in Ann Arbor Mich. "Soup to Nuts" featuring Katrin E., Jeff Schewe, Russell Brown and Thomas Knoll. So I asked Tom K about this question. This is what I understood him to say:

That there are 'no camera profiles'. That Photoshop will read only the exif tags (on files coming from the camera) - the binary choice of "sRGB" or "Uncalibrated" and will (if the boxes are checked in Color settings) prompt for mismatches based on those tags along with the workspace settings.

That some camera manufacturers, when faced with tagging the camera's files with either "sRGB" or "Uncalibrated" will in most cases, chose "sRGB" because they would prefer that their files not be tagged "Uncalibrated" - this was delivered with a wink and a nod - actually a smile - indicating that it was merely 'cosmetic'. AND that even an image file that was shot with one of the camera's "Adobe RGB" settings, may still be "tagged as" (not profiled as) "sRGB". And finally, that the color space of any camera image coming into Photoshop is determined ONLY by the working space in Photoshop once editing begins, and that there is no loss of data or gain of data, but only a translation of numbers into whatever workspace gamut that the user has set in Photoshop. That there may be some rounding errors but that is all and is not significant.

He also pointed out that this would Not be the case IF a file was brought into Photoshop where the color space was set to sRGB, for example, saved as an sRGB file, and then reopened into Photoshop with a working space of Adobe RGB. Iow, under Those conditions, there would be No gain of data (gamut) - that the earlier profile cannot be expanded from sRGB to Adobe RGB. Conversely, there Would be a loss of data changing from an Adobe RGB profiled image being converted into sRGB. But this is only the case where the file was brought into Photoshop first, and then saved with a profile.

Again, this is how I remember the conversation going. I reiterated many of the points by asking in this manner "I want to be perfectly clear" or "Ok, I understand that to mean...." etc. So, is this the way that You understand this data? Or do you think there may have been something that I didn't fully understand?

The reason I ask is that the subject has come up again and people simply believe that the settings in their cameras establish a profile. One person recently questioned the Adobe doc, saying that says "MOST cameras" do this, not All, therefore there could be a 'camera profile'. I'm wondering that perhaps even though the exif data can only have a binary choice in tags, doesn't necessarily mean that there might not be an actual profile that exists and that the exif data simply can't show that. However, Tom's description of the situation certainly doesn't point to that - in fact, he stated that 'beginning to edit in a Photoshop workspace setting is the only thing that determines color space' (with files coming from the camera). I don't need to be right here, I just want to be able to pass on correct information for people that ask.

Thanks for any light you could shed.

Regards,
Kent Christiansen
--
Kent
http://www.pbase.com/kentc
For prior discussions on most questions:
http://porg.4t.com/KentC.html
or d/l 'archives' at:
http://www.atncentral.com/Sites.htm