/*****************************************************************************
- * macroblock.h: h264 encoder library
+ * macroblock.h: macroblock encoding
*****************************************************************************
- * Copyright (C) 2003-2008 x264 project
+ * Copyright (C) 2003-2016 x264 project
*
* Authors: Loren Merritt <lorenm@u.washington.edu>
* Laurent Aimar <fenrir@via.ecp.fr>
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111, USA.
+ *
+ * This program is also available under a commercial proprietary license.
+ * For more information, contact us at licensing@x264.com.
*****************************************************************************/
#ifndef X264_ENCODER_MACROBLOCK_H
#include "common/macroblock.h"
-extern const int x264_lambda2_tab[52];
-extern const uint8_t x264_lambda_tab[52];
+extern const int x264_lambda2_tab[QP_MAX_MAX+1];
+extern const uint16_t x264_lambda_tab[QP_MAX_MAX+1];
void x264_rdo_init( void );
#define x264_macroblock_probe_bskip( h )\
x264_macroblock_probe_skip( h, 1 )
-void x264_predict_lossless_8x8_chroma( x264_t *h, int i_mode );
-void x264_predict_lossless_4x4( x264_t *h, pixel *p_dst, int idx, int i_mode );
-void x264_predict_lossless_8x8( x264_t *h, pixel *p_dst, int idx, int i_mode, pixel edge[33] );
-void x264_predict_lossless_16x16( x264_t *h, int i_mode );
+void x264_predict_lossless_4x4( x264_t *h, pixel *p_dst, int p, int idx, int i_mode );
+void x264_predict_lossless_8x8( x264_t *h, pixel *p_dst, int p, int idx, int i_mode, pixel edge[36] );
+void x264_predict_lossless_16x16( x264_t *h, int p, int i_mode );
+void x264_predict_lossless_chroma( x264_t *h, int i_mode );
void x264_macroblock_encode ( x264_t *h );
void x264_macroblock_write_cabac ( x264_t *h, x264_cabac_t *cb );
void x264_macroblock_encode_p8x8( x264_t *h, int i8 );
void x264_macroblock_encode_p4x4( x264_t *h, int i4 );
-void x264_mb_encode_i4x4( x264_t *h, int idx, int i_qp );
-void x264_mb_encode_i8x8( x264_t *h, int idx, int i_qp );
-void x264_mb_encode_8x8_chroma( x264_t *h, int b_inter, int i_qp );
+void x264_mb_encode_chroma( x264_t *h, int b_inter, int i_qp );
void x264_cabac_mb_skip( x264_t *h, int b_skip );
-int x264_quant_dc_trellis( x264_t *h, dctcoef *dct, int i_quant_cat,
- int i_qp, int i_ctxBlockCat, int b_intra, int b_chroma );
+int x264_quant_luma_dc_trellis( x264_t *h, dctcoef *dct, int i_quant_cat, int i_qp,
+ int ctx_block_cat, int b_intra, int idx );
+int x264_quant_chroma_dc_trellis( x264_t *h, dctcoef *dct, int i_qp, int b_intra, int idx );
int x264_quant_4x4_trellis( x264_t *h, dctcoef *dct, int i_quant_cat,
- int i_qp, int i_ctxBlockCat, int b_intra, int b_chroma, int idx );
+ int i_qp, int ctx_block_cat, int b_intra, int b_chroma, int idx );
int x264_quant_8x8_trellis( x264_t *h, dctcoef *dct, int i_quant_cat,
- int i_qp, int b_intra, int idx );
+ int i_qp, int ctx_block_cat, int b_intra, int b_chroma, int idx );
void x264_noise_reduction_update( x264_t *h );
+static ALWAYS_INLINE int x264_quant_4x4( x264_t *h, dctcoef dct[16], int i_qp, int ctx_block_cat, int b_intra, int p, int idx )
+{
+ int i_quant_cat = b_intra ? (p?CQM_4IC:CQM_4IY) : (p?CQM_4PC:CQM_4PY);
+ if( h->mb.b_noise_reduction )
+ h->quantf.denoise_dct( dct, h->nr_residual_sum[0+!!p*2], h->nr_offset[0+!!p*2], 16 );
+ if( h->mb.b_trellis )
+ return x264_quant_4x4_trellis( h, dct, i_quant_cat, i_qp, ctx_block_cat, b_intra, !!p, idx+p*16 );
+ else
+ return h->quantf.quant_4x4( dct, h->quant4_mf[i_quant_cat][i_qp], h->quant4_bias[i_quant_cat][i_qp] );
+}
+
+static ALWAYS_INLINE int x264_quant_8x8( x264_t *h, dctcoef dct[64], int i_qp, int ctx_block_cat, int b_intra, int p, int idx )
+{
+ int i_quant_cat = b_intra ? (p?CQM_8IC:CQM_8IY) : (p?CQM_8PC:CQM_8PY);
+ if( h->mb.b_noise_reduction )
+ h->quantf.denoise_dct( dct, h->nr_residual_sum[1+!!p*2], h->nr_offset[1+!!p*2], 64 );
+ if( h->mb.b_trellis )
+ return x264_quant_8x8_trellis( h, dct, i_quant_cat, i_qp, ctx_block_cat, b_intra, !!p, idx+p*4 );
+ else
+ return h->quantf.quant_8x8( dct, h->quant8_mf[i_quant_cat][i_qp], h->quant8_bias[i_quant_cat][i_qp] );
+}
+
+#define STORE_8x8_NNZ( p, idx, nz )\
+do\
+{\
+ M16( &h->mb.cache.non_zero_count[x264_scan8[p*16+idx*4]+0] ) = (nz) * 0x0101;\
+ M16( &h->mb.cache.non_zero_count[x264_scan8[p*16+idx*4]+8] ) = (nz) * 0x0101;\
+} while(0)
+
+#define CLEAR_16x16_NNZ( p ) \
+do\
+{\
+ M32( &h->mb.cache.non_zero_count[x264_scan8[16*p] + 0*8] ) = 0;\
+ M32( &h->mb.cache.non_zero_count[x264_scan8[16*p] + 1*8] ) = 0;\
+ M32( &h->mb.cache.non_zero_count[x264_scan8[16*p] + 2*8] ) = 0;\
+ M32( &h->mb.cache.non_zero_count[x264_scan8[16*p] + 3*8] ) = 0;\
+} while(0)
+
+/* A special for loop that iterates branchlessly over each set
+ * bit in a 4-bit input. */
+#define FOREACH_BIT(idx,start,mask) for( int idx = start, msk = mask, skip; msk && (skip = x264_ctz_4bit(msk), idx += skip, msk >>= skip+1, 1); idx++ )
+
+static ALWAYS_INLINE void x264_mb_encode_i4x4( x264_t *h, int p, int idx, int i_qp, int i_mode, int b_predict )
+{
+ int nz;
+ pixel *p_src = &h->mb.pic.p_fenc[p][block_idx_xy_fenc[idx]];
+ pixel *p_dst = &h->mb.pic.p_fdec[p][block_idx_xy_fdec[idx]];
+ ALIGNED_ARRAY_N( dctcoef, dct4x4,[16] );
+
+ if( b_predict )
+ {
+ if( h->mb.b_lossless )
+ x264_predict_lossless_4x4( h, p_dst, p, idx, i_mode );
+ else
+ h->predict_4x4[i_mode]( p_dst );
+ }
+
+ if( h->mb.b_lossless )
+ {
+ nz = h->zigzagf.sub_4x4( h->dct.luma4x4[p*16+idx], p_src, p_dst );
+ h->mb.cache.non_zero_count[x264_scan8[p*16+idx]] = nz;
+ h->mb.i_cbp_luma |= nz<<(idx>>2);
+ return;
+ }
+
+ h->dctf.sub4x4_dct( dct4x4, p_src, p_dst );
+
+ nz = x264_quant_4x4( h, dct4x4, i_qp, ctx_cat_plane[DCT_LUMA_4x4][p], 1, p, idx );
+ h->mb.cache.non_zero_count[x264_scan8[p*16+idx]] = nz;
+ if( nz )
+ {
+ h->mb.i_cbp_luma |= 1<<(idx>>2);
+ h->zigzagf.scan_4x4( h->dct.luma4x4[p*16+idx], dct4x4 );
+ h->quantf.dequant_4x4( dct4x4, h->dequant4_mf[p?CQM_4IC:CQM_4IY], i_qp );
+ h->dctf.add4x4_idct( p_dst, dct4x4 );
+ }
+}
+
+static ALWAYS_INLINE void x264_mb_encode_i8x8( x264_t *h, int p, int idx, int i_qp, int i_mode, pixel *edge, int b_predict )
+{
+ int x = idx&1;
+ int y = idx>>1;
+ int nz;
+ pixel *p_src = &h->mb.pic.p_fenc[p][8*x + 8*y*FENC_STRIDE];
+ pixel *p_dst = &h->mb.pic.p_fdec[p][8*x + 8*y*FDEC_STRIDE];
+ ALIGNED_ARRAY_N( dctcoef, dct8x8,[64] );
+ ALIGNED_ARRAY_32( pixel, edge_buf,[36] );
+
+ if( b_predict )
+ {
+ if( !edge )
+ {
+ h->predict_8x8_filter( p_dst, edge_buf, h->mb.i_neighbour8[idx], x264_pred_i4x4_neighbors[i_mode] );
+ edge = edge_buf;
+ }
+
+ if( h->mb.b_lossless )
+ x264_predict_lossless_8x8( h, p_dst, p, idx, i_mode, edge );
+ else
+ h->predict_8x8[i_mode]( p_dst, edge );
+ }
+
+ if( h->mb.b_lossless )
+ {
+ nz = h->zigzagf.sub_8x8( h->dct.luma8x8[p*4+idx], p_src, p_dst );
+ STORE_8x8_NNZ( p, idx, nz );
+ h->mb.i_cbp_luma |= nz<<idx;
+ return;
+ }
+
+ h->dctf.sub8x8_dct8( dct8x8, p_src, p_dst );
+
+ nz = x264_quant_8x8( h, dct8x8, i_qp, ctx_cat_plane[DCT_LUMA_8x8][p], 1, p, idx );
+ if( nz )
+ {
+ h->mb.i_cbp_luma |= 1<<idx;
+ h->zigzagf.scan_8x8( h->dct.luma8x8[p*4+idx], dct8x8 );
+ h->quantf.dequant_8x8( dct8x8, h->dequant8_mf[p?CQM_8IC:CQM_8IY], i_qp );
+ h->dctf.add8x8_idct8( p_dst, dct8x8 );
+ STORE_8x8_NNZ( p, idx, 1 );
+ }
+ else
+ STORE_8x8_NNZ( p, idx, 0 );
+}
+
#endif
projects / x264 / blobdiff