# HG changeset patch
# User michael
# Date 1229568798 0
# Node ID e1b2422249704a5411c44167ffc625c880a54b9f
# Parent  de2509cf3c44c577b9c4262960ffdc1597a7f3bc
Use the new idct functions (except chroma as it was slower in benchmarks)
cathedral +0.5% speed
aladin +0.6% speed [note aladin has been cat-ed 10 times to reduce the influence
of init time]
Speedup also verified via START/STOP_TIMER (difference was very significant
for the changed parts)

diff -r de2509cf3c44 -r e1b242224970 h264.c
--- a/h264.c	Thu Dec 18 02:36:48 2008 +0000
+++ b/h264.c	Thu Dec 18 02:53:18 2008 +0000
@@ -2515,27 +2515,34 @@
         if(!IS_INTRA4x4(mb_type)){
             if(is_h264){
                 if(IS_INTRA16x16(mb_type)){
-                    if(transform_bypass && h->sps.profile_idc==244 && (h->intra16x16_pred_mode==VERT_PRED8x8 || h->intra16x16_pred_mode==HOR_PRED8x8)){
+                    if(transform_bypass){
+                        if(h->sps.profile_idc==244 && (h->intra16x16_pred_mode==VERT_PRED8x8 || h->intra16x16_pred_mode==HOR_PRED8x8)){
                         h->hpc.pred16x16_add[h->intra16x16_pred_mode](dest_y, block_offset, h->mb, linesize);
                     }else{
                         for(i=0; i<16; i++){
-                            if(h->non_zero_count_cache[ scan8[i] ])
+                            if(h->non_zero_count_cache[ scan8[i] ] || h->mb[i*16])
                                 idct_add   (dest_y + block_offset[i], h->mb + i*16, linesize);
-                            else if(h->mb[i*16])
-                                idct_dc_add(dest_y + block_offset[i], h->mb + i*16, linesize);
+                        }
                         }
+                    }else{
+                         s->dsp.h264_idct_add16intra(dest_y, block_offset, h->mb, linesize, h->non_zero_count_cache);
                     }
                 }else if(h->cbp&15){
+                    if(transform_bypass){
                     const int di = IS_8x8DCT(mb_type) ? 4 : 1;
                     for(i=0; i<16; i+=di){
                         int nnz = h->non_zero_count_cache[ scan8[i] ];
                         if(nnz){
-                            if(nnz==1 && h->mb[i*16])
-                                idct_dc_add(dest_y + block_offset[i], h->mb + i*16, linesize);
-                            else
                                 idct_add(dest_y + block_offset[i], h->mb + i*16, linesize);
                         }
                     }
+                    }else{
+                        if(IS_8x8DCT(mb_type)){
+                            s->dsp.h264_idct8_add4(dest_y, block_offset, h->mb, linesize, h->non_zero_count_cache);
+                        }else{
+                            s->dsp.h264_idct_add16(dest_y, block_offset, h->mb, linesize, h->non_zero_count_cache);
+                        }
+                    }
                 }
             }else{
                 for(i=0; i<16; i++){