I've checked several device data sets of RGB printers. The mapping techniqe of Argyll CMS is very good from Jab space point of view. But my request still active. Take a look to .ti3 attached. My test was: colprof -nc -S"sRGB.icc" LSM2484 xicclu -fb -ir -pj -s255 LSM2484.icm 0 0 0 [Jab] -> Lut -> 21.542882 15.857801 0.000150 [RGB] (clip) xicclu -ff -ia -pj -s255 LSM2484.icm 255 255 255 [RGB] -> Lut -> 92.590810 2.033623 -7.751497 [Jab] 0 0 0 [RGB] -> Lut -> 15.022740 -1.045231 -10.012886 [Jab] 21.542882 15.857801 0.000150 [RGB] -> Lut -> 16.663121 1.075238 -6.318114 [Jab] Here we can see, that relative-mapped black point isn't so dark, as R=G=B=0, but the level of lightness is less, than the printer native tone shift compensated by mapping. Let's check the perceptual intent: xicclu -fb -ip -pj -s255 LSM2484.icm 0.000000 0.000000 0.000000 [Jab] -> Lut -> 4.710601 4.732119 0.000313 [RGB] (clip) xicclu -ff -ia -pj -s255 LSM2484.icm 4.710601 4.732119 0.000313 [RGB] -> Lut -> 15.468991 -0.929117 -9.292939 [Jab] 0.000000 0.000000 0.000000 [RGB] -> Lut -> 15.022740 -1.045231 -10.012886 [Jab] As in relative intent case, the "mapper" does it's work good - the shifted black point is slightly closer to white point, than native bp, but at the cost of less black lightness increase. Generally, there are no any claims to "mapper", but the requested feature is needed for some special cases. For exaple text, artworks and other cases, where perceptual or relative color management is needed, but with accordance L=a=b=0 to R=B=G=0.