*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*
*/
MpegEncContext s;
int nal_ref_idc;
int nal_unit_type;
-#define NAL_SLICE 1
-#define NAL_DPA 2
-#define NAL_DPB 3
-#define NAL_DPC 4
-#define NAL_IDR_SLICE 5
-#define NAL_SEI 6
-#define NAL_SPS 7
-#define NAL_PPS 8
-#define NAL_AUD 9
+#define NAL_SLICE 1
+#define NAL_DPA 2
+#define NAL_DPB 3
+#define NAL_DPC 4
+#define NAL_IDR_SLICE 5
+#define NAL_SEI 6
+#define NAL_SPS 7
+#define NAL_PPS 8
+#define NAL_AUD 9
#define NAL_END_SEQUENCE 10
#define NAL_END_STREAM 11
#define NAL_FILLER_DATA 12
static void svq3_add_idct_c(uint8_t *dst, DCTELEM *block, int stride, int qp, int dc);
static void filter_mb( H264Context *h, int mb_x, int mb_y, uint8_t *img_y, uint8_t *img_cb, uint8_t *img_cr, unsigned int linesize, unsigned int uvlinesize);
-static inline uint32_t pack16to32(int a, int b){
+static always_inline uint32_t pack16to32(int a, int b){
#ifdef WORDS_BIGENDIAN
return (b&0xFFFF) + (a<<16);
#else
* @param w width of the rectangle, should be a constant
* @param size the size of val (1 or 4), should be a constant
*/
-static inline void fill_rectangle(void *vp, int w, int h, int stride, uint32_t val, int size){ //FIXME ensure this IS inlined
+static always_inline void fill_rectangle(void *vp, int w, int h, int stride, uint32_t val, int size){
uint8_t *p= (uint8_t*)vp;
assert(size==1 || size==4);
assert(0);
}
-static inline void fill_caches(H264Context *h, int mb_type, int for_deblock){
+static void fill_caches(H264Context *h, int mb_type, int for_deblock){
MpegEncContext * const s = &h->s;
const int mb_xy= s->mb_x + s->mb_y*s->mb_stride;
int topleft_xy, top_xy, topright_xy, left_xy[2];
}
if(IS_16X16(*mb_type)){
- fill_rectangle(&h->ref_cache[0][scan8[0]], 4, 4, 8, ref[0], 1);
- fill_rectangle(&h->ref_cache[1][scan8[0]], 4, 4, 8, ref[1], 1);
+ fill_rectangle(&h->ref_cache[0][scan8[0]], 4, 4, 8, (uint8_t)ref[0], 1);
+ fill_rectangle(&h->ref_cache[1][scan8[0]], 4, 4, 8, (uint8_t)ref[1], 1);
if(!IS_INTRA(mb_type_col)
&& ( (l1ref0[0] == 0 && ABS(l1mv0[0][0]) <= 1 && ABS(l1mv0[0][1]) <= 1)
|| (l1ref0[0] < 0 && l1ref1[0] == 0 && ABS(l1mv1[0][0]) <= 1 && ABS(l1mv1[0][1]) <= 1
fill_rectangle(&h->mv_cache[0][scan8[i8*4]], 2, 2, 8, pack16to32(mv[0][0],mv[0][1]), 4);
fill_rectangle(&h->mv_cache[1][scan8[i8*4]], 2, 2, 8, pack16to32(mv[1][0],mv[1][1]), 4);
- fill_rectangle(&h->ref_cache[0][scan8[i8*4]], 2, 2, 8, ref[0], 1);
- fill_rectangle(&h->ref_cache[1][scan8[i8*4]], 2, 2, 8, ref[1], 1);
+ fill_rectangle(&h->ref_cache[0][scan8[i8*4]], 2, 2, 8, (uint8_t)ref[0], 1);
+ fill_rectangle(&h->ref_cache[1][scan8[i8*4]], 2, 2, 8, (uint8_t)ref[1], 1);
/* col_zero_flag */
if(!IS_INTRA(mb_type_col) && ( l1ref0[x8 + y8*h->b8_stride] == 0
|| (l1ref0[x8 + y8*h->b8_stride] < 0 && l1ref1[x8 + y8*h->b8_stride] == 0
&& (h->x264_build>33 || !h->x264_build)))){
const int16_t (*l1mv)[2]= l1ref0[x8 + y8*h->b8_stride] == 0 ? l1mv0 : l1mv1;
+ if(IS_SUB_8X8(sub_mb_type)){
+ const int16_t *mv_col = l1mv[x8*3 + y8*3*h->b_stride];
+ if(ABS(mv_col[0]) <= 1 && ABS(mv_col[1]) <= 1){
+ if(ref[0] == 0)
+ fill_rectangle(&h->mv_cache[0][scan8[i8*4]], 2, 2, 8, 0, 4);
+ if(ref[1] == 0)
+ fill_rectangle(&h->mv_cache[1][scan8[i8*4]], 2, 2, 8, 0, 4);
+ }
+ }else
for(i4=0; i4<4; i4++){
const int16_t *mv_col = l1mv[x8*2 + (i4&1) + (y8*2 + (i4>>1))*h->b_stride];
if(ABS(mv_col[0]) <= 1 && ABS(mv_col[1]) <= 1){
fill_rectangle(&h->ref_cache[0][scan8[i8*4]], 2, 2, 8, ref0, 1);
fill_rectangle(&h->ref_cache[1][scan8[i8*4]], 2, 2, 8, 0, 1);
+ if(IS_SUB_8X8(sub_mb_type)){
+ const int16_t *mv_col = l1mv[x8*3 + y8*3*h->b_stride];
+ int mx = (dist_scale_factor * mv_col[0] + 128) >> 8;
+ int my = (dist_scale_factor * mv_col[1] + 128) >> 8;
+ fill_rectangle(&h->mv_cache[0][scan8[i8*4]], 2, 2, 8, pack16to32(mx,my), 4);
+ fill_rectangle(&h->mv_cache[1][scan8[i8*4]], 2, 2, 8, pack16to32(mx-mv_col[0],my-mv_col[1]), 4);
+ }else
for(i4=0; i4<4; i4++){
const int16_t *mv_col = l1mv[x8*2 + (i4&1) + (y8*2 + (i4>>1))*h->b_stride];
int16_t *mv_l0 = h->mv_cache[0][scan8[i8*4+i4]];
int i, si, di;
uint8_t *dst;
-// src[0]&0x80; //forbidden bit
+// src[0]&0x80; //forbidden bit
h->nal_ref_idc= src[0]>>5;
h->nal_unit_type= src[0]&0x1F;
if(h->use_weight == 2){
int weight0 = h->implicit_weight[refn0][refn1];
int weight1 = 64 - weight0;
- luma_weight_avg( dest_y, tmp_y, s-> linesize, 5, weight0, weight1, 0, 0);
- chroma_weight_avg(dest_cb, tmp_cb, s->uvlinesize, 5, weight0, weight1, 0, 0);
- chroma_weight_avg(dest_cr, tmp_cr, s->uvlinesize, 5, weight0, weight1, 0, 0);
+ luma_weight_avg( dest_y, tmp_y, s-> linesize, 5, weight0, weight1, 0);
+ chroma_weight_avg(dest_cb, tmp_cb, s->uvlinesize, 5, weight0, weight1, 0);
+ chroma_weight_avg(dest_cr, tmp_cr, s->uvlinesize, 5, weight0, weight1, 0);
}else{
luma_weight_avg(dest_y, tmp_y, s->linesize, h->luma_log2_weight_denom,
h->luma_weight[0][refn0], h->luma_weight[1][refn1],
- h->luma_offset[0][refn0], h->luma_offset[1][refn1]);
+ h->luma_offset[0][refn0] + h->luma_offset[1][refn1]);
chroma_weight_avg(dest_cb, tmp_cb, s->uvlinesize, h->chroma_log2_weight_denom,
h->chroma_weight[0][refn0][0], h->chroma_weight[1][refn1][0],
- h->chroma_offset[0][refn0][0], h->chroma_offset[1][refn1][0]);
+ h->chroma_offset[0][refn0][0] + h->chroma_offset[1][refn1][0]);
chroma_weight_avg(dest_cr, tmp_cr, s->uvlinesize, h->chroma_log2_weight_denom,
h->chroma_weight[0][refn0][1], h->chroma_weight[1][refn1][1],
- h->chroma_offset[0][refn0][1], h->chroma_offset[1][refn1][1]);
+ h->chroma_offset[0][refn0][1] + h->chroma_offset[1][refn1][1]);
}
}else{
int list = list1 ? 1 : 0;
static void init_dequant4_coeff_table(H264Context *h){
int i,j,q,x;
+ const int transpose = (h->s.dsp.h264_idct_add != ff_h264_idct_add_c); //FIXME ugly
for(i=0; i<6; i++ ){
h->dequant4_coeff[i] = h->dequant4_buffer[i];
for(j=0; j<i; j++){
int shift = div6[q] + 2;
int idx = rem6[q];
for(x=0; x<16; x++)
- h->dequant4_coeff[i][q][x] = ((uint32_t)dequant4_coeff_init[idx][(x&1) + ((x>>2)&1)] *
+ h->dequant4_coeff[i][q][transpose ? (x>>2)|((x<<2)&0xF) : x] =
+ ((uint32_t)dequant4_coeff_init[idx][(x&1) + ((x>>2)&1)] *
h->pps.scaling_matrix4[i][x]) << shift;
}
}
h->sub_mb_type[i]= b_sub_mb_type_info[ h->sub_mb_type[i] ].type;
}
if( IS_DIRECT(h->sub_mb_type[0]) || IS_DIRECT(h->sub_mb_type[1])
- || IS_DIRECT(h->sub_mb_type[2]) || IS_DIRECT(h->sub_mb_type[3]))
+ || IS_DIRECT(h->sub_mb_type[2]) || IS_DIRECT(h->sub_mb_type[3])) {
pred_direct_motion(h, &mb_type);
+ h->ref_cache[0][scan8[4]] =
+ h->ref_cache[1][scan8[4]] =
+ h->ref_cache[0][scan8[12]] =
+ h->ref_cache[1][scan8[12]] = PART_NOT_AVAILABLE;
+ }
}else{
assert(h->slice_type == P_TYPE || h->slice_type == SP_TYPE); //FIXME SP correct ?
for(i=0; i<4; i++){
if(ref_count == 0) continue;
for(i=0; i<4; i++){
- if(IS_DIRECT(h->sub_mb_type[i])) continue;
+ if(IS_DIRECT(h->sub_mb_type[i])) {
+ h->ref_cache[list][ scan8[4*i] ] = h->ref_cache[list][ scan8[4*i]+1 ];
+ continue;
+ }
h->ref_cache[list][ scan8[4*i] ]=h->ref_cache[list][ scan8[4*i]+1 ]=
h->ref_cache[list][ scan8[4*i]+8 ]=h->ref_cache[list][ scan8[4*i]+9 ]= ref[list][i];
return ctx + 4 * cat;
}
-static int inline decode_cabac_residual( H264Context *h, DCTELEM *block, int cat, int n, const uint8_t *scantable, const uint32_t *qmul, int max_coeff) {
+static int decode_cabac_residual( H264Context *h, DCTELEM *block, int cat, int n, const uint8_t *scantable, const uint32_t *qmul, int max_coeff) {
const int mb_xy = h->s.mb_x + h->s.mb_y*h->s.mb_stride;
static const int significant_coeff_flag_field_offset[2] = { 105, 277 };
static const int last_significant_coeff_flag_field_offset[2] = { 166, 338 };
return 0;
}
-static void decode_scaling_list(H264Context *h, uint8_t *factors, int size, const uint8_t *default_list){
+static void decode_scaling_list(H264Context *h, uint8_t *factors, int size,
