[argyllcms] Printer Profiling Patch Count and Profiling Generation Quality
- From: Nikolay Pokhilchenko <nikolay_po@xxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Tue, 05 Jan 2010 03:32:16 +0300
Pascal de Bruijn wrote:
> As some might have noticed, I'm the proud new owner of a ColorMunki,
> when generating targets for it, the standard target can do about 90
> patches, and the dense target can do about 210 patches on an A4 sized
> paper.
IMHO, It's rather small patch count per page. I'm successfully printing 690
patches on A5 and 1440 patches on A5 (6x6mm) for my i1Pro with negligible
tolerance decreasing (about 0.02dE RMS) in comparing with i1Pro native 12x10mm
patches.
> The first profile I did was with 90 patches and -ql which seemed to
> work reasonably well for satin photo inkjet papers, and less well for
> true-matt photo inkjet papers.
It's unbelievable! Generally the profiles from 200-300 patches for InkJets
aren't good.
> Now I read my first 210 patch target (which worked excellently), and
> I'm wondering whether it has any merit to use -q m?
>
> And what if I do a 420 patch target, or 630 or 840? At what patch
> counts are certain profile quality levels sensible?
The profile quality parameter is influencing on CLUT resolution. With higher
quality the profile has more points in its tables. So it can describe the
device behavior more precisely by cost of a computing time, profiler memory
consumption and file size.
> After inspecting a generated profile with iccdump, I can't make much of it:
>
> Input Channels = 3 (PCS=LAB?)
For A2B table yes.
> Output Channels = 3 (RGB?)
For A2B table yes, If Your printer is an "RGB". It's just a device channels.
> CLUT resolution = 9
It's the number of points in every dimension of a profile LUT. With higher
quality this number will be higher. Just imagine himself are the 9 optimal
placed points enough for describing Your printer behavior? If yes, than the
quality (-q) is enough. If the printer response is highly non-linear, the
number of points (the profile quality and patch number) must be increased.
I use 690 patches for RGB-printer with -qm quality minimum. 2880 patches with
-qu quality gives the profile with 1.6 dE2000 maximum error while checking A2B
table with different random verification chart and 1.5 dE max for B2A while
invprofcheck.
> That leaves the CLUT resolution and table entries... How do these fit
> together? Besides the interpolation between entries, how does a table
> lookup actually work?
The profile structure and principles are described in ICC specification.
I'm using -v option with colprof and collink every time. It can help to
understand what the profiler do.
Best regards,
Nikolay.
Other related posts:
