[FFmpeg-cvslog] VC1: don't use vc1_put_block() in vc1_decode_i_blocks_adv().
Ronald S. Bultje
git
Sun Feb 20 19:28:25 CET 2011
ffmpeg | branch: master | Ronald S. Bultje <rsbultje at gmail.com> | Sat Feb 19 19:07:29 2011 -0500| [a8858ee11cf4f0ae22e0a9df57bec4ec5dd02f80] | committer: Michael Niedermayer
VC1: don't use vc1_put_block() in vc1_decode_i_blocks_adv().
Advanced profile never uses "range reduction", so vc1_put_block() quite
literally just calls put_pixels_clamped() from vc1_decode_i_blocks_adv().
By inlining the function, we can prevent calling IDCT8x8 if
CODEC_FLAG_GRAY is set, and we don't have to scale the coeffs in the
[0,256] range, but can instead use put_signed_pixels_clamped().
(cherry picked from commit 70aa916e4630bcec14439a2d703074b6d4c890a8)
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=a8858ee11cf4f0ae22e0a9df57bec4ec5dd02f80
---
libavcodec/vc1dec.c | 14 +++++++++++---
1 files changed, 11 insertions(+), 3 deletions(-)
diff --git a/libavcodec/vc1dec.c b/libavcodec/vc1dec.c
index 7eb9576..aff8514 100644
--- a/libavcodec/vc1dec.c
+++ b/libavcodec/vc1dec.c
@@ -2704,7 +2704,7 @@ static void vc1_decode_i_blocks(VC1Context *v)
*/
static void vc1_decode_i_blocks_adv(VC1Context *v)
{
- int k, j;
+ int k;
MpegEncContext *s = &v->s;
int cbp, val;
uint8_t *coded_val;
@@ -2747,7 +2747,14 @@ static void vc1_decode_i_blocks_adv(VC1Context *v)
s->mb_x = 0;
ff_init_block_index(s);
for(;s->mb_x < s->mb_width; s->mb_x++) {
+ uint8_t *dst[6];
ff_update_block_index(s);
+ dst[0] = s->dest[0];
+ dst[1] = dst[0] + 8;
+ dst[2] = s->dest[0] + s->linesize * 8;
+ dst[3] = dst[2] + 8;
+ dst[4] = s->dest[1];
+ dst[5] = s->dest[2];
s->dsp.clear_blocks(s->block[0]);
mb_pos = s->mb_x + s->mb_y * s->mb_stride;
s->current_picture.mb_type[mb_pos] = MB_TYPE_INTRA;
@@ -2791,11 +2798,12 @@ static void vc1_decode_i_blocks_adv(VC1Context *v)
vc1_decode_i_block_adv(v, s->block[k], k, val, (k<4)? v->codingset : v->codingset2, mquant);
+ if (k > 3 && (s->flags & CODEC_FLAG_GRAY)) continue;
v->vc1dsp.vc1_inv_trans_8x8(s->block[k]);
- for(j = 0; j < 64; j++) s->block[k][j] += 128;
+ s->dsp.put_signed_pixels_clamped(s->block[k], dst[k],
+ k & 4 ? s->uvlinesize : s->linesize);
}
- vc1_put_block(v, s->block);
if(overlap) {
if(s->mb_x) {
v->vc1dsp.vc1_h_overlap(s->dest[0], s->linesize);
More information about the ffmpeg-cvslog
mailing list