[argyllcms] Re: Perceptual intent

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Wed, 17 Dec 2014 10:12:55 +1100

Hi Martin,

> Perceptual conversions to sheet fed print profiles like FOGRA39 are generally 
> OK but
> there is a tendency to boost the saturation in yellows. This can look 
> pleasing in pale
> skin but nasty in already saturated, darker skin tones (around the neck for 
> example).
> 
> Warm tones in shadows are also boosted so that they are easily noticed and 
> commented on
> when clients are viewing contract proofs. In version 1.6 the warmth in the 
> shadows
> became so pronounced that we were unable to use it at all and still use 1.5 
> for RGB to
> CMYK conversions.

It would be really useful if you could send me an example of this, so that I
can look into it. A .ti3 + profiling parameters + image is what I need.

> Perceptual conversions to high speed press profiles and/or those using 
> uncoated media
> aren’t good at all. For a profile like PSO Uncoated I see results that are 
> along the
> lines of the OP’s observations — so lacking in contrast that we could never 
> send them
> to press.

Same with this. I need some obvious examples if I am to be able to get to the
bottom of things.

> I’m guessing that the problem here is that the numbers are mathematically 
> sound but the
> physics of ink and dots on paper requires another layer of processing… and 
> the sort of
> bending, distortion and clipping that would give some people nightmares :-)

I wouldn't expect so - the details of the printing process are typically 
irrelevant
once the space has been sampled and reduced to a 4D Lut. The gamut limits and
shape of the two gamuts, and how the mapping relates to important visual 
elements
is typically the driving factors in the gamut mapping itself.

> I’m sure that a high quality print made with an Argyll profile looks amazing, 
> but if
> the artist needs a CMYK conversion that will be printed in an uncoated 
> exhibition
> catalogue I would use an alternative method.

Testing to small gamut destinations such as news print is something I've done
in tuning the gamut mapping, so I'm surprised that you report such reservations.
I need specific examples to be able to change things though.

Graeme Gill.

Other related posts: