I keep doing some Wavelet NR (thanks Jack !!) and DBE to each of the RGB files BEFORE combining them...
Are folks ABSOLUTELY sure this is not neccessary/a waste of time..?
It really is NOT necessary to mess around with the individual channels prior to the initial combine. I 'used to' to this, but realised that the new processes 'work better' on the full-colour data.
DBE is a classic example - there is just NO NEED to define three separate sets of sample points for each of the channels. Just define ONE set of sample points, click the button, and watch the magic (you probably won't even have time for a 'sip' of coffee, far less a 'cup')
I can't fully comment about WaveletNR - as I don't use it. I prefer to work in the realm of ACDNR - and that works in TWO domains, Chrominance and Luminance. Which makes things simple again. If you already have your three-channel image, then ACDNR can work on the 'colour-side' of the data as one step, and the 'luminance-side' of the data as a second step - and the most that you will have to do is click a couple of selction boxes to change between the two.
And, I have found that it IS better to handle the ACDNR process as two discrete steps - simply because you are dealing with two totally different, and unique, 'noise sources' ('chrominance' and 'luminance') - and they really do benefit from different approaches in my experience.
But - admit it - it IS getting 'easier', isn't it
Cheers,
