Nikolay Pokhilchenko wrote:
When the glare value is computed correctly, the profile dE will be minimum.
I was thinking along those lines. Create a glare model, and fit it by searching for the glare that gives minimum sample error for a matrix & gamma model.
Yes, It's reasonable. But we shouldn't forgot the sensor bias. The bias is the constant part of error at the darks.
Hmm. I would have expected that to be automatically subtracted in the sensor processing. Graeme Gill.