As far as I can see, using 'auto profile' based on the actual image should produce a more accurate / acceptable result than using set noise profiles derived from a calibration target.
Each time I run this 'auto profile' on different images of the same ISO, it gives completely different results in the amount of noise it sees in the rgb channels. Is this because it is using such a small area of the image to calculate the profile, or it really is seeing differing values.
ie, am I better off with the device noise profiles made from a calibration sheet.
If not, is there a way to set 'auto profiling' to run in batch mode ?
Auto Profile V4.00
Building a profile from the very image you are processing is of course better than using ready-made profile - either built with the Calibration Target or with regular images. Using the input image to build profile is the best way to go IF the input image contains good flat featureless areas that Auto Profile tries to find. You have to check what it is finding, it may be fooled.
And yes, you can do this in Batch, there is an option for Auto Profiling in Batch window.
Vlad
And yes, you can do this in Batch, there is an option for Auto Profiling in Batch window.
Vlad