DV Artifact Removal
by Dale Paquette on Dec 28, 2009 at 1:01:30 pm
As we all know, HDV is hardly the best format to work with due, in part, to color sub-sampling. That said, many recommend converting to YUV, blurring the U and V, and converting back to RGB to help compensate a little for the sub-sampling. I even have a preset for this which, I think, came with AE CS3. Anyway, I have a question about the blurring process. As I understand it, HDV color sampling goes by the "name" 4:2:0 (though many say 4:1:1 which I think is wrong). There is a difference. In 4:1:1, the R and B channels are sampled in groups of 4 pixels in a row whereas in 4:2:0 the R and B are sampled in groups of 4 arranged in a box. It would seem to me that during the blurring process, 4:1:1 would be blurred horizontally only whereas with 4:2:0 the blurring should be both horizontal and vertical as well. The question is - which to do or does it make any difference at all in terms of trying to squeak just a little more out of HDV?