Is the magenta where I am losing integrity?
The profile I sent was the most recent one. What are you using to make your profiles?
Refer to the earlier 2D gamut plot.
The test image (from you) was converted with RelCol (Relative Colorimetric, CS5) and its xyz co-ordinates were plotted in the xyY space.
a) Points in RED represent the original data.
b) Points in GREEN (mostly convered up by RED) represent the data after converting with jc1 profile (see below, Profile4)
c) Points in MAGENTA were data after converting with your profile (see below, Profile1). They were shifted slightly towards the red primary, compared with the original data points. IMHO, the shift towards the Red primary is small and might be insignificant. I compared it with profiles built by other profilers, generally they are very similar (Gamut boundary, neutrality and so on).
The profile that I created earlier was derived from the AtoB tag of the your profile, and that was before the measurement data was received. The measured data file was non-standard and I edited and recreated it for further use. The RGB reference was derived from the supplied target tiff image.More Tests
Profile1: custom profile by nick
Profile2: profile generated with PM5, with measured data supplied by nick
Profile3: profile generated with another profiler, with measured data supplied by nick
Profile4. profile generated with the profiler used to creat Profile3, with measured data derived from AtoB of Profiler1.
Test patches: 4096, RGB with 17-pitch, similar to thisComparison of the 4 profiles
Avg 90% 95% Max error
Profile1 0 0 0 0 (reference)
Profile2 1.02 1.67 1.93 3.98
Profile3 0.68 1.04 1.14 2.07
Profile4 0.08 0.18 0.23 0.86
1. No flaw was detected on Profile1 with the test sample, though its (absolute) gamut with Relcol exhibits abnormal behavior.
Note: gamut observed with Perceptual intent is normal for profile1.
2. Only profile1 exhibits the boundary effect with the test image.
Profiles are available upon request.