Camera profiling target evaluation

[torger]

If the LUT gets aggressive localized stretches in it (easiest to see by plotting), the profile will be extremely sensitive no white balance setting. If the white balance is a little off from the designed, the colors can get dragged away by a correction that actually belongs to a neighboring color. Maybe there's something like that going on. A matrix-only profile is the most robust type of profile you can make, both in terms of measurement errors in input data but also how it behaves under different white balances than it was designed for.

[hurodal]

Hi there,

It's my first post into this forum. My name is Hugo and I work mainly teaching technical photography to advanced users and professionals. Part of my work is about color management, and camera profiling is always interesting to me.

This discussion is very interesting. Although I normally don't use command line tools, I'll definitely look at DCamprof.

I have made dozens of experiments and tests with camera profiling and I have been profiling cameras since 2003, but always with real data. I have been always very picky and I've been refining my procedure so I'm sure my color target captures are really of very high quality in every aspect.
Never tried to do that with spectral data, but looks interesting. However, does spectral data takes into account the characteristic curve of the sensor or does it assume it is absolutely linear?

Is there any chance that you make some user friendly window with all options of dcamprof?

Best regards from Barcelona,