|
Authored by: Anonymous on Tuesday, March 26 2013 @ 04:44 PM EDT |
It isn't dithering -- it is deblocking. The idea is each video still frame is
encoded as a matrix of rectangular blocks of pixels. Some blocks are encoded
directly like for a jpeg image, while others are encoded by reference to other
frames of the video. Depending on the codec, there can be several different
possibilities for exactly how a block is calculated. Unfortunately in the
decoded image ugly block boundaries are visible. To mitigate this, some
smoothing or filtering can be done on top of the original block calculation.
This tries to airbrush away the ugly block boundaries. All of this so far is
prior art background to the claimed invention. The patent claims as new a
strategy to use different deblocking filters depending on the calculation mode
of surrounding blocks.
[ Reply to This | Parent | # ]
|
|
|
|
|