I only remove those that are "way out of whack". Since I use the 51-patch random test file, it becomes very clear when a point isn't even close to the model curve in Excel. I use the strip mode in ColorMunki to read the patches, and sometimes I get an anomoly. After reading some of the other posts in this thread, I have to admit that I'm sometimes not real careful about making sure the sheet is completely flat, that the Munki exactly follows the centerline of the patches, or doesn't tilt a little during the scan. I've also found that the speed at which the Munki is moved over the patches can have an effect on the final output; too slow and it sometimes misses a patch completely. I'm getting better at it, and will pay more attention to these things.
Message
[Digital BW] Re: Invalid linearize curve -- not constantly increasing
2012-06-01 by wolverinemsu