diff options
| author | Ronald S. Bultje <rsbultje@gmail.com> | 2013-01-15 08:38:54 -0800 |
|---|---|---|
| committer | Michael Niedermayer <michaelni@gmx.at> | 2013-01-20 02:43:06 +0100 |
| commit | d369cc83aa763a780488c9b511507c8624d51c69 (patch) | |
| tree | f9bd526ef156cc536ffa25cb57f55ca895b9f7dd | |
| parent | ac6b2fecf4731ddaac546a38294918be39ff190a (diff) | |
| download | ffmpeg-d369cc83aa763a780488c9b511507c8624d51c69.tar.gz | |
h264: add 3 pixels below for subpixel filter wait position.
If the motion vector is at a subpixel position, we need 3 pixels below
the motion vector's wholepel position available, not 2, since the MC
filter is a sixtap filter for the hpel position, and then a bilin filter
for the qpel position.
This patch fixes highly irreproducible (0.1%) fate failures in frame 2
and 4 of h264-conformance-cama2_vtc_b (e.g. first P-frame, first field,
last line of MB x=40,y=2 and second field and last lines of MBs x=39-40,
y=3). These used pre-loopfilter instead of post-loopfilter data because
the await_progress() waited for one line too little in that field, and
the motion vector of these particular MBs happened to align exactly to a
position where that demonstrates the bug.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
(cherry picked from commit 2c85d7c015489e169759e52c21499579dae5c2b1)
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
| -rw-r--r-- | libavcodec/h264.c | 7 |
1 files changed, 4 insertions, 3 deletions
diff --git a/libavcodec/h264.c b/libavcodec/h264.c index f2006ec633..c8c3974ee7 100644 --- a/libavcodec/h264.c +++ b/libavcodec/h264.c @@ -310,10 +310,11 @@ static inline int get_lowest_part_list_y(H264Context *h, Picture *pic, int n, int height, int y_offset, int list) { int raw_my = h->mv_cache[list][scan8[n]][1]; - int filter_height = (raw_my & 3) ? 2 : 0; + int filter_height_up = (raw_my & 3) ? 2 : 0; + int filter_height_down = (raw_my & 3) ? 3 : 0; int full_my = (raw_my >> 2) + y_offset; - int top = full_my - filter_height; - int bottom = full_my + filter_height + height; + int top = full_my - filter_height_up; + int bottom = full_my + filter_height_down + height; return FFMAX(abs(top), bottom); } |