Merge remote-tracking branch 'qatar/master'

[ffmpeg] / libavcodec / h264pred_template.c

diff --git a/libavcodec/h264pred_template.c b/libavcodec/h264pred_template.c
index 80218953781221af638f03328201dd3009a53886..3a1b1cf94ecadd8a906a10cc63d065fc5f7c3ee4 100644 (file)
--- a/libavcodec/h264pred_template.c
+++ b/libavcodec/h264pred_template.c
@@ -388,9 +388,9 @@ static void FUNCC(pred16x16_##n##_dc)(uint8_t *_src, int stride){\
     PREDICT_16x16_DC(PIXEL_SPLAT_X4(v));\
 }
 
-PRED16x16_X(127, (1<<(BIT_DEPTH-1))-1);
-PRED16x16_X(128, (1<<(BIT_DEPTH-1))+0);
-PRED16x16_X(129, (1<<(BIT_DEPTH-1))+1);
+PRED16x16_X(127, (1<<(BIT_DEPTH-1))-1)
+PRED16x16_X(128, (1<<(BIT_DEPTH-1))+0)
+PRED16x16_X(129, (1<<(BIT_DEPTH-1))+1)
 
 static inline void FUNCC(pred16x16_plane_compat)(uint8_t *p_src, int p_stride, const int svq3, const int rv40){
   int i, j, k;
@@ -502,9 +502,9 @@ static void FUNCC(pred8x8_##n##_dc)(uint8_t *_src, int stride){\
     }\
 }
 
-PRED8x8_X(127, (1<<(BIT_DEPTH-1))-1);
-PRED8x8_X(128, (1<<(BIT_DEPTH-1))+0);
-PRED8x8_X(129, (1<<(BIT_DEPTH-1))+1);
+PRED8x8_X(127, (1<<(BIT_DEPTH-1))-1)
+PRED8x8_X(128, (1<<(BIT_DEPTH-1))+0)
+PRED8x8_X(129, (1<<(BIT_DEPTH-1))+1)
 
 static void FUNCC(pred8x16_128_dc)(uint8_t *_src, int stride){
     FUNCC(pred8x8_128_dc)(_src, stride);
@@ -663,23 +663,45 @@ static void FUNC(pred8x8_mad_cow_dc_l0t)(uint8_t *src, int stride){
     FUNCC(pred4x4_dc)(src, NULL, stride);
 }
 
+static void FUNC(pred8x16_mad_cow_dc_l0t)(uint8_t *src, int stride){
+    FUNCC(pred8x16_top_dc)(src, stride);
+    FUNCC(pred4x4_dc)(src, NULL, stride);
+}
+
 static void FUNC(pred8x8_mad_cow_dc_0lt)(uint8_t *src, int stride){
     FUNCC(pred8x8_dc)(src, stride);
     FUNCC(pred4x4_top_dc)(src, NULL, stride);
 }
 
+static void FUNC(pred8x16_mad_cow_dc_0lt)(uint8_t *src, int stride){
+    FUNCC(pred8x16_dc)(src, stride);
+    FUNCC(pred4x4_top_dc)(src, NULL, stride);
+}
+
 static void FUNC(pred8x8_mad_cow_dc_l00)(uint8_t *src, int stride){
     FUNCC(pred8x8_left_dc)(src, stride);
     FUNCC(pred4x4_128_dc)(src + 4*stride                  , NULL, stride);
     FUNCC(pred4x4_128_dc)(src + 4*stride + 4*sizeof(pixel), NULL, stride);
 }
 
+static void FUNC(pred8x16_mad_cow_dc_l00)(uint8_t *src, int stride){
+    FUNCC(pred8x16_left_dc)(src, stride);
+    FUNCC(pred4x4_128_dc)(src + 4*stride                  , NULL, stride);
+    FUNCC(pred4x4_128_dc)(src + 4*stride + 4*sizeof(pixel), NULL, stride);
+}
+
 static void FUNC(pred8x8_mad_cow_dc_0l0)(uint8_t *src, int stride){
     FUNCC(pred8x8_left_dc)(src, stride);
     FUNCC(pred4x4_128_dc)(src                  , NULL, stride);
     FUNCC(pred4x4_128_dc)(src + 4*sizeof(pixel), NULL, stride);
 }
 
+static void FUNC(pred8x16_mad_cow_dc_0l0)(uint8_t *src, int stride){
+    FUNCC(pred8x16_left_dc)(src, stride);
+    FUNCC(pred4x4_128_dc)(src                  , NULL, stride);
+    FUNCC(pred4x4_128_dc)(src + 4*sizeof(pixel), NULL, stride);
+}
+
 static void FUNCC(pred8x8_plane)(uint8_t *_src, int _stride){
   int j, k;
   int a;
@@ -1126,8 +1148,24 @@ static void FUNCC(pred8x8_vertical_add)(uint8_t *pix, const int *block_offset, c
         FUNCC(pred4x4_vertical_add)(pix + block_offset[i], block + i*16*sizeof(pixel), stride);
 }
 
+static void FUNCC(pred8x16_vertical_add)(uint8_t *pix, const int *block_offset, const DCTELEM *block, int stride){
+    int i;
+    for(i=0; i<4; i++)
+        FUNCC(pred4x4_vertical_add)(pix + block_offset[i], block + i*16*sizeof(pixel), stride);
+    for(i=4; i<8; i++)
+        FUNCC(pred4x4_vertical_add)(pix + block_offset[i+4], block + i*16*sizeof(pixel), stride);
+}
+
 static void FUNCC(pred8x8_horizontal_add)(uint8_t *pix, const int *block_offset, const DCTELEM *block, int stride){
     int i;
     for(i=0; i<4; i++)
         FUNCC(pred4x4_horizontal_add)(pix + block_offset[i], block + i*16*sizeof(pixel), stride);
 }
+
+static void FUNCC(pred8x16_horizontal_add)(uint8_t *pix, const int *block_offset, const DCTELEM *block, int stride){
+    int i;
+    for(i=0; i<4; i++)
+        FUNCC(pred4x4_horizontal_add)(pix + block_offset[i], block + i*16*sizeof(pixel), stride);
+    for(i=4; i<8; i++)
+        FUNCC(pred4x4_horizontal_add)(pix + block_offset[i+4], block + i*16*sizeof(pixel), stride);
+}