[HDR-photo] Gamut choice forHDR.

[Royce Howland]

Jeffrey Hartog wrote:
> Actually I stripped the profile accidentally on saving out of Photoshop [...]

That happens. Whenever I hear a) someone is using multiple color spaces 
other than sRGB; and b) color is changing radically; then c) I 
immediately check color spaces, followed by app color management 
settings. :)

> I personally use an Eizo monitor which is 
> pretty close to AdobeRGB.

In that case your view of the ProPhoto RGB version of the image should 
look even better than my view of it. My Dell 24" LCD is pretty much a 
straight sRGB device. I've got it dialed in quite well through 
calibration, but at best it's a moderately good consumer-level display.

Even so, with the two test files you sent, the tonal resolution in the 
saturated red-yellows looks better to me on the ProPhoto RGB version of 
the file than on the sRGB version of it. But the caveat is that the app 
has to be properly interpreting the color space. If you just load the 
ProPhoto RGB JPEG file into Photomatix or some other app that isn't 
fully color managed, it won't look right even if the color space tag is 
in the file.

> I think what I take from this is that I might look at the gamut of my 
> original images prior to HDR ( I usually start in Prophoto) and consider 
> what my output objectives are, print, web etc. [...]

Certainly nothing wrong with that policy at all. It's advocated by 
Joseph Holmes, for example, which is why he created a series of 5 nice, 
congruent, synthetic working color spaces based on increasing gamut 
volumes. I'd be using his DCam color spaces myself except that it's 
already tricky enough dealing with ProPhoto RGB in some areas, and the 
Adobe RAW conversion tools do not support arbitrary color spaces as the 
output target, much to my continued annoyance.

http://www.josephholmes.com/profiles.html

Personally, I've never found use of ProPhoto RGB to cause me any issues, 
as long as I stay within a 16-bit workflow, and recognize that some 
colors may be (or become) pushed out of gamut of my monitor. I soft 
proof stuff pretty religiously and check color at critical points in 
development, usually towards the end once I've determine my output 
formats and so on.

> I think what is interesting 
> in this case is that the Prophoto result is actually less saturated 
> which is counterintuitive, but makes sense when one looks at the 
> ColorThink graphics, and one sees how the out of gamut component in the 
> HDR Prophoto image has been magnified from the original TIFF.

Err, maybe we're saying the same thing in different ways but it's not 
clear to me if so. :) The ProPhoto RGB version of the file is not 
actually less saturated. It only appears that way when its color space 
is stripped, and/or it is rendered by an application that can't 
understand the color space tag, and so defaults to sRGB instead. This 
doesn't have to do at all with the modifications caused by the HDR tone 
mapping.

The RGB numbers in the image file do not represent absolute color, 
right? They position the colors of each pixel relative to the RGB color 
primaries of the associated color space. The primaries of ProPhoto RGB 
are much more saturated than those of sRGB, which is why the gamut 
volume of ProPhoto RGB is so much larger. So if you take RGB numbers 
from a file that is supposed to be interpreted as ProPhoto RGB, and 
instead interpret it as sRGB -- wham, instant drop in saturation. It 
doesn't mean the image is less saturated, it just means the 
interpretation of the numbers is incorrect.

If you already got that, sorry... just being absolutely clear. :)

Royce Howland
Calgary, Alberta
http://www.vividaspect.com/