New print profiling functionality

[Richard Hughes]

On 23 February 2010 17:44, Pascal de Bruijn <pmjdebruijn at pcode.nl> wrote:
> I just tried the brand spanking new print profiling functionality...
> And of course there's something to bitch about...

Heh, good. :-)

> * It only generated 45 patches, while on A4 I can fit 210 ColorMunki
> patches in double density mode (-h).

Right, this is something I need help with. How many patches do we use
for quick, normal and slow profiling? I'm already using -h for
ColorMunki.

> * Some patches got clipped (because of margins), because the generated
> image isn't centered (or it is getting stretched for no reason, since
> the generated image doesn't include margins).

Ahh, I'm just blitting it to the cairo surface, without preserving
aspect. I guess we need to do that. I'm using
gdk_pixbuf_new_from_file_at_scale() which is probably a bad plan.

> * The generated image is pretty low resolution, the text looks like
> crap, consider (printtarg -T 200 or 300)?

Hmm, it should already be using 300DPI thanks to the -t300 option.

> * GCM should not report "reading the patches" during self-calibration,
> we need a seperate dialog for this!

Sure, I can't get my ColorMunki to self calibrate using chartread (see
the Argyll m/l) so I've not tested past this point.

Richard.