Can someone smarter than me tell me exactly why X-Rite's camera profiles are so bad?
Can someone smarter than me tell me exactly why X-Rite's camera profiles are so bad?
- Subject: Can someone smarter than me tell me exactly why X-Rite's camera profiles are so bad?
- From: Nathan Duran <email@hidden>
- Date: Tue, 14 Jun 2005 23:04:38 -0700
I realize that camera profiling is largely pointless exercise, and that it
rarely if ever yields attractive images for a wide variety of reasons, but I
am of the opinion that it can at least assist in color correcting images
shot under atypical illuminants, so when X-Rite began shipping their new
Pulse suite of tools that included a camera profiling module, I was anxious
to try it out. I have since found Pulse's camera profiles to be astoundingly
inferior to those produced by other packages however, and since I have
received no response from their tech support department when I sent them a
very similar email to this one, I'm not entirely sure what the actual
problem is.
I shot a GMB ColorChecker DC target under fluorescent lighting (which was
admittedly not perfectly even, but that's what the large array of neutral
patches around the edge of the target are there for, right?), and built a
profile from the resultant image with the demo version of Fuji ColorKit's
camera profiling utility, and while assigning this profile to the image in
Photoshop CS2 made the blacks a little muddy, the colors are pretty much
spot on throughout, and a minor levels tweak is all that is necessary to fix
the muddiness. The unadjusted image is here:
http://img177.echo.cx/img177/8614/fuji1gq.jpg
I then made a profile from the exact same unmodified TIFF file and the exact
same sample data file that I used for the Fuji utility with Pulse, and while
assigning it yields a darker and more attractively saturated look overall,
it also degrades the image severely:
http://img137.echo.cx/img137/5425/pulse2ss.jpg
There is an obvious green cast to all the blacks, there is some kind of
posterized scum all over the place (especially obvious near the title at the
top), patch B10 is ridiculously oversaturated, and J8 is noticeably darker
than L8, which is the exact opposite of how it should be. In practice with
real images shot under the same lighting, the same horrible posterization is
seen in nearly every shadow, and highlights frequently seem to invert
themselves to become shadows when the Pulse profile is assigned, almost as
though a Sabattier filter had been applied.
I've reshot and rebulit this thing more times than I can count, and the
results never improve. The differences I have noticed in the profiles
themselves are as follows:
1) Fuji specifies the Relative Colorimetric intent as the default, whereas
Pulse opts for Perceptual.
2) Fuji specifies a black point tag, Pulse does not.
3) Fuji creates larger lookup tables (199732 bytes vs 96948), and when
viewed in GamutWorks, they seem to produce a much more jagged and realistic
looking volume than Pulse's optimistically smooth one.
4) Fuji ignores the glossy patches entirely. I don't know if Pulse does or
not. Since GMB's own software provides this as an option, it would seem that
it is likely necessary to omit those patches under certain circumstances.
Fuji certainly appears to be the better product here, but I'm not about to
shell out another $600 for something I will use so rarely; I'd greatly
prefer it if the software I already bought would simply do what it is
supposed to since I have been reasonably pleased with all the other sorts of
non-camera related profiles I have made with it. Can anybody tell exactly
what is going wrong from looking at these images and whether or not it is
some glaring error on my part?
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden
- Follow-Ups:
- ACR
- From: Roger Breton <email@hidden>