[argyllcms] Re: Firefox qcms vs. lcms 2.0 vs. Argyll imdi

  • From: Florian Höch <lists+argyllcms@xxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Sat, 13 Nov 2010 16:44:52 +0100

Am 12.11.2010 01:20, schrieb Graeme Gill:
At a hint from Marti that he has added a special case for matrix
only profiles, I've repeated the benchmark using lcms V2.0.

Matrix profiles:

lcms 73
argyll 65
qcms 32

cLUT profiles:

lcms 357
argyll 69
qcms n/a

So for the special case that only qcms supports, lcms V2 is just as fast
as Argyll,
and a factor of 2 slower than qcms.

It's nice to see that the Argyll code is doing so well.
And I have now dubbed qcms 'quirkcms', for obvious reasons ;)

Whether this is significant in
practice for
use in a browser probably depends more on whether color management is
being applied efficiently, or whether it is being applied badly
(Consider: even a large display only has around 4 Million pixels, so
repainting the whole screen should take about 15msec's worth of color

Different source profiles used in a webpage (for images applied via stylesheets or img tags) should also play a role, as well as the page complexity (in terms of overlaying images, transparency etc). Somewhat related, I wonder which method is used by the 'color-aware' webrowsers: Transforming the whole webpage (even non-visible parts outside the viewport), or just updating the viewport dynamically (and only transforming the visible parts on the fly) as the page is scrolled.

Florian Höch

Other related posts: