BTW, the second part of the paper with the denoising/degraining algorithms also looks highly promising, I never actually took much notice of it before as when I first got into this I was only interested in the colour matching - but that noise reduction could be quite something.