+ const uint8_t *jvt_list, const uint8_t *fallback_list){
MpegEncContext * const s = &h->s;
int i, last = 8, next = 8;
const uint8_t *scan = size == 16 ? zigzag_scan : zigzag_scan8x8;
- if(!get_bits1(&s->gb)) /* matrix not written, we use the default one */
- memcpy(factors, default_list, size*sizeof(uint8_t));
+ if(!get_bits1(&s->gb)) /* matrix not written, we use the predicted one */
+ memcpy(factors, fallback_list, size*sizeof(uint8_t));
else
for(i=0;i<size;i++){
if(next)
next = (last + get_se_golomb(&s->gb)) & 0xff;
- if(!i && !next){ /* matrix not written, we use the default one */
- memcpy(factors, default_list, size*sizeof(uint8_t));
+ if(!i && !next){ /* matrix not written, we use the preset one */
+ memcpy(factors, jvt_list, size*sizeof(uint8_t));
break;
}
last = factors[scan[i]] = next ? next : last;
};
if(get_bits1(&s->gb)){
sps->scaling_matrix_present |= is_sps;
- decode_scaling_list(h,scaling_matrix4[0],16,fallback[0]); // Intra, Y
- decode_scaling_list(h,scaling_matrix4[1],16,scaling_matrix4[0]); // Intra, Cr
- decode_scaling_list(h,scaling_matrix4[2],16,scaling_matrix4[1]); // Intra, Cb
- decode_scaling_list(h,scaling_matrix4[3],16,fallback[1]); // Inter, Y
- decode_scaling_list(h,scaling_matrix4[4],16,scaling_matrix4[3]); // Inter, Cr
- decode_scaling_list(h,scaling_matrix4[5],16,scaling_matrix4[4]); // Inter, Cb
+ decode_scaling_list(h,scaling_matrix4[0],16,default_scaling4[0],fallback[0]); // Intra, Y
+ decode_scaling_list(h,scaling_matrix4[1],16,default_scaling4[0],scaling_matrix4[0]); // Intra, Cr
+ decode_scaling_list(h,scaling_matrix4[2],16,default_scaling4[0],scaling_matrix4[1]); // Intra, Cb
+ decode_scaling_list(h,scaling_matrix4[3],16,default_scaling4[1],fallback[1]); // Inter, Y
+ decode_scaling_list(h,scaling_matrix4[4],16,default_scaling4[1],scaling_matrix4[3]); // Inter, Cr
+ decode_scaling_list(h,scaling_matrix4[5],16,default_scaling4[1],scaling_matrix4[4]); // Inter, Cb
if(is_sps || pps->transform_8x8_mode){
- decode_scaling_list(h,scaling_matrix8[0],64,fallback[2]); // Intra, Y
- decode_scaling_list(h,scaling_matrix8[1],64,fallback[3]); // Inter, Y
+ decode_scaling_list(h,scaling_matrix8[0],64,default_scaling8[0],fallback[2]); // Intra, Y
+ decode_scaling_list(h,scaling_matrix8[1],64,default_scaling8[1],fallback[3]); // Inter, Y
}
} else if(fallback_sps) {
memcpy(scaling_matrix4, sps->scaling_matrix4, 6*16*sizeof(uint8_t));
case NAL_SPS_EXT:
case NAL_AUXILIARY_SLICE:
break;
- default:
- av_log(avctx, AV_LOG_ERROR, "Unknown NAL code: %d\n", h->nal_unit_type);
+ default:
+ av_log(avctx, AV_LOG_ERROR, "Unknown NAL code: %d\n", h->nal_unit_type);
}
}