MiniDV is a compressed format. It could be far worse - it's still very watchable on TV but the colour blocking is very visible when viewed on a large screen. However the luminance information (Y) is stored at much higher quality than the colour information (Cr and Cb) which provides the DV Rebel with a way of reducing the artefacts. Here's a close-up and closer-up before processing:
A noodle network to separate and quad blur the two colour channels leaving the luminance (Y) channel untouched. As luminance is stored at higher quality, the detail in the image is retained, although if you blur the colours too much you can still get overspill, loss of contrast and very very slow rendering.
The close-up shows less colour blocking and the picture will stand up to slightly more scrutiny. Quad blur is slightly better than flat blur at preserving sharper edges.
A noodle network to separate and quad blur the two colour channels leaving the luminance (Y) channel untouched. As luminance is stored at higher quality, the detail in the image is retained, although if you blur the colours too much you can still get overspill, loss of contrast and very very slow rendering. 
The close-up shows less colour blocking and the picture will stand up to slightly more scrutiny. Quad blur is slightly better than flat blur at preserving sharper edges.
Posts
No comments:
Post a Comment