[argyllcms] Re: V1.1.0 Perseptual rending Qst
- From: Nikolay Pokhilchenko <nikolay_po@xxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Wed, 20 Jan 2010 17:21:35 +0300
Date: Wed, 20 Jan 2010 12:22:47 +1100
Graeme Gill wrote:
>
> Hi,
> the measurement values for the device seem rather strange.
> For instance:
>
> SAMPLE_ID CMYK_C CMYK_M CMYK_Y CMYK_K XYZ_X XYZ_Y XYZ_Z
> 515 100.00 100.00 100.00 0.0000 -1.0366 -1.5231e-006 -0.23692
>
> A Y value of 0.0 would be highly unusual, a value less than 0.0
> is truly miraculous !
>
> Since you seem to be using scanin, I'm presuming you are using
> a calibrated scanner as a substitute for an instrument. I would
> guess that the test chart is exceeding the range over which the
> scanner is profiled (ie. the scanner reference chart has insufficient
> dynamic range), and that the resulting extrapolation
> is resulting in unrealistic values near black. (I'd
> need the necessary files for the scanin step to check
> exactly what's happening though.) (Perhaps you're using a matrix
> profile for the scanner where a cLUT profile would give better accuracy ?)
>
I was unable to do any realistic printer profile by scanner reading with
LUT-based scanner profile. The better (but not good) results I've got only with
XYZ matrix&gamma scanner profile.
Other related posts:
