dsputil.c
上传用户:shlianrong
上传日期:2022-07-08
资源大小:309k
文件大小:174k
- }
- }
- static void OPNAME ## h264_qpel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){
- const int w=8;
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
- int i;
- for(i=0; i<w; i++)
- {
- const int srcB= src[-2*srcStride];
- const int srcA= src[-1*srcStride];
- const int src0= src[0 *srcStride];
- const int src1= src[1 *srcStride];
- const int src2= src[2 *srcStride];
- const int src3= src[3 *srcStride];
- const int src4= src[4 *srcStride];
- const int src5= src[5 *srcStride];
- const int src6= src[6 *srcStride];
- const int src7= src[7 *srcStride];
- const int src8= src[8 *srcStride];
- const int src9= src[9 *srcStride];
- const int src10=src[10*srcStride];
- OP(dst[0*dstStride], (src0+src1)*20 - (srcA+src2)*5 + (srcB+src3));
- OP(dst[1*dstStride], (src1+src2)*20 - (src0+src3)*5 + (srcA+src4));
- OP(dst[2*dstStride], (src2+src3)*20 - (src1+src4)*5 + (src0+src5));
- OP(dst[3*dstStride], (src3+src4)*20 - (src2+src5)*5 + (src1+src6));
- OP(dst[4*dstStride], (src4+src5)*20 - (src3+src6)*5 + (src2+src7));
- OP(dst[5*dstStride], (src5+src6)*20 - (src4+src7)*5 + (src3+src8));
- OP(dst[6*dstStride], (src6+src7)*20 - (src5+src8)*5 + (src4+src9));
- OP(dst[7*dstStride], (src7+src8)*20 - (src6+src9)*5 + (src5+src10));
- dst++;
- src++;
- }
- }
- static void OPNAME ## h264_qpel8_hv_lowpass(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){
- const int h=8;
- const int w=8;
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
- int i;
- src -= 2*srcStride;
- for(i=0; i<h+5; i++)
- {
- tmp[0]= (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3 ]);
- tmp[1]= (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4 ]);
- tmp[2]= (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5 ]);
- tmp[3]= (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6 ]);
- tmp[4]= (src[4]+src[5])*20 - (src[3 ]+src[6])*5 + (src[2 ]+src[7 ]);
- tmp[5]= (src[5]+src[6])*20 - (src[4 ]+src[7])*5 + (src[3 ]+src[8 ]);
- tmp[6]= (src[6]+src[7])*20 - (src[5 ]+src[8])*5 + (src[4 ]+src[9 ]);
- tmp[7]= (src[7]+src[8])*20 - (src[6 ]+src[9])*5 + (src[5 ]+src[10]);
- tmp+=tmpStride;
- src+=srcStride;
- }
- tmp -= tmpStride*(h+5-2);
- for(i=0; i<w; i++)
- {
- const int tmpB= tmp[-2*tmpStride];
- const int tmpA= tmp[-1*tmpStride];
- const int tmp0= tmp[0 *tmpStride];
- const int tmp1= tmp[1 *tmpStride];
- const int tmp2= tmp[2 *tmpStride];
- const int tmp3= tmp[3 *tmpStride];
- const int tmp4= tmp[4 *tmpStride];
- const int tmp5= tmp[5 *tmpStride];
- const int tmp6= tmp[6 *tmpStride];
- const int tmp7= tmp[7 *tmpStride];
- const int tmp8= tmp[8 *tmpStride];
- const int tmp9= tmp[9 *tmpStride];
- const int tmp10=tmp[10*tmpStride];
- OP2(dst[0*dstStride], (tmp0+tmp1)*20 - (tmpA+tmp2)*5 + (tmpB+tmp3));
- OP2(dst[1*dstStride], (tmp1+tmp2)*20 - (tmp0+tmp3)*5 + (tmpA+tmp4));
- OP2(dst[2*dstStride], (tmp2+tmp3)*20 - (tmp1+tmp4)*5 + (tmp0+tmp5));
- OP2(dst[3*dstStride], (tmp3+tmp4)*20 - (tmp2+tmp5)*5 + (tmp1+tmp6));
- OP2(dst[4*dstStride], (tmp4+tmp5)*20 - (tmp3+tmp6)*5 + (tmp2+tmp7));
- OP2(dst[5*dstStride], (tmp5+tmp6)*20 - (tmp4+tmp7)*5 + (tmp3+tmp8));
- OP2(dst[6*dstStride], (tmp6+tmp7)*20 - (tmp5+tmp8)*5 + (tmp4+tmp9));
- OP2(dst[7*dstStride], (tmp7+tmp8)*20 - (tmp6+tmp9)*5 + (tmp5+tmp10));
- dst++;
- tmp++;
- }
- }
- static void OPNAME ## h264_qpel16_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){
- OPNAME ## h264_qpel8_v_lowpass(dst , src , dstStride, srcStride);
- OPNAME ## h264_qpel8_v_lowpass(dst+8, src+8, dstStride, srcStride);
- src += 8*srcStride;
- dst += 8*dstStride;
- OPNAME ## h264_qpel8_v_lowpass(dst , src , dstStride, srcStride);
- OPNAME ## h264_qpel8_v_lowpass(dst+8, src+8, dstStride, srcStride);
- }
- static void OPNAME ## h264_qpel16_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){
- OPNAME ## h264_qpel8_h_lowpass(dst , src , dstStride, srcStride);
- OPNAME ## h264_qpel8_h_lowpass(dst+8, src+8, dstStride, srcStride);
- src += 8*srcStride;
- dst += 8*dstStride;
- OPNAME ## h264_qpel8_h_lowpass(dst , src , dstStride, srcStride);
- OPNAME ## h264_qpel8_h_lowpass(dst+8, src+8, dstStride, srcStride);
- }
- static void OPNAME ## h264_qpel16_hv_lowpass(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){
- OPNAME ## h264_qpel8_hv_lowpass(dst , tmp , src , dstStride, tmpStride, srcStride);
- OPNAME ## h264_qpel8_hv_lowpass(dst+8, tmp+8, src+8, dstStride, tmpStride, srcStride);
- src += 8*srcStride;
- dst += 8*dstStride;
- OPNAME ## h264_qpel8_hv_lowpass(dst , tmp , src , dstStride, tmpStride, srcStride);
- OPNAME ## h264_qpel8_hv_lowpass(dst+8, tmp+8, src+8, dstStride, tmpStride, srcStride);
- }
- #define H264_MC(OPNAME, SIZE)
- static void OPNAME ## h264_qpel ## SIZE ## _mc00_c (uint8_t *dst, uint8_t *src, int stride){
- OPNAME ## pixels ## SIZE ## _c(dst, src, stride, SIZE);
- }
- static void OPNAME ## h264_qpel ## SIZE ## _mc10_c(uint8_t *dst, uint8_t *src, int stride){
- uint8_t half[SIZE*SIZE];
- put_h264_qpel ## SIZE ## _h_lowpass(half, src, SIZE, stride);
- OPNAME ## pixels ## SIZE ## _l2(dst, src, half, stride, stride, SIZE, SIZE);
- }
- static void OPNAME ## h264_qpel ## SIZE ## _mc20_c(uint8_t *dst, uint8_t *src, int stride){
- OPNAME ## h264_qpel ## SIZE ## _h_lowpass(dst, src, stride, stride);
- }
- static void OPNAME ## h264_qpel ## SIZE ## _mc30_c(uint8_t *dst, uint8_t *src, int stride){
- uint8_t half[SIZE*SIZE];
- put_h264_qpel ## SIZE ## _h_lowpass(half, src, SIZE, stride);
- OPNAME ## pixels ## SIZE ## _l2(dst, src+1, half, stride, stride, SIZE, SIZE);
- }
- static void OPNAME ## h264_qpel ## SIZE ## _mc01_c(uint8_t *dst, uint8_t *src, int stride){
- uint8_t full[SIZE*(SIZE+5)];
- uint8_t * const full_mid= full + SIZE*2;
- uint8_t half[SIZE*SIZE];
- copy_block ## SIZE (full, src - stride*2, SIZE, stride, SIZE + 5);
- put_h264_qpel ## SIZE ## _v_lowpass(half, full_mid, SIZE, SIZE);
- OPNAME ## pixels ## SIZE ## _l2(dst, full_mid, half, stride, SIZE, SIZE, SIZE);
- }
- static void OPNAME ## h264_qpel ## SIZE ## _mc02_c(uint8_t *dst, uint8_t *src, int stride){
- uint8_t full[SIZE*(SIZE+5)];
- uint8_t * const full_mid= full + SIZE*2;
- copy_block ## SIZE (full, src - stride*2, SIZE, stride, SIZE + 5);
- OPNAME ## h264_qpel ## SIZE ## _v_lowpass(dst, full_mid, stride, SIZE);
- }
- static void OPNAME ## h264_qpel ## SIZE ## _mc03_c(uint8_t *dst, uint8_t *src, int stride){
- uint8_t full[SIZE*(SIZE+5)];
- uint8_t * const full_mid= full + SIZE*2;
- uint8_t half[SIZE*SIZE];
- copy_block ## SIZE (full, src - stride*2, SIZE, stride, SIZE + 5);
- put_h264_qpel ## SIZE ## _v_lowpass(half, full_mid, SIZE, SIZE);
- OPNAME ## pixels ## SIZE ## _l2(dst, full_mid+SIZE, half, stride, SIZE, SIZE, SIZE);
- }
- static void OPNAME ## h264_qpel ## SIZE ## _mc11_c(uint8_t *dst, uint8_t *src, int stride){
- uint8_t full[SIZE*(SIZE+5)];
- uint8_t * const full_mid= full + SIZE*2;
- uint8_t halfH[SIZE*SIZE];
- uint8_t halfV[SIZE*SIZE];
- put_h264_qpel ## SIZE ## _h_lowpass(halfH, src, SIZE, stride);
- copy_block ## SIZE (full, src - stride*2, SIZE, stride, SIZE + 5);
- put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);
- OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfV, stride, SIZE, SIZE, SIZE);
- }
- static void OPNAME ## h264_qpel ## SIZE ## _mc31_c(uint8_t *dst, uint8_t *src, int stride){
- uint8_t full[SIZE*(SIZE+5)];
- uint8_t * const full_mid= full + SIZE*2;
- uint8_t halfH[SIZE*SIZE];
- uint8_t halfV[SIZE*SIZE];
- put_h264_qpel ## SIZE ## _h_lowpass(halfH, src, SIZE, stride);
- copy_block ## SIZE (full, src - stride*2 + 1, SIZE, stride, SIZE + 5);
- put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);
- OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfV, stride, SIZE, SIZE, SIZE);
- }
- static void OPNAME ## h264_qpel ## SIZE ## _mc13_c(uint8_t *dst, uint8_t *src, int stride){
- uint8_t full[SIZE*(SIZE+5)];
- uint8_t * const full_mid= full + SIZE*2;
- uint8_t halfH[SIZE*SIZE];
- uint8_t halfV[SIZE*SIZE];
- put_h264_qpel ## SIZE ## _h_lowpass(halfH, src + stride, SIZE, stride);
- copy_block ## SIZE (full, src - stride*2, SIZE, stride, SIZE + 5);
- put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);
- OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfV, stride, SIZE, SIZE, SIZE);
- }
- static void OPNAME ## h264_qpel ## SIZE ## _mc33_c(uint8_t *dst, uint8_t *src, int stride){
- uint8_t full[SIZE*(SIZE+5)];
- uint8_t * const full_mid= full + SIZE*2;
- uint8_t halfH[SIZE*SIZE];
- uint8_t halfV[SIZE*SIZE];
- put_h264_qpel ## SIZE ## _h_lowpass(halfH, src + stride, SIZE, stride);
- copy_block ## SIZE (full, src - stride*2 + 1, SIZE, stride, SIZE + 5);
- put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);
- OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfV, stride, SIZE, SIZE, SIZE);
- }
- static void OPNAME ## h264_qpel ## SIZE ## _mc22_c(uint8_t *dst, uint8_t *src, int stride){
- int16_t tmp[SIZE*(SIZE+5)];
- OPNAME ## h264_qpel ## SIZE ## _hv_lowpass(dst, tmp, src, stride, SIZE, stride);
- }
- static void OPNAME ## h264_qpel ## SIZE ## _mc21_c(uint8_t *dst, uint8_t *src, int stride){
- int16_t tmp[SIZE*(SIZE+5)];
- uint8_t halfH[SIZE*SIZE];
- uint8_t halfHV[SIZE*SIZE];
- put_h264_qpel ## SIZE ## _h_lowpass(halfH, src, SIZE, stride);
- put_h264_qpel ## SIZE ## _hv_lowpass(halfHV, tmp, src, SIZE, SIZE, stride);
- OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfHV, stride, SIZE, SIZE, SIZE);
- }
- static void OPNAME ## h264_qpel ## SIZE ## _mc23_c(uint8_t *dst, uint8_t *src, int stride){
- int16_t tmp[SIZE*(SIZE+5)];
- uint8_t halfH[SIZE*SIZE];
- uint8_t halfHV[SIZE*SIZE];
- put_h264_qpel ## SIZE ## _h_lowpass(halfH, src + stride, SIZE, stride);
- put_h264_qpel ## SIZE ## _hv_lowpass(halfHV, tmp, src, SIZE, SIZE, stride);
- OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfHV, stride, SIZE, SIZE, SIZE);
- }
- static void OPNAME ## h264_qpel ## SIZE ## _mc12_c(uint8_t *dst, uint8_t *src, int stride){
- uint8_t full[SIZE*(SIZE+5)];
- uint8_t * const full_mid= full + SIZE*2;
- int16_t tmp[SIZE*(SIZE+5)];
- uint8_t halfV[SIZE*SIZE];
- uint8_t halfHV[SIZE*SIZE];
- copy_block ## SIZE (full, src - stride*2, SIZE, stride, SIZE + 5);
- put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);
- put_h264_qpel ## SIZE ## _hv_lowpass(halfHV, tmp, src, SIZE, SIZE, stride);
- OPNAME ## pixels ## SIZE ## _l2(dst, halfV, halfHV, stride, SIZE, SIZE, SIZE);
- }
- static void OPNAME ## h264_qpel ## SIZE ## _mc32_c(uint8_t *dst, uint8_t *src, int stride){
- uint8_t full[SIZE*(SIZE+5)];
- uint8_t * const full_mid= full + SIZE*2;
- int16_t tmp[SIZE*(SIZE+5)];
- uint8_t halfV[SIZE*SIZE];
- uint8_t halfHV[SIZE*SIZE];
- copy_block ## SIZE (full, src - stride*2 + 1, SIZE, stride, SIZE + 5);
- put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);
- put_h264_qpel ## SIZE ## _hv_lowpass(halfHV, tmp, src, SIZE, SIZE, stride);
- OPNAME ## pixels ## SIZE ## _l2(dst, halfV, halfHV, stride, SIZE, SIZE, SIZE);
- }
- #define op_avg(a, b) a = (((a)+cm[((b) + 16)>>5]+1)>>1)
- #define op2_avg(a, b) a = (((a)+cm[((b) + 512)>>10]+1)>>1)
- #define op_put(a, b) a = cm[((b) + 16)>>5]
- #define op2_put(a, b) a = cm[((b) + 512)>>10]
- H264_LOWPASS(avg_ , op_avg, op2_avg)
- H264_LOWPASS(put_ , op_put, op2_put)
- static av_unused void put_h264_qpel2_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride)
- {
- const int h=2;
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
- int i;
- for(i=0; i<h; i++)
- {
- /*OP(dst[0], (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]));*/
- dst[0] = (((dst[0])+cm[(((src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3])) + 16)>>5]+1)>>1);
- /*OP(dst[1], (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]));*/
- dst[1] = (((dst[1])+cm[(((src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4])) + 16)>>5]+1)>>1);
- dst+=dstStride;
- src+=srcStride;
- }
- }
- static av_unused void put_h264_qpel2_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride)
- {
- const int w=2;
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
- int i;
- for(i=0; i<w; i++)
- {
- const int srcB= src[-2*srcStride];
- const int srcA= src[-1*srcStride];
- const int src0= src[0 *srcStride];
- const int src1= src[1 *srcStride];
- const int src2= src[2 *srcStride];
- const int src3= src[3 *srcStride];
- const int src4= src[4 *srcStride];
- /*OP(dst[0*dstStride], (src0+src1)*20 - (srcA+src2)*5 + (srcB+src3));*/
- dst[0*dstStride] = (((dst[0*dstStride])+cm[(((src0+src1)*20 - (srcA+src2)*5 + (srcB+src3)) + 16)>>5]+1)>>1);
- /*OP(dst[1*dstStride], (src1+src2)*20 - (src0+src3)*5 + (srcA+src4));*/
- dst[1*dstStride] = (((dst[1*dstStride])+cm[(((src1+src2)*20 - (src0+src3)*5 + (srcA+src4)) + 16)>>5]+1)>>1);
- dst++;
- src++;
- }
- }
- static av_unused void put_h264_qpel2_hv_lowpass(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride)
- {
- const int h=2;
- const int w=2;
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
- int i;
- src -= 2*srcStride;
- for(i=0; i<h+5; i++)
- {
- tmp[0]= (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]);
- tmp[1]= (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]);
- tmp+=tmpStride;
- src+=srcStride;
- }
- tmp -= tmpStride*(h+5-2);
- for(i=0; i<w; i++)
- {
- const int tmpB= tmp[-2*tmpStride];
- const int tmpA= tmp[-1*tmpStride];
- const int tmp0= tmp[0 *tmpStride];
- const int tmp1= tmp[1 *tmpStride];
- const int tmp2= tmp[2 *tmpStride];
- const int tmp3= tmp[3 *tmpStride];
- const int tmp4= tmp[4 *tmpStride];
- /*OP2(dst[0*dstStride], (tmp0+tmp1)*20 - (tmpA+tmp2)*5 + (tmpB+tmp3));*/
- dst[0*dstStride] = (((dst[0*dstStride])+cm[(((tmp0+tmp1)*20 - (tmpA+tmp2)*5 + (tmpB+tmp3)) + 512)>>10]+1)>>1);
- /*OP2(dst[1*dstStride], (tmp1+tmp2)*20 - (tmp0+tmp3)*5 + (tmpA+tmp4));*/
- dst[1*dstStride] = (((dst[1*dstStride])+cm[(((tmp1+tmp2)*20 - (tmp0+tmp3)*5 + (tmpA+tmp4)) + 512)>>10]+1)>>1);
- dst++;
- tmp++;
- }
- }
- H264_MC(put_, 2)
- H264_MC(put_, 4)
- H264_MC(put_, 8)
- H264_MC(put_, 16)
- H264_MC(avg_, 4)
- H264_MC(avg_, 8)
- H264_MC(avg_, 16)
- #undef op_avg
- #undef op_put
- #undef op2_avg
- #undef op2_put
- #endif
- #define op_scale1(x) block[x] = av_clip_uint8( (block[x]*weight + offset) >> log2_denom )
- #define op_scale2(x) dst[x] = av_clip_uint8( (src[x]*weights + dst[x]*weightd + offset) >> (log2_denom+1))
- #define H264_WEIGHT(W,H)
- static void weight_h264_pixels ## W ## x ## H ## _c(uint8_t *block, int stride, int log2_denom, int weight, int offset){
- int y;
- offset <<= log2_denom;
- if(log2_denom) offset += 1<<(log2_denom-1);
- for(y=0; y<H; y++, block += stride){
- op_scale1(0);
- op_scale1(1);
- if(W==2) continue;
- op_scale1(2);
- op_scale1(3);
- if(W==4) continue;
- op_scale1(4);
- op_scale1(5);
- op_scale1(6);
- op_scale1(7);
- if(W==8) continue;
- op_scale1(8);
- op_scale1(9);
- op_scale1(10);
- op_scale1(11);
- op_scale1(12);
- op_scale1(13);
- op_scale1(14);
- op_scale1(15);
- }
- }
- static void biweight_h264_pixels ## W ## x ## H ## _c(uint8_t *dst, uint8_t *src, int stride, int log2_denom, int weightd, int weights, int offset){
- int y;
- offset = ((offset + 1) | 1) << log2_denom;
- for(y=0; y<H; y++, dst += stride, src += stride){
- op_scale2(0);
- op_scale2(1);
- if(W==2) continue;
- op_scale2(2);
- op_scale2(3);
- if(W==4) continue;
- op_scale2(4);
- op_scale2(5);
- op_scale2(6);
- op_scale2(7);
- if(W==8) continue;
- op_scale2(8);
- op_scale2(9);
- op_scale2(10);
- op_scale2(11);
- op_scale2(12);
- op_scale2(13);
- op_scale2(14);
- op_scale2(15);
- }
- }
- H264_WEIGHT(16,16)
- H264_WEIGHT(16,8)
- H264_WEIGHT(8,16)
- H264_WEIGHT(8,8)
- H264_WEIGHT(8,4)
- H264_WEIGHT(4,8)
- H264_WEIGHT(4,4)
- H264_WEIGHT(4,2)
- H264_WEIGHT(2,4)
- H264_WEIGHT(2,2)
- #undef op_scale1
- #undef op_scale2
- #undef H264_WEIGHT
- static void wmv2_mspel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
- int i;
- for(i=0; i<h; i++){
- dst[0]= cm[(9*(src[0] + src[1]) - (src[-1] + src[2]) + 8)>>4];
- dst[1]= cm[(9*(src[1] + src[2]) - (src[ 0] + src[3]) + 8)>>4];
- dst[2]= cm[(9*(src[2] + src[3]) - (src[ 1] + src[4]) + 8)>>4];
- dst[3]= cm[(9*(src[3] + src[4]) - (src[ 2] + src[5]) + 8)>>4];
- dst[4]= cm[(9*(src[4] + src[5]) - (src[ 3] + src[6]) + 8)>>4];
- dst[5]= cm[(9*(src[5] + src[6]) - (src[ 4] + src[7]) + 8)>>4];
- dst[6]= cm[(9*(src[6] + src[7]) - (src[ 5] + src[8]) + 8)>>4];
- dst[7]= cm[(9*(src[7] + src[8]) - (src[ 6] + src[9]) + 8)>>4];
- dst+=dstStride;
- src+=srcStride;
- }
- }
- #ifdef CONFIG_CAVS_DECODER
- /* AVS specific */
- void ff_cavsdsp_init(DSPContext* c, AVCodecContext *avctx);
- void ff_put_cavs_qpel8_mc00_c(uint8_t *dst, uint8_t *src, int stride) {
- put_pixels8_c(dst, src, stride, 8);
- }
- void ff_avg_cavs_qpel8_mc00_c(uint8_t *dst, uint8_t *src, int stride) {
- avg_pixels8_c(dst, src, stride, 8);
- }
- void ff_put_cavs_qpel16_mc00_c(uint8_t *dst, uint8_t *src, int stride) {
- put_pixels16_c(dst, src, stride, 16);
- }
- void ff_avg_cavs_qpel16_mc00_c(uint8_t *dst, uint8_t *src, int stride) {
- avg_pixels16_c(dst, src, stride, 16);
- }
- #endif /* CONFIG_CAVS_DECODER */
- #if defined(CONFIG_VC1_DECODER) || defined(CONFIG_WMV3_DECODER)
- /* VC-1 specific */
- void ff_vc1dsp_init(DSPContext* c, AVCodecContext *avctx);
- void ff_put_vc1_mspel_mc00_c(uint8_t *dst, uint8_t *src, int stride, int rnd) {
- put_pixels8_c(dst, src, stride, 8);
- }
- #endif /* CONFIG_VC1_DECODER||CONFIG_WMV3_DECODER */
- void ff_intrax8dsp_init(DSPContext* c, AVCodecContext *avctx);
- /* H264 specific */
- void ff_h264dspenc_init(DSPContext* c, AVCodecContext *avctx);
- #if defined(CONFIG_RV30_DECODER)
- void ff_rv30dsp_init(DSPContext* c, AVCodecContext *avctx);
- #endif /* CONFIG_RV30_DECODER */
- #if defined(CONFIG_RV40_DECODER)
- static void put_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride){
- put_pixels16_xy2_c(dst, src, stride, 16);
- }
- static void avg_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride){
- avg_pixels16_xy2_c(dst, src, stride, 16);
- }
- static void put_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride){
- put_pixels8_xy2_c(dst, src, stride, 8);
- }
- static void avg_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride){
- avg_pixels8_xy2_c(dst, src, stride, 8);
- }
- void ff_rv40dsp_init(DSPContext* c, AVCodecContext *avctx);
- #endif /* CONFIG_RV40_DECODER */
- static void wmv2_mspel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int w){
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
- int i;
- for(i=0; i<w; i++){
- const int src_1= src[ -srcStride];
- const int src0 = src[0 ];
- const int src1 = src[ srcStride];
- const int src2 = src[2*srcStride];
- const int src3 = src[3*srcStride];
- const int src4 = src[4*srcStride];
- const int src5 = src[5*srcStride];
- const int src6 = src[6*srcStride];
- const int src7 = src[7*srcStride];
- const int src8 = src[8*srcStride];
- const int src9 = src[9*srcStride];
- dst[0*dstStride]= cm[(9*(src0 + src1) - (src_1 + src2) + 8)>>4];
- dst[1*dstStride]= cm[(9*(src1 + src2) - (src0 + src3) + 8)>>4];
- dst[2*dstStride]= cm[(9*(src2 + src3) - (src1 + src4) + 8)>>4];
- dst[3*dstStride]= cm[(9*(src3 + src4) - (src2 + src5) + 8)>>4];
- dst[4*dstStride]= cm[(9*(src4 + src5) - (src3 + src6) + 8)>>4];
- dst[5*dstStride]= cm[(9*(src5 + src6) - (src4 + src7) + 8)>>4];
- dst[6*dstStride]= cm[(9*(src6 + src7) - (src5 + src8) + 8)>>4];
- dst[7*dstStride]= cm[(9*(src7 + src8) - (src6 + src9) + 8)>>4];
- src++;
- dst++;
- }
- }
- static void put_mspel8_mc00_c (uint8_t *dst, uint8_t *src, int stride){
- put_pixels8_c(dst, src, stride, 8);
- }
- static void put_mspel8_mc10_c(uint8_t *dst, uint8_t *src, int stride){
- uint8_t half[64];
- wmv2_mspel8_h_lowpass(half, src, 8, stride, 8);
- put_pixels8_l2(dst, src, half, stride, stride, 8, 8);
- }
- static void put_mspel8_mc20_c(uint8_t *dst, uint8_t *src, int stride){
- wmv2_mspel8_h_lowpass(dst, src, stride, stride, 8);
- }
- static void put_mspel8_mc30_c(uint8_t *dst, uint8_t *src, int stride){
- uint8_t half[64];
- wmv2_mspel8_h_lowpass(half, src, 8, stride, 8);
- put_pixels8_l2(dst, src+1, half, stride, stride, 8, 8);
- }
- static void put_mspel8_mc02_c(uint8_t *dst, uint8_t *src, int stride){
- wmv2_mspel8_v_lowpass(dst, src, stride, stride, 8);
- }
- static void put_mspel8_mc12_c(uint8_t *dst, uint8_t *src, int stride){
- uint8_t halfH[88];
- uint8_t halfV[64];
- uint8_t halfHV[64];
- wmv2_mspel8_h_lowpass(halfH, src-stride, 8, stride, 11);
- wmv2_mspel8_v_lowpass(halfV, src, 8, stride, 8);
- wmv2_mspel8_v_lowpass(halfHV, halfH+8, 8, 8, 8);
- put_pixels8_l2(dst, halfV, halfHV, stride, 8, 8, 8);
- }
- static void put_mspel8_mc32_c(uint8_t *dst, uint8_t *src, int stride){
- uint8_t halfH[88];
- uint8_t halfV[64];
- uint8_t halfHV[64];
- wmv2_mspel8_h_lowpass(halfH, src-stride, 8, stride, 11);
- wmv2_mspel8_v_lowpass(halfV, src+1, 8, stride, 8);
- wmv2_mspel8_v_lowpass(halfHV, halfH+8, 8, 8, 8);
- put_pixels8_l2(dst, halfV, halfHV, stride, 8, 8, 8);
- }
- static void put_mspel8_mc22_c(uint8_t *dst, uint8_t *src, int stride){
- uint8_t halfH[88];
- wmv2_mspel8_h_lowpass(halfH, src-stride, 8, stride, 11);
- wmv2_mspel8_v_lowpass(dst, halfH+8, stride, 8, 8);
- }
- /*
- static void h263_v_loop_filter_c(uint8_t *src, int stride, int qscale){
- if(ENABLE_ANY_H263) {
- int x;
- const int strength= ff_h263_loop_filter_strength[qscale];
- for(x=0; x<8; x++){
- int d1, d2, ad1;
- int p0= src[x-2*stride];
- int p1= src[x-1*stride];
- int p2= src[x+0*stride];
- int p3= src[x+1*stride];
- int d = (p0 - p3 + 4*(p2 - p1)) / 8;
- if (d<-2*strength) d1= 0;
- else if(d<- strength) d1=-2*strength - d;
- else if(d< strength) d1= d;
- else if(d< 2*strength) d1= 2*strength - d;
- else d1= 0;
- p1 += d1;
- p2 -= d1;
- if(p1&256) p1= ~(p1>>31);
- if(p2&256) p2= ~(p2>>31);
- src[x-1*stride] = p1;
- src[x+0*stride] = p2;
- ad1= FFABS(d1)>>1;
- d2= av_clip((p0-p3)/4, -ad1, ad1);
- src[x-2*stride] = p0 - d2;
- src[x+ stride] = p3 + d2;
- }
- }
- }
- static void h263_h_loop_filter_c(uint8_t *src, int stride, int qscale){
- if(ENABLE_ANY_H263) {
- int y;
- const int strength= ff_h263_loop_filter_strength[qscale];
- for(y=0; y<8; y++){
- int d1, d2, ad1;
- int p0= src[y*stride-2];
- int p1= src[y*stride-1];
- int p2= src[y*stride+0];
- int p3= src[y*stride+1];
- int d = (p0 - p3 + 4*(p2 - p1)) / 8;
- if (d<-2*strength) d1= 0;
- else if(d<- strength) d1=-2*strength - d;
- else if(d< strength) d1= d;
- else if(d< 2*strength) d1= 2*strength - d;
- else d1= 0;
- p1 += d1;
- p2 -= d1;
- if(p1&256) p1= ~(p1>>31);
- if(p2&256) p2= ~(p2>>31);
- src[y*stride-1] = p1;
- src[y*stride+0] = p2;
- ad1= FFABS(d1)>>1;
- d2= av_clip((p0-p3)/4, -ad1, ad1);
- src[y*stride-2] = p0 - d2;
- src[y*stride+1] = p3 + d2;
- }
- }
- }
- */
- static void h261_loop_filter_c(uint8_t *src, int stride){
- int x,y,xy,yz;
- int temp[64];
- for(x=0; x<8; x++){
- temp[x ] = 4*src[x ];
- temp[x + 7*8] = 4*src[x + 7*stride];
- }
- for(y=1; y<7; y++){
- for(x=0; x<8; x++){
- xy = y * stride + x;
- yz = y * 8 + x;
- temp[yz] = src[xy - stride] + 2*src[xy] + src[xy + stride];
- }
- }
- for(y=0; y<8; y++){
- src[ y*stride] = (temp[ y*8] + 2)>>2;
- src[7+y*stride] = (temp[7+y*8] + 2)>>2;
- for(x=1; x<7; x++){
- xy = y * stride + x;
- yz = y * 8 + x;
- src[xy] = (temp[yz-1] + 2*temp[yz] + temp[yz+1] + 8)>>4;
- }
- }
- }
- static inline void h264_loop_filter_luma_c(uint8_t *pix, int xstride, int ystride, int alpha, int beta, int8_t *tc0)
- {
- int i, d;
- for( i = 0; i < 4; i++ ) {
- if( tc0[i] < 0 ) {
- pix += 4*ystride;
- continue;
- }
- for( d = 0; d < 4; d++ ) {
- const int p0 = pix[-1*xstride];
- const int p1 = pix[-2*xstride];
- const int p2 = pix[-3*xstride];
- const int q0 = pix[0];
- const int q1 = pix[1*xstride];
- const int q2 = pix[2*xstride];
- if( FFABS( p0 - q0 ) < alpha &&
- FFABS( p1 - p0 ) < beta &&
- FFABS( q1 - q0 ) < beta ) {
- int tc = tc0[i];
- int i_delta;
- if( FFABS( p2 - p0 ) < beta ) {
- pix[-2*xstride] = p1 + av_clip( (( p2 + ( ( p0 + q0 + 1 ) >> 1 ) ) >> 1) - p1, -tc0[i], tc0[i] );
- tc++;
- }
- if( FFABS( q2 - q0 ) < beta ) {
- pix[ xstride] = q1 + av_clip( (( q2 + ( ( p0 + q0 + 1 ) >> 1 ) ) >> 1) - q1, -tc0[i], tc0[i] );
- tc++;
- }
- i_delta = av_clip( (((q0 - p0 ) << 2) + (p1 - q1) + 4) >> 3, -tc, tc );
- pix[-xstride] = av_clip_uint8( p0 + i_delta ); /* p0' */
- pix[0] = av_clip_uint8( q0 - i_delta ); /* q0' */
- }
- pix += ystride;
- }
- }
- }
- static void h264_v_loop_filter_luma_c(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
- {
- h264_loop_filter_luma_c(pix, stride, 1, alpha, beta, tc0);
- }
- static void h264_h_loop_filter_luma_c(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
- {
- h264_loop_filter_luma_c(pix, 1, stride, alpha, beta, tc0);
- }
- static inline void h264_loop_filter_luma_intra_c(uint8_t *pix, int xstride, int ystride, int alpha, int beta)
- {
- int d;
- for( d = 0; d < 16; d++ ) {
- const int p2 = pix[-3*xstride];
- const int p1 = pix[-2*xstride];
- const int p0 = pix[-1*xstride];
- const int q0 = pix[ 0*xstride];
- const int q1 = pix[ 1*xstride];
- const int q2 = pix[ 2*xstride];
- if( FFABS( p0 - q0 ) < alpha &&
- FFABS( p1 - p0 ) < beta &&
- FFABS( q1 - q0 ) < beta ) {
- if(FFABS( p0 - q0 ) < (( alpha >> 2 ) + 2 )){
- if( FFABS( p2 - p0 ) < beta)
- {
- const int p3 = pix[-4*xstride];
- /* p0', p1', p2' */
- pix[-1*xstride] = ( p2 + 2*p1 + 2*p0 + 2*q0 + q1 + 4 ) >> 3;
- pix[-2*xstride] = ( p2 + p1 + p0 + q0 + 2 ) >> 2;
- pix[-3*xstride] = ( 2*p3 + 3*p2 + p1 + p0 + q0 + 4 ) >> 3;
- } else {
- /* p0' */
- pix[-1*xstride] = ( 2*p1 + p0 + q1 + 2 ) >> 2;
- }
- if( FFABS( q2 - q0 ) < beta)
- {
- const int q3 = pix[3*xstride];
- /* q0', q1', q2' */
- pix[0*xstride] = ( p1 + 2*p0 + 2*q0 + 2*q1 + q2 + 4 ) >> 3;
- pix[1*xstride] = ( p0 + q0 + q1 + q2 + 2 ) >> 2;
- pix[2*xstride] = ( 2*q3 + 3*q2 + q1 + q0 + p0 + 4 ) >> 3;
- } else {
- /* q0' */
- pix[0*xstride] = ( 2*q1 + q0 + p1 + 2 ) >> 2;
- }
- }else{
- /* p0', q0' */
- pix[-1*xstride] = ( 2*p1 + p0 + q1 + 2 ) >> 2;
- pix[ 0*xstride] = ( 2*q1 + q0 + p1 + 2 ) >> 2;
- }
- }
- pix += ystride;
- }
- }
- static void h264_v_loop_filter_luma_intra_c(uint8_t *pix, int stride, int alpha, int beta)
- {
- h264_loop_filter_luma_intra_c(pix, stride, 1, alpha, beta);
- }
- static void h264_h_loop_filter_luma_intra_c(uint8_t *pix, int stride, int alpha, int beta)
- {
- h264_loop_filter_luma_intra_c(pix, 1, stride, alpha, beta);
- }
- static inline void h264_loop_filter_chroma_c(uint8_t *pix, int xstride, int ystride, int alpha, int beta, int8_t *tc0)
- {
- int i, d;
- for( i = 0; i < 4; i++ ) {
- const int tc = tc0[i];
- if( tc <= 0 ) {
- pix += 2*ystride;
- continue;
- }
- for( d = 0; d < 2; d++ ) {
- const int p0 = pix[-1*xstride];
- const int p1 = pix[-2*xstride];
- const int q0 = pix[0];
- const int q1 = pix[1*xstride];
- if( FFABS( p0 - q0 ) < alpha &&
- FFABS( p1 - p0 ) < beta &&
- FFABS( q1 - q0 ) < beta ) {
- int delta = av_clip( (((q0 - p0 ) << 2) + (p1 - q1) + 4) >> 3, -tc, tc );
- pix[-xstride] = av_clip_uint8( p0 + delta ); /* p0' */
- pix[0] = av_clip_uint8( q0 - delta ); /* q0' */
- }
- pix += ystride;
- }
- }
- }
- static void h264_v_loop_filter_chroma_c(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
- {
- h264_loop_filter_chroma_c(pix, stride, 1, alpha, beta, tc0);
- }
- static void h264_h_loop_filter_chroma_c(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
- {
- h264_loop_filter_chroma_c(pix, 1, stride, alpha, beta, tc0);
- }
- static inline void h264_loop_filter_chroma_intra_c(uint8_t *pix, int xstride, int ystride, int alpha, int beta)
- {
- int d;
- for( d = 0; d < 8; d++ ) {
- const int p0 = pix[-1*xstride];
- const int p1 = pix[-2*xstride];
- const int q0 = pix[0];
- const int q1 = pix[1*xstride];
- if( FFABS( p0 - q0 ) < alpha &&
- FFABS( p1 - p0 ) < beta &&
- FFABS( q1 - q0 ) < beta ) {
- pix[-xstride] = ( 2*p1 + p0 + q1 + 2 ) >> 2; /* p0' */
- pix[0] = ( 2*q1 + q0 + p1 + 2 ) >> 2; /* q0' */
- }
- pix += ystride;
- }
- }
- static void h264_v_loop_filter_chroma_intra_c(uint8_t *pix, int stride, int alpha, int beta)
- {
- h264_loop_filter_chroma_intra_c(pix, stride, 1, alpha, beta);
- }
- static void h264_h_loop_filter_chroma_intra_c(uint8_t *pix, int stride, int alpha, int beta)
- {
- h264_loop_filter_chroma_intra_c(pix, 1, stride, alpha, beta);
- }
- static inline int pix_abs16_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
- {
- int s, i;
- s = 0;
- for(i=0;i<h;i++) {
- s += abs(pix1[0] - pix2[0]);
- s += abs(pix1[1] - pix2[1]);
- s += abs(pix1[2] - pix2[2]);
- s += abs(pix1[3] - pix2[3]);
- s += abs(pix1[4] - pix2[4]);
- s += abs(pix1[5] - pix2[5]);
- s += abs(pix1[6] - pix2[6]);
- s += abs(pix1[7] - pix2[7]);
- s += abs(pix1[8] - pix2[8]);
- s += abs(pix1[9] - pix2[9]);
- s += abs(pix1[10] - pix2[10]);
- s += abs(pix1[11] - pix2[11]);
- s += abs(pix1[12] - pix2[12]);
- s += abs(pix1[13] - pix2[13]);
- s += abs(pix1[14] - pix2[14]);
- s += abs(pix1[15] - pix2[15]);
- pix1 += line_size;
- pix2 += line_size;
- }
- return s;
- }
- static int pix_abs16_x2_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
- {
- int s, i;
- s = 0;
- for(i=0;i<h;i++) {
- s += abs(pix1[0] - avg2(pix2[0], pix2[1]));
- s += abs(pix1[1] - avg2(pix2[1], pix2[2]));
- s += abs(pix1[2] - avg2(pix2[2], pix2[3]));
- s += abs(pix1[3] - avg2(pix2[3], pix2[4]));
- s += abs(pix1[4] - avg2(pix2[4], pix2[5]));
- s += abs(pix1[5] - avg2(pix2[5], pix2[6]));
- s += abs(pix1[6] - avg2(pix2[6], pix2[7]));
- s += abs(pix1[7] - avg2(pix2[7], pix2[8]));
- s += abs(pix1[8] - avg2(pix2[8], pix2[9]));
- s += abs(pix1[9] - avg2(pix2[9], pix2[10]));
- s += abs(pix1[10] - avg2(pix2[10], pix2[11]));
- s += abs(pix1[11] - avg2(pix2[11], pix2[12]));
- s += abs(pix1[12] - avg2(pix2[12], pix2[13]));
- s += abs(pix1[13] - avg2(pix2[13], pix2[14]));
- s += abs(pix1[14] - avg2(pix2[14], pix2[15]));
- s += abs(pix1[15] - avg2(pix2[15], pix2[16]));
- pix1 += line_size;
- pix2 += line_size;
- }
- return s;
- }
- static int pix_abs16_y2_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
- {
- int s, i;
- uint8_t *pix3 = pix2 + line_size;
- s = 0;
- for(i=0;i<h;i++) {
- s += abs(pix1[0] - avg2(pix2[0], pix3[0]));
- s += abs(pix1[1] - avg2(pix2[1], pix3[1]));
- s += abs(pix1[2] - avg2(pix2[2], pix3[2]));
- s += abs(pix1[3] - avg2(pix2[3], pix3[3]));
- s += abs(pix1[4] - avg2(pix2[4], pix3[4]));
- s += abs(pix1[5] - avg2(pix2[5], pix3[5]));
- s += abs(pix1[6] - avg2(pix2[6], pix3[6]));
- s += abs(pix1[7] - avg2(pix2[7], pix3[7]));
- s += abs(pix1[8] - avg2(pix2[8], pix3[8]));
- s += abs(pix1[9] - avg2(pix2[9], pix3[9]));
- s += abs(pix1[10] - avg2(pix2[10], pix3[10]));
- s += abs(pix1[11] - avg2(pix2[11], pix3[11]));
- s += abs(pix1[12] - avg2(pix2[12], pix3[12]));
- s += abs(pix1[13] - avg2(pix2[13], pix3[13]));
- s += abs(pix1[14] - avg2(pix2[14], pix3[14]));
- s += abs(pix1[15] - avg2(pix2[15], pix3[15]));
- pix1 += line_size;
- pix2 += line_size;
- pix3 += line_size;
- }
