[argyllcms] Re: Green Color Cast with LUT profiles

  • From: Pascal de Bruijn <pmjdebruijn@xxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Tue, 15 Sep 2009 21:50:03 +0200

2009/9/15 Nikolay Pokhilchenko <nikolay_po@xxxxxxx>:
> Pascal de Bruijn wrote:
>
>> > What instrument You have use?
>>
>> A huey :)
>>
>> > If You have try a colorimeter, it may have a shift itself, because 
>> > colorimeter matrix may be not optimized for your certain display(s).
>>
>> I do not have a funky LED backlit display or something like that, so
>> that should be no issue...
>>
>
> I've noticed, that on some generic LCD displays with classical CCFL backlit 
> get a green cast when calibrated with i1 Display 2 (a colorimeter). It's 
> because most of consumer colorimeter have far not exact spectral sensitivity 
> as in described in CIE standard for standard observer. The correction matrix 
> are applied to colorimeter data for every kind of display type in software 
> (LCD CCFL, LCD LED, Wide GAMUT LCD, Projectors). If Your kind of display 
> wasn't corrected by matrix by manufacturer, the results may be quite far from 
> wanted.

Does this correction happen in hardware? Anyway I though the Huey was
a stupid device... As in just a passive sensor... Since it has no
firmware...

>> And it _does_ work with single shaper matrix profiles.
>
> Yes, It may work because of very restricted fitting capabilities of single 
> shaper matrix. In general, there are rotary dispersion effect in the 
> subpixels of LCD. The light spectrum, which passes through a liquid crysrals, 
> are depend of crystals rotation angle. The effect is more distinctive at a 
> maximum subpixel transparency level. So the averaged by gamut volume shaper 
> data may not to be greenish, but fitted LUT data in certain transparencies 
> levels can.
> Do You try a shaper matrix, not a single shaper?

I've uploaded some sample files here:

  http://files.pcode.nl/test/argyll_greencast.zip

It's the original ti3, all the generated profiles, the profiles are
named as the options used for colprof. I also converted some JPEGs
using jpegicc from lcms...

Please note that only the single shaper matrix profile looks good on
my screen... Even the multiple shaper matrix looks a bit greenish..

I'd appreciate any pointers to what might be the issue...

Regards,
Pascal de Bruijn

Other related posts: