FreeLists / argyllcms / [argyllcms] cctiff -d (was: relativ volume of gamut -- vrml comparison?)

[Klaus Karcher <lists@xxxxxxxxxxxxxxxxx>]

Preben Soeberg wrote:
> There seems to be a problem with the -p flag. See
> http://img242.imageshack.us/img242/412/timagetxsp6071fadobergbxc3.jpg and
> http://img149.imageshack.us/img149/2008/timagexp6071fadobergbmz2.jpg
/very/ strange.

I can't reproduce your bug.

I only noticed that your camera profile is one of these "one way 
profiles" (like many other camera profiles -- bothersome).
iccdump reveals that it has tag table entries for the device2PCS 
direction: A2B0, A2B1 and A2B2. All of them point to the same offset 
(i.e it suppots only one RI). There are no B2Ax-Tags -- you can't use it 
for PCS2device transformations.
this transformation works for me:

timage -t -x timaget.tif
cctiff -p -ir 6071_F.icc -ir sRGB.icc timaget.tif canon2srgb_r_p.tif

-> everything is ok so far, the converted TIFF looks ok.

but when you try to use the profile in the opposite direction, cctiff 
returns a (not exactly constructive ;-) error message:
cctiff -p -ir sRGB.icc -ir 6071_F.icc timaget.tif srgb2canon_r_p.tif
cctiff: Error - 2, icc_read_tag: Tag 'Media Black Point' not found from 
'6071_F.icc'
Also Photoshop nags when you try to paste data to an image with that 
profile (see screenshot).
I don't understand why neither cctiff nor Photoshop uses the matrix part 
of the profile for PCS2device transforms - it should be no problem to 
use it. Or have I overlooked something?
Klaus