[argyllcms] Re: Argyll + Firefox

  • From: Michel Joly de Lotbinière <michel@xxxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Wed, 9 Jun 2010 14:53:02 -0400

I'm one of the people still using Firefox 3.0.20 (last version of the
now obsolescent 3.0.x series), due to not yet taking the time to see
if all the extensions I use are still compatible with more recent
versions. ArgyllCMS XYZLUT+MATRIX display profiles work with that old
version and use the XYZLUT part of the profile, tested with the GBR
swapped matrix version of such a profile. Embedded version 4 ICC
profiles also seem to work, at least the examples on color.org

I just downloaded the portable versions of Firefox 3.5.9 and 3.6.3,
and tried them out with the same swapped matrix version of the
XYZLUT+MATRIX display profile: they both display images correctly
using the XYZLUT part of the profile. (I used
as a convenient source of color patches photos with various embedded
profiles; I also used the special test ICC profile available on that
page to convert a photo's colours: the result displayed correctly once
the test profile was embedded.)

Side-by-side comparisons of photos in 3.0.20, 3.5.9 and 3.6.3 browser
windows showed no significant differences with
gfx.color_management.mode set to 1 in the more recent browser versions
(i.e. all graphics are colour managed and untagged images are managed
as if in sRGB, which is the default with 3.0.20). But I didn't spend
too long doing these tests--caveat lector.

These recent versions still cannot deal with version 4 ICC profiles,
at least those embedded in test images on color.org.

So there seems to be progress on the new Firefox CMM--it began as a
rapid cut-and-paste affair out of LCMS (according to the relevant
Bugzilla bug report, whose number I've forgotten), but it has evolved.

On Wed, Jun 9, 2010 at 9:30 AM, Graeme Gill <graeme@xxxxxxxxxxxxx> wrote:
> Magnus Berg wrote:
>> Does or doesn't Firefox 3.5.9 work with Argyll profiles?
> It's hard to say. It probably works with matrix based display profiles.
> [Summary: Firefox used to use Lcms and hence it worked with most profiles.
> In response to an imagined security issue though, some firefox developers
> went crazy and created their own CMM to replace it. Their CMM is
> rather primitive. Feel free to encourage them to revert to sanity,
> or at least bring their CMM up to some level of usability.]
>> I had had my Firefox unmanaged quite a long time now and it gets more and
>> more
>> frustrating. I had a XYZ LUT + matrix profile that work with Gimp and
>> Geany but not
>> with Firefox. It's really no fun then I looking at pictures on
>> photographic sites with
>> my wide gammut display. I have read the thread "XYZ LUT profile unable to
>> be installed
>> on Windows XP or used with Firefox". (
>> //www.freelists.org/post/argyllcms/XYZ-LUT-profile-unable-to-be-installed-on-Windows-XP-or-used-with-Firefox-35
>>  ) I don't understand how to make such a shaper + matrix profile that is
>> said to work
>> with Firefox. " "Just try -g or -s keys for colprof. The profile will be
>> shaper&matrix
>>  in the case and FireFox will handle such profiles well."
>> The example in the thread is: "colprof -v -qh -ax -A Sony -M GDM-F520 -C
>> "(c) 2009" -D
>> "12-20-2009" -S "sRGB.icm" "12-20-2009"
> The example creates a cLut based profile, which was the main topic of the
> thread. For a matrix based profile you need to use -as or -ag or some other
> matrix option (-aG, -aS).
> So at its most basic:
>        colprof -v -as -D "Description" basename
> which reads in the basename.ti3 and spits out basename.icc
> Graeme Gill.

Other related posts: