Mercurial > libavcodec.hg
diff h264.c @ 5638:4a26dc4ca11d libavcodec
Move H.264 intra prediction functions into their own context
| author | kostya |
|---|---|
| date | Wed, 05 Sep 2007 05:30:08 +0000 |
| parents | 550e37ec2204 |
| children | d2598034f2a9 |
line wrap: on
line diff
--- a/h264.c Wed Sep 05 01:49:14 2007 +0000 +++ b/h264.c Wed Sep 05 05:30:08 2007 +0000 @@ -1681,719 +1681,6 @@ return last_non_zero; } -static void pred4x4_vertical_c(uint8_t *src, uint8_t *topright, int stride){ - const uint32_t a= ((uint32_t*)(src-stride))[0]; - ((uint32_t*)(src+0*stride))[0]= a; - ((uint32_t*)(src+1*stride))[0]= a; - ((uint32_t*)(src+2*stride))[0]= a; - ((uint32_t*)(src+3*stride))[0]= a; -} - -static void pred4x4_horizontal_c(uint8_t *src, uint8_t *topright, int stride){ - ((uint32_t*)(src+0*stride))[0]= src[-1+0*stride]*0x01010101; - ((uint32_t*)(src+1*stride))[0]= src[-1+1*stride]*0x01010101; - ((uint32_t*)(src+2*stride))[0]= src[-1+2*stride]*0x01010101; - ((uint32_t*)(src+3*stride))[0]= src[-1+3*stride]*0x01010101; -} - -static void pred4x4_dc_c(uint8_t *src, uint8_t *topright, int stride){ - const int dc= ( src[-stride] + src[1-stride] + src[2-stride] + src[3-stride] - + src[-1+0*stride] + src[-1+1*stride] + src[-1+2*stride] + src[-1+3*stride] + 4) >>3; - - ((uint32_t*)(src+0*stride))[0]= - ((uint32_t*)(src+1*stride))[0]= - ((uint32_t*)(src+2*stride))[0]= - ((uint32_t*)(src+3*stride))[0]= dc* 0x01010101; -} - -static void pred4x4_left_dc_c(uint8_t *src, uint8_t *topright, int stride){ - const int dc= ( src[-1+0*stride] + src[-1+1*stride] + src[-1+2*stride] + src[-1+3*stride] + 2) >>2; - - ((uint32_t*)(src+0*stride))[0]= - ((uint32_t*)(src+1*stride))[0]= - ((uint32_t*)(src+2*stride))[0]= - ((uint32_t*)(src+3*stride))[0]= dc* 0x01010101; -} - -static void pred4x4_top_dc_c(uint8_t *src, uint8_t *topright, int stride){ - const int dc= ( src[-stride] + src[1-stride] + src[2-stride] + src[3-stride] + 2) >>2; - - ((uint32_t*)(src+0*stride))[0]= - ((uint32_t*)(src+1*stride))[0]= - ((uint32_t*)(src+2*stride))[0]= - ((uint32_t*)(src+3*stride))[0]= dc* 0x01010101; -} - -static void pred4x4_128_dc_c(uint8_t *src, uint8_t *topright, int stride){ - ((uint32_t*)(src+0*stride))[0]= - ((uint32_t*)(src+1*stride))[0]= - ((uint32_t*)(src+2*stride))[0]= - ((uint32_t*)(src+3*stride))[0]= 128U*0x01010101U; -} - - -#define LOAD_TOP_RIGHT_EDGE\ - const int av_unused t4= topright[0];\ - const int av_unused t5= topright[1];\ - const int av_unused t6= topright[2];\ - const int av_unused t7= topright[3];\ - -#define LOAD_LEFT_EDGE\ - const int av_unused l0= src[-1+0*stride];\ - const int av_unused l1= src[-1+1*stride];\ - const int av_unused l2= src[-1+2*stride];\ - const int av_unused l3= src[-1+3*stride];\ - -#define LOAD_TOP_EDGE\ - const int av_unused t0= src[ 0-1*stride];\ - const int av_unused t1= src[ 1-1*stride];\ - const int av_unused t2= src[ 2-1*stride];\ - const int av_unused t3= src[ 3-1*stride];\ - -static void pred4x4_down_right_c(uint8_t *src, uint8_t *topright, int stride){ - const int lt= src[-1-1*stride]; - LOAD_TOP_EDGE - LOAD_LEFT_EDGE - - src[0+3*stride]=(l3 + 2*l2 + l1 + 2)>>2; - src[0+2*stride]= - src[1+3*stride]=(l2 + 2*l1 + l0 + 2)>>2; - src[0+1*stride]= - src[1+2*stride]= - src[2+3*stride]=(l1 + 2*l0 + lt + 2)>>2; - src[0+0*stride]= - src[1+1*stride]= - src[2+2*stride]= - src[3+3*stride]=(l0 + 2*lt + t0 + 2)>>2; - src[1+0*stride]= - src[2+1*stride]= - src[3+2*stride]=(lt + 2*t0 + t1 + 2)>>2; - src[2+0*stride]= - src[3+1*stride]=(t0 + 2*t1 + t2 + 2)>>2; - src[3+0*stride]=(t1 + 2*t2 + t3 + 2)>>2; -} - -static void pred4x4_down_left_c(uint8_t *src, uint8_t *topright, int stride){ - LOAD_TOP_EDGE - LOAD_TOP_RIGHT_EDGE -// LOAD_LEFT_EDGE - - src[0+0*stride]=(t0 + t2 + 2*t1 + 2)>>2; - src[1+0*stride]= - src[0+1*stride]=(t1 + t3 + 2*t2 + 2)>>2; - src[2+0*stride]= - src[1+1*stride]= - src[0+2*stride]=(t2 + t4 + 2*t3 + 2)>>2; - src[3+0*stride]= - src[2+1*stride]= - src[1+2*stride]= - src[0+3*stride]=(t3 + t5 + 2*t4 + 2)>>2; - src[3+1*stride]= - src[2+2*stride]= - src[1+3*stride]=(t4 + t6 + 2*t5 + 2)>>2; - src[3+2*stride]= - src[2+3*stride]=(t5 + t7 + 2*t6 + 2)>>2; - src[3+3*stride]=(t6 + 3*t7 + 2)>>2; -} - -static void pred4x4_vertical_right_c(uint8_t *src, uint8_t *topright, int stride){ - const int lt= src[-1-1*stride]; - LOAD_TOP_EDGE - LOAD_LEFT_EDGE - - src[0+0*stride]= - src[1+2*stride]=(lt + t0 + 1)>>1; - src[1+0*stride]= - src[2+2*stride]=(t0 + t1 + 1)>>1; - src[2+0*stride]= - src[3+2*stride]=(t1 + t2 + 1)>>1; - src[3+0*stride]=(t2 + t3 + 1)>>1; - src[0+1*stride]= - src[1+3*stride]=(l0 + 2*lt + t0 + 2)>>2; - src[1+1*stride]= - src[2+3*stride]=(lt + 2*t0 + t1 + 2)>>2; - src[2+1*stride]= - src[3+3*stride]=(t0 + 2*t1 + t2 + 2)>>2; - src[3+1*stride]=(t1 + 2*t2 + t3 + 2)>>2; - src[0+2*stride]=(lt + 2*l0 + l1 + 2)>>2; - src[0+3*stride]=(l0 + 2*l1 + l2 + 2)>>2; -} - -static void pred4x4_vertical_left_c(uint8_t *src, uint8_t *topright, int stride){ - LOAD_TOP_EDGE - LOAD_TOP_RIGHT_EDGE - - src[0+0*stride]=(t0 + t1 + 1)>>1; - src[1+0*stride]= - src[0+2*stride]=(t1 + t2 + 1)>>1; - src[2+0*stride]= - src[1+2*stride]=(t2 + t3 + 1)>>1; - src[3+0*stride]= - src[2+2*stride]=(t3 + t4+ 1)>>1; - src[3+2*stride]=(t4 + t5+ 1)>>1; - src[0+1*stride]=(t0 + 2*t1 + t2 + 2)>>2; - src[1+1*stride]= - src[0+3*stride]=(t1 + 2*t2 + t3 + 2)>>2; - src[2+1*stride]= - src[1+3*stride]=(t2 + 2*t3 + t4 + 2)>>2; - src[3+1*stride]= - src[2+3*stride]=(t3 + 2*t4 + t5 + 2)>>2; - src[3+3*stride]=(t4 + 2*t5 + t6 + 2)>>2; -} - -static void pred4x4_horizontal_up_c(uint8_t *src, uint8_t *topright, int stride){ - LOAD_LEFT_EDGE - - src[0+0*stride]=(l0 + l1 + 1)>>1; - src[1+0*stride]=(l0 + 2*l1 + l2 + 2)>>2; - src[2+0*stride]= - src[0+1*stride]=(l1 + l2 + 1)>>1; - src[3+0*stride]= - src[1+1*stride]=(l1 + 2*l2 + l3 + 2)>>2; - src[2+1*stride]= - src[0+2*stride]=(l2 + l3 + 1)>>1; - src[3+1*stride]= - src[1+2*stride]=(l2 + 2*l3 + l3 + 2)>>2; - src[3+2*stride]= - src[1+3*stride]= - src[0+3*stride]= - src[2+2*stride]= - src[2+3*stride]= - src[3+3*stride]=l3; -} - -static void pred4x4_horizontal_down_c(uint8_t *src, uint8_t *topright, int stride){ - const int lt= src[-1-1*stride]; - LOAD_TOP_EDGE - LOAD_LEFT_EDGE - - src[0+0*stride]= - src[2+1*stride]=(lt + l0 + 1)>>1; - src[1+0*stride]= - src[3+1*stride]=(l0 + 2*lt + t0 + 2)>>2; - src[2+0*stride]=(lt + 2*t0 + t1 + 2)>>2; - src[3+0*stride]=(t0 + 2*t1 + t2 + 2)>>2; - src[0+1*stride]= - src[2+2*stride]=(l0 + l1 + 1)>>1; - src[1+1*stride]= - src[3+2*stride]=(lt + 2*l0 + l1 + 2)>>2; - src[0+2*stride]= - src[2+3*stride]=(l1 + l2+ 1)>>1; - src[1+2*stride]= - src[3+3*stride]=(l0 + 2*l1 + l2 + 2)>>2; - src[0+3*stride]=(l2 + l3 + 1)>>1; - src[1+3*stride]=(l1 + 2*l2 + l3 + 2)>>2; -} - -void ff_pred16x16_vertical_c(uint8_t *src, int stride){ - int i; - const uint32_t a= ((uint32_t*)(src-stride))[0]; - const uint32_t b= ((uint32_t*)(src-stride))[1]; - const uint32_t c= ((uint32_t*)(src-stride))[2]; - const uint32_t d= ((uint32_t*)(src-stride))[3]; - - for(i=0; i<16; i++){ - ((uint32_t*)(src+i*stride))[0]= a; - ((uint32_t*)(src+i*stride))[1]= b; - ((uint32_t*)(src+i*stride))[2]= c; - ((uint32_t*)(src+i*stride))[3]= d; - } -} - -void ff_pred16x16_horizontal_c(uint8_t *src, int stride){ - int i; - - for(i=0; i<16; i++){ - ((uint32_t*)(src+i*stride))[0]= - ((uint32_t*)(src+i*stride))[1]= - ((uint32_t*)(src+i*stride))[2]= - ((uint32_t*)(src+i*stride))[3]= src[-1+i*stride]*0x01010101; - } -} - -void ff_pred16x16_dc_c(uint8_t *src, int stride){ - int i, dc=0; - - for(i=0;i<16; i++){ - dc+= src[-1+i*stride]; - } - - for(i=0;i<16; i++){ - dc+= src[i-stride]; - } - - dc= 0x01010101*((dc + 16)>>5); - - for(i=0; i<16; i++){ - ((uint32_t*)(src+i*stride))[0]= - ((uint32_t*)(src+i*stride))[1]= - ((uint32_t*)(src+i*stride))[2]= - ((uint32_t*)(src+i*stride))[3]= dc; - } -} - -void ff_pred16x16_left_dc_c(uint8_t *src, int stride){ - int i, dc=0; - - for(i=0;i<16; i++){ - dc+= src[-1+i*stride]; - } - - dc= 0x01010101*((dc + 8)>>4); - - for(i=0; i<16; i++){ - ((uint32_t*)(src+i*stride))[0]= - ((uint32_t*)(src+i*stride))[1]= - ((uint32_t*)(src+i*stride))[2]= - ((uint32_t*)(src+i*stride))[3]= dc; - } -} - -void ff_pred16x16_top_dc_c(uint8_t *src, int stride){ - int i, dc=0; - - for(i=0;i<16; i++){ - dc+= src[i-stride]; - } - dc= 0x01010101*((dc + 8)>>4); - - for(i=0; i<16; i++){ - ((uint32_t*)(src+i*stride))[0]= - ((uint32_t*)(src+i*stride))[1]= - ((uint32_t*)(src+i*stride))[2]= - ((uint32_t*)(src+i*stride))[3]= dc; - } -} - -void ff_pred16x16_128_dc_c(uint8_t *src, int stride){ - int i; - - for(i=0; i<16; i++){ - ((uint32_t*)(src+i*stride))[0]= - ((uint32_t*)(src+i*stride))[1]= - ((uint32_t*)(src+i*stride))[2]= - ((uint32_t*)(src+i*stride))[3]= 0x01010101U*128U; - } -} - -static inline void pred16x16_plane_compat_c(uint8_t *src, int stride, const int svq3){ - int i, j, k; - int a; - uint8_t *cm = ff_cropTbl + MAX_NEG_CROP; - const uint8_t * const src0 = src+7-stride; - const uint8_t *src1 = src+8*stride-1; - const uint8_t *src2 = src1-2*stride; // == src+6*stride-1; - int H = src0[1] - src0[-1]; - int V = src1[0] - src2[ 0]; - for(k=2; k<=8; ++k) { - src1 += stride; src2 -= stride; - H += k*(src0[k] - src0[-k]); - V += k*(src1[0] - src2[ 0]); - } - if(svq3){ - H = ( 5*(H/4) ) / 16; - V = ( 5*(V/4) ) / 16; - - /* required for 100% accuracy */ - i = H; H = V; V = i; - }else{ - H = ( 5*H+32 ) >> 6; - V = ( 5*V+32 ) >> 6; - } - - a = 16*(src1[0] + src2[16] + 1) - 7*(V+H); - for(j=16; j>0; --j) { - int b = a; - a += V; - for(i=-16; i<0; i+=4) { - src[16+i] = cm[ (b ) >> 5 ]; - src[17+i] = cm[ (b+ H) >> 5 ]; - src[18+i] = cm[ (b+2*H) >> 5 ]; - src[19+i] = cm[ (b+3*H) >> 5 ]; - b += 4*H; - } - src += stride; - } -} - -void ff_pred16x16_plane_c(uint8_t *src, int stride){ - pred16x16_plane_compat_c(src, stride, 0); -} - -void ff_pred8x8_vertical_c(uint8_t *src, int stride){ - int i; - const uint32_t a= ((uint32_t*)(src-stride))[0]; - const uint32_t b= ((uint32_t*)(src-stride))[1]; - - for(i=0; i<8; i++){ - ((uint32_t*)(src+i*stride))[0]= a; - ((uint32_t*)(src+i*stride))[1]= b; - } -} - -void ff_pred8x8_horizontal_c(uint8_t *src, int stride){ - int i; - - for(i=0; i<8; i++){ - ((uint32_t*)(src+i*stride))[0]= - ((uint32_t*)(src+i*stride))[1]= src[-1+i*stride]*0x01010101; - } -} - -void ff_pred8x8_128_dc_c(uint8_t *src, int stride){ - int i; - - for(i=0; i<8; i++){ - ((uint32_t*)(src+i*stride))[0]= - ((uint32_t*)(src+i*stride))[1]= 0x01010101U*128U; - } -} - -void ff_pred8x8_left_dc_c(uint8_t *src, int stride){ - int i; - int dc0, dc2; - - dc0=dc2=0; - for(i=0;i<4; i++){ - dc0+= src[-1+i*stride]; - dc2+= src[-1+(i+4)*stride]; - } - dc0= 0x01010101*((dc0 + 2)>>2); - dc2= 0x01010101*((dc2 + 2)>>2); - - for(i=0; i<4; i++){ - ((uint32_t*)(src+i*stride))[0]= - ((uint32_t*)(src+i*stride))[1]= dc0; - } - for(i=4; i<8; i++){ - ((uint32_t*)(src+i*stride))[0]= - ((uint32_t*)(src+i*stride))[1]= dc2; - } -} - -void ff_pred8x8_top_dc_c(uint8_t *src, int stride){ - int i; - int dc0, dc1; - - dc0=dc1=0; - for(i=0;i<4; i++){ - dc0+= src[i-stride]; - dc1+= src[4+i-stride]; - } - dc0= 0x01010101*((dc0 + 2)>>2); - dc1= 0x01010101*((dc1 + 2)>>2); - - for(i=0; i<4; i++){ - ((uint32_t*)(src+i*stride))[0]= dc0; - ((uint32_t*)(src+i*stride))[1]= dc1; - } - for(i=4; i<8; i++){ - ((uint32_t*)(src+i*stride))[0]= dc0; - ((uint32_t*)(src+i*stride))[1]= dc1; - } -} - - -void ff_pred8x8_dc_c(uint8_t *src, int stride){ - int i; - int dc0, dc1, dc2, dc3; - - dc0=dc1=dc2=0; - for(i=0;i<4; i++){ - dc0+= src[-1+i*stride] + src[i-stride]; - dc1+= src[4+i-stride]; - dc2+= src[-1+(i+4)*stride]; - } - dc3= 0x01010101*((dc1 + dc2 + 4)>>3); - dc0= 0x01010101*((dc0 + 4)>>3); - dc1= 0x01010101*((dc1 + 2)>>2); - dc2= 0x01010101*((dc2 + 2)>>2); - - for(i=0; i<4; i++){ - ((uint32_t*)(src+i*stride))[0]= dc0; - ((uint32_t*)(src+i*stride))[1]= dc1; - } - for(i=4; i<8; i++){ - ((uint32_t*)(src+i*stride))[0]= dc2; - ((uint32_t*)(src+i*stride))[1]= dc3; - } -} - -void ff_pred8x8_plane_c(uint8_t *src, int stride){ - int j, k; - int a; - uint8_t *cm = ff_cropTbl + MAX_NEG_CROP; - const uint8_t * const src0 = src+3-stride; - const uint8_t *src1 = src+4*stride-1; - const uint8_t *src2 = src1-2*stride; // == src+2*stride-1; - int H = src0[1] - src0[-1]; - int V = src1[0] - src2[ 0]; - for(k=2; k<=4; ++k) { - src1 += stride; src2 -= stride; - H += k*(src0[k] - src0[-k]); - V += k*(src1[0] - src2[ 0]); - } - H = ( 17*H+16 ) >> 5; - V = ( 17*V+16 ) >> 5; - - a = 16*(src1[0] + src2[8]+1) - 3*(V+H); - for(j=8; j>0; --j) { - int b = a; - a += V; - src[0] = cm[ (b ) >> 5 ]; - src[1] = cm[ (b+ H) >> 5 ]; - src[2] = cm[ (b+2*H) >> 5 ]; - src[3] = cm[ (b+3*H) >> 5 ]; - src[4] = cm[ (b+4*H) >> 5 ]; - src[5] = cm[ (b+5*H) >> 5 ]; - src[6] = cm[ (b+6*H) >> 5 ]; - src[7] = cm[ (b+7*H) >> 5 ]; - src += stride; - } -} - -#define SRC(x,y) src[(x)+(y)*stride] -#define PL(y) \ - const int l##y = (SRC(-1,y-1) + 2*SRC(-1,y) + SRC(-1,y+1) + 2) >> 2; -#define PREDICT_8x8_LOAD_LEFT \ - const int l0 = ((has_topleft ? SRC(-1,-1) : SRC(-1,0)) \ - + 2*SRC(-1,0) + SRC(-1,1) + 2) >> 2; \ - PL(1) PL(2) PL(3) PL(4) PL(5) PL(6) \ - const int l7 av_unused = (SRC(-1,6) + 3*SRC(-1,7) + 2) >> 2 - -#define PT(x) \ - const int t##x = (SRC(x-1,-1) + 2*SRC(x,-1) + SRC(x+1,-1) + 2) >> 2; -#define PREDICT_8x8_LOAD_TOP \ - const int t0 = ((has_topleft ? SRC(-1,-1) : SRC(0,-1)) \ - + 2*SRC(0,-1) + SRC(1,-1) + 2) >> 2; \ - PT(1) PT(2) PT(3) PT(4) PT(5) PT(6) \ - const int t7 av_unused = ((has_topright ? SRC(8,-1) : SRC(7,-1)) \ - + 2*SRC(7,-1) + SRC(6,-1) + 2) >> 2 - -#define PTR(x) \ - t##x = (SRC(x-1,-1) + 2*SRC(x,-1) + SRC(x+1,-1) + 2) >> 2; -#define PREDICT_8x8_LOAD_TOPRIGHT \ - int t8, t9, t10, t11, t12, t13, t14, t15; \ - if(has_topright) { \ - PTR(8) PTR(9) PTR(10) PTR(11) PTR(12) PTR(13) PTR(14) \ - t15 = (SRC(14,-1) + 3*SRC(15,-1) + 2) >> 2; \ - } else t8=t9=t10=t11=t12=t13=t14=t15= SRC(7,-1); - -#define PREDICT_8x8_LOAD_TOPLEFT \ - const int lt = (SRC(-1,0) + 2*SRC(-1,-1) + SRC(0,-1) + 2) >> 2 - -#define PREDICT_8x8_DC(v) \ - int y; \ - for( y = 0; y < 8; y++ ) { \ - ((uint32_t*)src)[0] = \ - ((uint32_t*)src)[1] = v; \ - src += stride; \ - } - -static void pred8x8l_128_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride) -{ - PREDICT_8x8_DC(0x80808080); -} -static void pred8x8l_left_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride) -{ - PREDICT_8x8_LOAD_LEFT; - const uint32_t dc = ((l0+l1+l2+l3+l4+l5+l6+l7+4) >> 3) * 0x01010101; - PREDICT_8x8_DC(dc); -} -static void pred8x8l_top_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride) -{ - PREDICT_8x8_LOAD_TOP; - const uint32_t dc = ((t0+t1+t2+t3+t4+t5+t6+t7+4) >> 3) * 0x01010101; - PREDICT_8x8_DC(dc); -} -static void pred8x8l_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride) -{ - PREDICT_8x8_LOAD_LEFT; - PREDICT_8x8_LOAD_TOP; - const uint32_t dc = ((l0+l1+l2+l3+l4+l5+l6+l7 - +t0+t1+t2+t3+t4+t5+t6+t7+8) >> 4) * 0x01010101; - PREDICT_8x8_DC(dc); -} -static void pred8x8l_horizontal_c(uint8_t *src, int has_topleft, int has_topright, int stride) -{ - PREDICT_8x8_LOAD_LEFT; -#define ROW(y) ((uint32_t*)(src+y*stride))[0] =\ - ((uint32_t*)(src+y*stride))[1] = 0x01010101 * l##y - ROW(0); ROW(1); ROW(2); ROW(3); ROW(4); ROW(5); ROW(6); ROW(7); -#undef ROW -} -static void pred8x8l_vertical_c(uint8_t *src, int has_topleft, int has_topright, int stride) -{ - int y; - PREDICT_8x8_LOAD_TOP; - src[0] = t0; - src[1] = t1; - src[2] = t2; - src[3] = t3; - src[4] = t4; - src[5] = t5; - src[6] = t6; - src[7] = t7; - for( y = 1; y < 8; y++ ) - *(uint64_t*)(src+y*stride) = *(uint64_t*)src; -} -static void pred8x8l_down_left_c(uint8_t *src, int has_topleft, int has_topright, int stride) -{ - PREDICT_8x8_LOAD_TOP; - PREDICT_8x8_LOAD_TOPRIGHT; - SRC(0,0)= (t0 + 2*t1 + t2 + 2) >> 2; - SRC(0,1)=SRC(1,0)= (t1 + 2*t2 + t3 + 2) >> 2; - SRC(0,2)=SRC(1,1)=SRC(2,0)= (t2 + 2*t3 + t4 + 2) >> 2; - SRC(0,3)=SRC(1,2)=SRC(2,1)=SRC(3,0)= (t3 + 2*t4 + t5 + 2) >> 2; - SRC(0,4)=SRC(1,3)=SRC(2,2)=SRC(3,1)=SRC(4,0)= (t4 + 2*t5 + t6 + 2) >> 2; - SRC(0,5)=SRC(1,4)=SRC(2,3)=SRC(3,2)=SRC(4,1)=SRC(5,0)= (t5 + 2*t6 + t7 + 2) >> 2; - SRC(0,6)=SRC(1,5)=SRC(2,4)=SRC(3,3)=SRC(4,2)=SRC(5,1)=SRC(6,0)= (t6 + 2*t7 + t8 + 2) >> 2; - SRC(0,7)=SRC(1,6)=SRC(2,5)=SRC(3,4)=SRC(4,3)=SRC(5,2)=SRC(6,1)=SRC(7,0)= (t7 + 2*t8 + t9 + 2) >> 2; - SRC(1,7)=SRC(2,6)=SRC(3,5)=SRC(4,4)=SRC(5,3)=SRC(6,2)=SRC(7,1)= (t8 + 2*t9 + t10 + 2) >> 2; - SRC(2,7)=SRC(3,6)=SRC(4,5)=SRC(5,4)=SRC(6,3)=SRC(7,2)= (t9 + 2*t10 + t11 + 2) >> 2; - SRC(3,7)=SRC(4,6)=SRC(5,5)=SRC(6,4)=SRC(7,3)= (t10 + 2*t11 + t12 + 2) >> 2; - SRC(4,7)=SRC(5,6)=SRC(6,5)=SRC(7,4)= (t11 + 2*t12 + t13 + 2) >> 2; - SRC(5,7)=SRC(6,6)=SRC(7,5)= (t12 + 2*t13 + t14 + 2) >> 2; - SRC(6,7)=SRC(7,6)= (t13 + 2*t14 + t15 + 2) >> 2; - SRC(7,7)= (t14 + 3*t15 + 2) >> 2; -} -static void pred8x8l_down_right_c(uint8_t *src, int has_topleft, int has_topright, int stride) -{ - PREDICT_8x8_LOAD_TOP; - PREDICT_8x8_LOAD_LEFT; - PREDICT_8x8_LOAD_TOPLEFT; - SRC(0,7)= (l7 + 2*l6 + l5 + 2) >> 2; - SRC(0,6)=SRC(1,7)= (l6 + 2*l5 + l4 + 2) >> 2; - SRC(0,5)=SRC(1,6)=SRC(2,7)= (l5 + 2*l4 + l3 + 2) >> 2; - SRC(0,4)=SRC(1,5)=SRC(2,6)=SRC(3,7)= (l4 + 2*l3 + l2 + 2) >> 2; - SRC(0,3)=SRC(1,4)=SRC(2,5)=SRC(3,6)=SRC(4,7)= (l3 + 2*l2 + l1 + 2) >> 2; - SRC(0,2)=SRC(1,3)=SRC(2,4)=SRC(3,5)=SRC(4,6)=SRC(5,7)= (l2 + 2*l1 + l0 + 2) >> 2; - SRC(0,1)=SRC(1,2)=SRC(2,3)=SRC(3,4)=SRC(4,5)=SRC(5,6)=SRC(6,7)= (l1 + 2*l0 + lt + 2) >> 2; - SRC(0,0)=SRC(1,1)=SRC(2,2)=SRC(3,3)=SRC(4,4)=SRC(5,5)=SRC(6,6)=SRC(7,7)= (l0 + 2*lt + t0 + 2) >> 2; - SRC(1,0)=SRC(2,1)=SRC(3,2)=SRC(4,3)=SRC(5,4)=SRC(6,5)=SRC(7,6)= (lt + 2*t0 + t1 + 2) >> 2; - SRC(2,0)=SRC(3,1)=SRC(4,2)=SRC(5,3)=SRC(6,4)=SRC(7,5)= (t0 + 2*t1 + t2 + 2) >> 2; - SRC(3,0)=SRC(4,1)=SRC(5,2)=SRC(6,3)=SRC(7,4)= (t1 + 2*t2 + t3 + 2) >> 2; - SRC(4,0)=SRC(5,1)=SRC(6,2)=SRC(7,3)= (t2 + 2*t3 + t4 + 2) >> 2; - SRC(5,0)=SRC(6,1)=SRC(7,2)= (t3 + 2*t4 + t5 + 2) >> 2; - SRC(6,0)=SRC(7,1)= (t4 + 2*t5 + t6 + 2) >> 2; - SRC(7,0)= (t5 + 2*t6 + t7 + 2) >> 2; - -} -static void pred8x8l_vertical_right_c(uint8_t *src, int has_topleft, int has_topright, int stride) -{ - PREDICT_8x8_LOAD_TOP; - PREDICT_8x8_LOAD_LEFT; - PREDICT_8x8_LOAD_TOPLEFT; - SRC(0,6)= (l5 + 2*l4 + l3 + 2) >> 2; - SRC(0,7)= (l6 + 2*l5 + l4 + 2) >> 2; - SRC(0,4)=SRC(1,6)= (l3 + 2*l2 + l1 + 2) >> 2; - SRC(0,5)=SRC(1,7)= (l4 + 2*l3 + l2 + 2) >> 2; - SRC(0,2)=SRC(1,4)=SRC(2,6)= (l1 + 2*l0 + lt + 2) >> 2; - SRC(0,3)=SRC(1,5)=SRC(2,7)= (l2 + 2*l1 + l0 + 2) >> 2; - SRC(0,1)=SRC(1,3)=SRC(2,5)=SRC(3,7)= (l0 + 2*lt + t0 + 2) >> 2; - SRC(0,0)=SRC(1,2)=SRC(2,4)=SRC(3,6)= (lt + t0 + 1) >> 1; - SRC(1,1)=SRC(2,3)=SRC(3,5)=SRC(4,7)= (lt + 2*t0 + t1 + 2) >> 2; - SRC(1,0)=SRC(2,2)=SRC(3,4)=SRC(4,6)= (t0 + t1 + 1) >> 1; - SRC(2,1)=SRC(3,3)=SRC(4,5)=SRC(5,7)= (t0 + 2*t1 + t2 + 2) >> 2; - SRC(2,0)=SRC(3,2)=SRC(4,4)=SRC(5,6)= (t1 + t2 + 1) >> 1; - SRC(3,1)=SRC(4,3)=SRC(5,5)=SRC(6,7)= (t1 + 2*t2 + t3 + 2) >> 2; - SRC(3,0)=SRC(4,2)=SRC(5,4)=SRC(6,6)= (t2 + t3 + 1) >> 1; - SRC(4,1)=SRC(5,3)=SRC(6,5)=SRC(7,7)= (t2 + 2*t3 + t4 + 2) >> 2; - SRC(4,0)=SRC(5,2)=SRC(6,4)=SRC(7,6)= (t3 + t4 + 1) >> 1; - SRC(5,1)=SRC(6,3)=SRC(7,5)= (t3 + 2*t4 + t5 + 2) >> 2; - SRC(5,0)=SRC(6,2)=SRC(7,4)= (t4 + t5 + 1) >> 1; - SRC(6,1)=SRC(7,3)= (t4 + 2*t5 + t6 + 2) >> 2; - SRC(6,0)=SRC(7,2)= (t5 + t6 + 1) >> 1; - SRC(7,1)= (t5 + 2*t6 + t7 + 2) >> 2; - SRC(7,0)= (t6 + t7 + 1) >> 1; -} -static void pred8x8l_horizontal_down_c(uint8_t *src, int has_topleft, int has_topright, int stride) -{ - PREDICT_8x8_LOAD_TOP; - PREDICT_8x8_LOAD_LEFT; - PREDICT_8x8_LOAD_TOPLEFT; - SRC(0,7)= (l6 + l7 + 1) >> 1; - SRC(1,7)= (l5 + 2*l6 + l7 + 2) >> 2; - SRC(0,6)=SRC(2,7)= (l5 + l6 + 1) >> 1; - SRC(1,6)=SRC(3,7)= (l4 + 2*l5 + l6 + 2) >> 2; - SRC(0,5)=SRC(2,6)=SRC(4,7)= (l4 + l5 + 1) >> 1; - SRC(1,5)=SRC(3,6)=SRC(5,7)= (l3 + 2*l4 + l5 + 2) >> 2; - SRC(0,4)=SRC(2,5)=SRC(4,6)=SRC(6,7)= (l3 + l4 + 1) >> 1; - SRC(1,4)=SRC(3,5)=SRC(5,6)=SRC(7,7)= (l2 + 2*l3 + l4 + 2) >> 2; - SRC(0,3)=SRC(2,4)=SRC(4,5)=SRC(6,6)= (l2 + l3 + 1) >> 1; - SRC(1,3)=SRC(3,4)=SRC(5,5)=SRC(7,6)= (l1 + 2*l2 + l3 + 2) >> 2; - SRC(0,2)=SRC(2,3)=SRC(4,4)=SRC(6,5)= (l1 + l2 + 1) >> 1; - SRC(1,2)=SRC(3,3)=SRC(5,4)=SRC(7,5)= (l0 + 2*l1 + l2 + 2) >> 2; - SRC(0,1)=SRC(2,2)=SRC(4,3)=SRC(6,4)= (l0 + l1 + 1) >> 1; - SRC(1,1)=SRC(3,2)=SRC(5,3)=SRC(7,4)= (lt + 2*l0 + l1 + 2) >> 2; - SRC(0,0)=SRC(2,1)=SRC(4,2)=SRC(6,3)= (lt + l0 + 1) >> 1; - SRC(1,0)=SRC(3,1)=SRC(5,2)=SRC(7,3)= (l0 + 2*lt + t0 + 2) >> 2; - SRC(2,0)=SRC(4,1)=SRC(6,2)= (t1 + 2*t0 + lt + 2) >> 2; - SRC(3,0)=SRC(5,1)=SRC(7,2)= (t2 + 2*t1 + t0 + 2) >> 2; - SRC(4,0)=SRC(6,1)= (t3 + 2*t2 + t1 + 2) >> 2; - SRC(5,0)=SRC(7,1)= (t4 + 2*t3 + t2 + 2) >> 2; - SRC(6,0)= (t5 + 2*t4 + t3 + 2) >> 2; - SRC(7,0)= (t6 + 2*t5 + t4 + 2) >> 2; -} -static void pred8x8l_vertical_left_c(uint8_t *src, int has_topleft, int has_topright, int stride) -{ - PREDICT_8x8_LOAD_TOP; - PREDICT_8x8_LOAD_TOPRIGHT; - SRC(0,0)= (t0 + t1 + 1) >> 1; - SRC(0,1)= (t0 + 2*t1 + t2 + 2) >> 2; - SRC(0,2)=SRC(1,0)= (t1 + t2 + 1) >> 1; - SRC(0,3)=SRC(1,1)= (t1 + 2*t2 + t3 + 2) >> 2; - SRC(0,4)=SRC(1,2)=SRC(2,0)= (t2 + t3 + 1) >> 1; - SRC(0,5)=SRC(1,3)=SRC(2,1)= (t2 + 2*t3 + t4 + 2) >> 2; - SRC(0,6)=SRC(1,4)=SRC(2,2)=SRC(3,0)= (t3 + t4 + 1) >> 1; - SRC(0,7)=SRC(1,5)=SRC(2,3)=SRC(3,1)= (t3 + 2*t4 + t5 + 2) >> 2; - SRC(1,6)=SRC(2,4)=SRC(3,2)=SRC(4,0)= (t4 + t5 + 1) >> 1; - SRC(1,7)=SRC(2,5)=SRC(3,3)=SRC(4,1)= (t4 + 2*t5 + t6 + 2) >> 2; - SRC(2,6)=SRC(3,4)=SRC(4,2)=SRC(5,0)= (t5 + t6 + 1) >> 1; - SRC(2,7)=SRC(3,5)=SRC(4,3)=SRC(5,1)= (t5 + 2*t6 + t7 + 2) >> 2; - SRC(3,6)=SRC(4,4)=SRC(5,2)=SRC(6,0)= (t6 + t7 + 1) >> 1; - SRC(3,7)=SRC(4,5)=SRC(5,3)=SRC(6,1)= (t6 + 2*t7 + t8 + 2) >> 2; - SRC(4,6)=SRC(5,4)=SRC(6,2)=SRC(7,0)= (t7 + t8 + 1) >> 1; - SRC(4,7)=SRC(5,5)=SRC(6,3)=SRC(7,1)= (t7 + 2*t8 + t9 + 2) >> 2; - SRC(5,6)=SRC(6,4)=SRC(7,2)= (t8 + t9 + 1) >> 1; - SRC(5,7)=SRC(6,5)=SRC(7,3)= (t8 + 2*t9 + t10 + 2) >> 2; - SRC(6,6)=SRC(7,4)= (t9 + t10 + 1) >> 1; - SRC(6,7)=SRC(7,5)= (t9 + 2*t10 + t11 + 2) >> 2; - SRC(7,6)= (t10 + t11 + 1) >> 1; - SRC(7,7)= (t10 + 2*t11 + t12 + 2) >> 2; -} -static void pred8x8l_horizontal_up_c(uint8_t *src, int has_topleft, int has_topright, int stride) -{ - PREDICT_8x8_LOAD_LEFT; - SRC(0,0)= (l0 + l1 + 1) >> 1; - SRC(1,0)= (l0 + 2*l1 + l2 + 2) >> 2; - SRC(0,1)=SRC(2,0)= (l1 + l2 + 1) >> 1; - SRC(1,1)=SRC(3,0)= (l1 + 2*l2 + l3 + 2) >> 2; - SRC(0,2)=SRC(2,1)=SRC(4,0)= (l2 + l3 + 1) >> 1; - SRC(1,2)=SRC(3,1)=SRC(5,0)= (l2 + 2*l3 + l4 + 2) >> 2; - SRC(0,3)=SRC(2,2)=SRC(4,1)=SRC(6,0)= (l3 + l4 + 1) >> 1; - SRC(1,3)=SRC(3,2)=SRC(5,1)=SRC(7,0)= (l3 + 2*l4 + l5 + 2) >> 2; - SRC(0,4)=SRC(2,3)=SRC(4,2)=SRC(6,1)= (l4 + l5 + 1) >> 1; - SRC(1,4)=SRC(3,3)=SRC(5,2)=SRC(7,1)= (l4 + 2*l5 + l6 + 2) >> 2; - SRC(0,5)=SRC(2,4)=SRC(4,3)=SRC(6,2)= (l5 + l6 + 1) >> 1; - SRC(1,5)=SRC(3,4)=SRC(5,3)=SRC(7,2)= (l5 + 2*l6 + l7 + 2) >> 2; - SRC(0,6)=SRC(2,5)=SRC(4,4)=SRC(6,3)= (l6 + l7 + 1) >> 1; - SRC(1,6)=SRC(3,5)=SRC(5,4)=SRC(7,3)= (l6 + 3*l7 + 2) >> 2; - SRC(0,7)=SRC(1,7)=SRC(2,6)=SRC(2,7)=SRC(3,6)= - SRC(3,7)=SRC(4,5)=SRC(4,6)=SRC(4,7)=SRC(5,5)= - SRC(5,6)=SRC(5,7)=SRC(6,4)=SRC(6,5)=SRC(6,6)= - SRC(6,7)=SRC(7,4)=SRC(7,5)=SRC(7,6)=SRC(7,7)= l7; -} -#undef PREDICT_8x8_LOAD_LEFT -#undef PREDICT_8x8_LOAD_TOP -#undef PREDICT_8x8_LOAD_TOPLEFT -#undef PREDICT_8x8_LOAD_TOPRIGHT -#undef PREDICT_8x8_DC -#undef PTR -#undef PT -#undef PL -#undef SRC - static inline void mc_dir_part(H264Context *h, Picture *pic, int n, int square, int chroma_height, int delta, int list, uint8_t *dest_y, uint8_t *dest_cb, uint8_t *dest_cr, int src_x_offset, int src_y_offset, @@ -2716,55 +2003,6 @@ } } -/** - * Sets the intra prediction function pointers. - */ -static void init_pred_ptrs(H264Context *h){ -// MpegEncContext * const s = &h->s; - - h->pred4x4[VERT_PRED ]= pred4x4_vertical_c; - h->pred4x4[HOR_PRED ]= pred4x4_horizontal_c; - h->pred4x4[DC_PRED ]= pred4x4_dc_c; - h->pred4x4[DIAG_DOWN_LEFT_PRED ]= pred4x4_down_left_c; - h->pred4x4[DIAG_DOWN_RIGHT_PRED]= pred4x4_down_right_c; - h->pred4x4[VERT_RIGHT_PRED ]= pred4x4_vertical_right_c; - h->pred4x4[HOR_DOWN_PRED ]= pred4x4_horizontal_down_c; - h->pred4x4[VERT_LEFT_PRED ]= pred4x4_vertical_left_c; - h->pred4x4[HOR_UP_PRED ]= pred4x4_horizontal_up_c; - h->pred4x4[LEFT_DC_PRED ]= pred4x4_left_dc_c; - h->pred4x4[TOP_DC_PRED ]= pred4x4_top_dc_c; - h->pred4x4[DC_128_PRED ]= pred4x4_128_dc_c; - - h->pred8x8l[VERT_PRED ]= pred8x8l_vertical_c; - h->pred8x8l[HOR_PRED ]= pred8x8l_horizontal_c; - h->pred8x8l[DC_PRED ]= pred8x8l_dc_c; - h->pred8x8l[DIAG_DOWN_LEFT_PRED ]= pred8x8l_down_left_c; - h->pred8x8l[DIAG_DOWN_RIGHT_PRED]= pred8x8l_down_right_c; - h->pred8x8l[VERT_RIGHT_PRED ]= pred8x8l_vertical_right_c; - h->pred8x8l[HOR_DOWN_PRED ]= pred8x8l_horizontal_down_c; - h->pred8x8l[VERT_LEFT_PRED ]= pred8x8l_vertical_left_c; - h->pred8x8l[HOR_UP_PRED ]= pred8x8l_horizontal_up_c; - h->pred8x8l[LEFT_DC_PRED ]= pred8x8l_left_dc_c; - h->pred8x8l[TOP_DC_PRED ]= pred8x8l_top_dc_c; - h->pred8x8l[DC_128_PRED ]= pred8x8l_128_dc_c; - - h->pred8x8[DC_PRED8x8 ]= ff_pred8x8_dc_c; - h->pred8x8[VERT_PRED8x8 ]= ff_pred8x8_vertical_c; - h->pred8x8[HOR_PRED8x8 ]= ff_pred8x8_horizontal_c; - h->pred8x8[PLANE_PRED8x8 ]= ff_pred8x8_plane_c; - h->pred8x8[LEFT_DC_PRED8x8]= ff_pred8x8_left_dc_c; - h->pred8x8[TOP_DC_PRED8x8 ]= ff_pred8x8_top_dc_c; - h->pred8x8[DC_128_PRED8x8 ]= ff_pred8x8_128_dc_c; - - h->pred16x16[DC_PRED8x8 ]= ff_pred16x16_dc_c; - h->pred16x16[VERT_PRED8x8 ]= ff_pred16x16_vertical_c; - h->pred16x16[HOR_PRED8x8 ]= ff_pred16x16_horizontal_c; - h->pred16x16[PLANE_PRED8x8 ]= ff_pred16x16_plane_c; - h->pred16x16[LEFT_DC_PRED8x8]= ff_pred16x16_left_dc_c; - h->pred16x16[TOP_DC_PRED8x8 ]= ff_pred16x16_top_dc_c; - h->pred16x16[DC_128_PRED8x8 ]= ff_pred16x16_128_dc_c; -} - static void free_tables(H264Context *h){ int i; av_freep(&h->intra4x4_pred_mode); @@ -2914,7 +2152,7 @@ s->height = s->avctx->height; s->codec_id= s->avctx->codec->id; - init_pred_ptrs(h); + ff_h264_pred_init(&h->hpc, s->codec_id); h->dequant_coeff_pps= -1; s->unrestricted_mv=1; @@ -3270,8 +2508,8 @@ xchg_mb_border(h, dest_y, dest_cb, dest_cr, linesize, uvlinesize, 1, simple); if(simple || !ENABLE_GRAY || !(s->flags&CODEC_FLAG_GRAY)){ - h->pred8x8[ h->chroma_pred_mode ](dest_cb, uvlinesize); - h->pred8x8[ h->chroma_pred_mode ](dest_cr, uvlinesize); + h->hpc.pred8x8[ h->chroma_pred_mode ](dest_cb, uvlinesize); + h->hpc.pred8x8[ h->chroma_pred_mode ](dest_cr, uvlinesize); } if(IS_INTRA4x4(mb_type)){ @@ -3281,7 +2519,7 @@ uint8_t * const ptr= dest_y + block_offset[i]; const int dir= h->intra4x4_pred_mode_cache[ scan8[i] ]; const int nnz = h->non_zero_count_cache[ scan8[i] ]; - h->pred8x8l[ dir ](ptr, (h->topleft_samples_available<<i)&0x8000, + h->hpc.pred8x8l[ dir ](ptr, (h->topleft_samples_available<<i)&0x8000, (h->topright_samples_available<<i)&0x4000, linesize); if(nnz){ if(nnz == 1 && h->mb[i*16]) @@ -3308,7 +2546,7 @@ }else topright= NULL; - h->pred4x4[ dir ](ptr, topright, linesize); + h->hpc.pred4x4[ dir ](ptr, topright, linesize); nnz = h->non_zero_count_cache[ scan8[i] ]; if(nnz){ if(is_h264){ @@ -3322,7 +2560,7 @@ } } }else{ - h->pred16x16[ h->intra16x16_pred_mode ](dest_y , linesize); + h->hpc.pred16x16[ h->intra16x16_pred_mode ](dest_y , linesize); if(is_h264){ if(!transform_bypass) h264_luma_dc_dequant_idct_c(h->mb, s->qscale, h->dequant4_coeff[IS_INTRA(mb_type) ? 0:3][s->qscale][0]);