[argyllcms] Re: perceptual black too light

Gerhard Fuernkranz wrote:

xicc/xicclu -ip -fb -pl CHP410-1200.icm
0.000000 0.000000 0.000000 [Lab] -> Lut -> 0.999641 0.101813 0.132937 1.000000 [CMYK]

Note that Lab 0 0 0 returns the closest point from the perceptual table,
not the black point.

Sorry for my ignorance - aren't you using the the line segment between bkpt..wtpt (in CIECAM space, thus it is a slightly bent curve in CIELAB) as perceptual black axis? And doesn't perceptual intent map the PCS gray axis Lab=[0,0,0]..[100,0,0] to the perceptual gray axis? If so, why doesn't PCS [0,0,0] map to the bkpt? 

Sorry, I was overlooking the point that sRGB' black point is 0,0,0,
and is therefore within the range of the gamut mapping, and so
is not necessarily being clipped. (For source profiles with more normal
black points, this is more likely to be the case.)

The explanation as to why the output doesn't necessarily match the
destination black point, is that with the new gamut mapping, I'm not
assuming black point adaptation, so the src black point doesn't
map to the destination black point. The source black point will
be adapted to any src->dst shift in the white points, and will
be adapted in J to have the same J as the destination black point,
but target grey axis direction is not affected by this. As a result,
the desired output value corresponding to the input black point
may get clipped to the output gamut, raising the J value,
and moving the a & b towards the darkest point in the destination gamut.

(As far as I know, a PCS black point >= 0 is only defined in V4 (-> reference medium), while V2 profiles rather seem generally to assume Lab=[0,0,0] as PCS BP for perceptual intent). 

Yes, there was some tweaking after V2.3. I haven't really researched the
implications.

Graeme Gill.

Other related posts: