esby
10th February 2003, 15:43
Well i'm trying to see if it is possible to detect Lumina Blocks (LB) created by div3.11 encoding, (sbc or not).
There are two cases:
when we have the original frame.
when we don't have it.
For the first case, i think i have something working.
(using statistics methods on luma difference).
But for the second case, i'm wondering if anybody tried to code something.
Or has some ideas to detect or/and correct the LB(in a non-compressed form of course).
esby
PS: i know that Dup in 'chrominance mode only' could solve some of these block, when the frame corrupted is identical except the block to the next frame. But for the rest of the LBed Frames dup would not work i think.
There are two cases:
when we have the original frame.
when we don't have it.
For the first case, i think i have something working.
(using statistics methods on luma difference).
But for the second case, i'm wondering if anybody tried to code something.
Or has some ideas to detect or/and correct the LB(in a non-compressed form of course).
esby
PS: i know that Dup in 'chrominance mode only' could solve some of these block, when the frame corrupted is identical except the block to the next frame. But for the rest of the LBed Frames dup would not work i think.