[argyllcms] Re: Profiling a Minilab

On Sat, Jul 3, 2010 at 9:39 PM, Gerhard Fuernkranz <nospam456@xxxxxx> wrote:
> Am 03.07.2010 19:19, schrieb Pascal de Bruijn:
>> Well, the Lab in this case is foto.com, they can print in three 
>> configurations:
>>
>> FOTO.com Classic (Kodak Edge Glossy)
>> FOTO.com Professional (Kodak Royal Matte)
>> FOTO.com Professional (Kodak Royal Glossy)
>
> I guess that big photo services like foto.com and others do not only
> have a single, but many machines. Can you ensure that subsequent print
> jobs will be printed on the same machine that was used for printing the
> target?

Nope... Though one would think that they have all of their machines
calibrated to a very similar response... Which should mean an
undetailed profile should be usable on their farm.

Obviously this doesn't get you ultimate accuracy, but it should be
able to improve things...

>> They do offer an "uncorrected mode"...
>
> ... which is basically supposed to omit image-dependent enhancements,
> but likely the print pipeline is still color managed, using perceptual
> intent, and expecting sRGB images as input (some photo services possibly
> also honor profiles embedded in the images, instead of generally
> assuming sRGB). So profiling this workflow (for instance because you
> don't like their perceptual transformation) would eventually result in
> double profiling. I assume that some minilabs (i.e. the machines)
> basically would support some kind of raw mode as well, which may be
> better siuted for profiling, but as Edmund said, likely you need a
> personal relationship with the operator in order to get it enabled.

I don't know about this... It's a good point though...

I would have expected them to calibrate all of their equipment to a
similar responsive, but not a fully color managed pipeline...

Anyway, the results can be found here

  http://files.pcode.nl/icc/fotocom/

All fotocom* files are licensed CreativeCommons Attribution-ShareAlike
(for as far it's possible to license this data anyway).

The resulting profiles "look" pretty legit. I still have to submit
some test images though... I want to submit a test set of 5 or 10 test
images, all processed in these ways:

1. Plain sRGB
2. Some BGR profile (tagged) (to unambiguously check if tagged
profiles are read and honored)
3. My FOTO.com profile applied (untagged)
4. My FOTO.com profile applied (tagged)

And check the difference between the results.

Regards,
Pascal de Bruijn

Other related posts: