[FFmpeg-cvslog] H.264: fix bug in lossless 4:4:4 decoding

Jason Garrett-Glaser git at videolan.org
Wed Jul 27 12:49:36 CEST 2011


ffmpeg | branch: release/0.7 | Jason Garrett-Glaser <jason at x264.com> | Mon Jun 20 15:56:35 2011 -0700| [e4071fa04c213ab3ddf4f89953daf4e3788a9d0f] | committer: Reinhard Tartler

H.264: fix bug in lossless 4:4:4 decoding

Coefficient test for i16x16 add_pixels4 assumed luma plane.
(cherry picked from commit 3b79f2e2e928eb346dbea85cc89393dcdf010ca9)

Signed-off-by: Reinhard Tartler <siretart at tauware.de>

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=e4071fa04c213ab3ddf4f89953daf4e3788a9d0f
---

 libavcodec/h264.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/libavcodec/h264.c b/libavcodec/h264.c
index 13a6380..124f9a8 100644
--- a/libavcodec/h264.c
+++ b/libavcodec/h264.c
@@ -1744,7 +1744,7 @@ static av_always_inline void hl_decode_mb_idct_luma(H264Context *h, int mb_type,
                         h->hpc.pred16x16_add[h->intra16x16_pred_mode](dest_y, block_offset, h->mb + (p*256 << pixel_shift), linesize);
                     }else{
                         for(i=0; i<16; i++){
-                            if(h->non_zero_count_cache[ scan8[i+p*16] ] || dctcoef_get(h->mb, pixel_shift, i*16))
+                            if(h->non_zero_count_cache[ scan8[i+p*16] ] || dctcoef_get(h->mb, pixel_shift, i*16+p*256))
                                 s->dsp.add_pixels4(dest_y + block_offset[i], h->mb + (i*16+p*256 << pixel_shift), linesize);
                         }
                     }



More information about the ffmpeg-cvslog mailing list