- return s;
- }
- static int pix_abs16_xy2_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
- {
- int s, i;
- uint8_t *pix3 = pix2 + line_size;
- s = 0;
- for(i=0;i<h;i++) {
- s += abs(pix1[0] - avg4(pix2[0], pix2[1], pix3[0], pix3[1]));
- s += abs(pix1[1] - avg4(pix2[1], pix2[2], pix3[1], pix3[2]));
- s += abs(pix1[2] - avg4(pix2[2], pix2[3], pix3[2], pix3[3]));
- s += abs(pix1[3] - avg4(pix2[3], pix2[4], pix3[3], pix3[4]));
- s += abs(pix1[4] - avg4(pix2[4], pix2[5], pix3[4], pix3[5]));
- s += abs(pix1[5] - avg4(pix2[5], pix2[6], pix3[5], pix3[6]));
- s += abs(pix1[6] - avg4(pix2[6], pix2[7], pix3[6], pix3[7]));
- s += abs(pix1[7] - avg4(pix2[7], pix2[8], pix3[7], pix3[8]));
- s += abs(pix1[8] - avg4(pix2[8], pix2[9], pix3[8], pix3[9]));
- s += abs(pix1[9] - avg4(pix2[9], pix2[10], pix3[9], pix3[10]));
- s += abs(pix1[10] - avg4(pix2[10], pix2[11], pix3[10], pix3[11]));
- s += abs(pix1[11] - avg4(pix2[11], pix2[12], pix3[11], pix3[12]));
- s += abs(pix1[12] - avg4(pix2[12], pix2[13], pix3[12], pix3[13]));
- s += abs(pix1[13] - avg4(pix2[13], pix2[14], pix3[13], pix3[14]));
- s += abs(pix1[14] - avg4(pix2[14], pix2[15], pix3[14], pix3[15]));
- s += abs(pix1[15] - avg4(pix2[15], pix2[16], pix3[15], pix3[16]));
- pix1 += line_size;
- pix2 += line_size;
- pix3 += line_size;
- }
- return s;
- }
- static inline int pix_abs8_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
- {
- int s, i;
- s = 0;
- for(i=0;i<h;i++) {
- s += abs(pix1[0] - pix2[0]);
- s += abs(pix1[1] - pix2[1]);
- s += abs(pix1[2] - pix2[2]);
- s += abs(pix1[3] - pix2[3]);
- s += abs(pix1[4] - pix2[4]);
- s += abs(pix1[5] - pix2[5]);
- s += abs(pix1[6] - pix2[6]);
- s += abs(pix1[7] - pix2[7]);
- pix1 += line_size;
- pix2 += line_size;
- }
- return s;
- }
- static int pix_abs8_x2_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
- {
- int s, i;
- s = 0;
- for(i=0;i<h;i++) {
- s += abs(pix1[0] - avg2(pix2[0], pix2[1]));
- s += abs(pix1[1] - avg2(pix2[1], pix2[2]));
- s += abs(pix1[2] - avg2(pix2[2], pix2[3]));
- s += abs(pix1[3] - avg2(pix2[3], pix2[4]));
- s += abs(pix1[4] - avg2(pix2[4], pix2[5]));
- s += abs(pix1[5] - avg2(pix2[5], pix2[6]));
- s += abs(pix1[6] - avg2(pix2[6], pix2[7]));
- s += abs(pix1[7] - avg2(pix2[7], pix2[8]));
- pix1 += line_size;
- pix2 += line_size;
- }
- return s;
- }
- static int pix_abs8_y2_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
- {
- int s, i;
- uint8_t *pix3 = pix2 + line_size;
- s = 0;
- for(i=0;i<h;i++) {
- s += abs(pix1[0] - avg2(pix2[0], pix3[0]));
- s += abs(pix1[1] - avg2(pix2[1], pix3[1]));
- s += abs(pix1[2] - avg2(pix2[2], pix3[2]));
- s += abs(pix1[3] - avg2(pix2[3], pix3[3]));
- s += abs(pix1[4] - avg2(pix2[4], pix3[4]));
- s += abs(pix1[5] - avg2(pix2[5], pix3[5]));
- s += abs(pix1[6] - avg2(pix2[6], pix3[6]));
- s += abs(pix1[7] - avg2(pix2[7], pix3[7]));
- pix1 += line_size;
- pix2 += line_size;
- pix3 += line_size;
- }
- return s;
- }
- static int pix_abs8_xy2_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
- {
- int s, i;
- uint8_t *pix3 = pix2 + line_size;
- s = 0;
- for(i=0;i<h;i++) {
- s += abs(pix1[0] - avg4(pix2[0], pix2[1], pix3[0], pix3[1]));
- s += abs(pix1[1] - avg4(pix2[1], pix2[2], pix3[1], pix3[2]));
- s += abs(pix1[2] - avg4(pix2[2], pix2[3], pix3[2], pix3[3]));
- s += abs(pix1[3] - avg4(pix2[3], pix2[4], pix3[3], pix3[4]));
- s += abs(pix1[4] - avg4(pix2[4], pix2[5], pix3[4], pix3[5]));
- s += abs(pix1[5] - avg4(pix2[5], pix2[6], pix3[5], pix3[6]));
- s += abs(pix1[6] - avg4(pix2[6], pix2[7], pix3[6], pix3[7]));
- s += abs(pix1[7] - avg4(pix2[7], pix2[8], pix3[7], pix3[8]));
- pix1 += line_size;
- pix2 += line_size;
- pix3 += line_size;
- }
- return s;
- }
- static int nsse16_c(void *v, uint8_t *s1, uint8_t *s2, int stride, int h){
- MpegEncContext *c = v;
- int score1=0;
- int score2=0;
- int x,y;
- for(y=0; y<h; y++){
- for(x=0; x<16; x++){
- score1+= (s1[x ] - s2[x ])*(s1[x ] - s2[x ]);
- }
- if(y+1<h){
- for(x=0; x<15; x++){
- score2+= FFABS( s1[x ] - s1[x +stride]
- - s1[x+1] + s1[x+1+stride])
- -FFABS( s2[x ] - s2[x +stride]
- - s2[x+1] + s2[x+1+stride]);
- }
- }
- s1+= stride;
- s2+= stride;
- }
- if(c) return score1 + FFABS(score2)*c->avctx->nsse_weight;
- else return score1 + FFABS(score2)*8;
- }
- static int nsse8_c(void *v, uint8_t *s1, uint8_t *s2, int stride, int h){
- MpegEncContext *c = v;
- int score1=0;
- int score2=0;
- int x,y;
- for(y=0; y<h; y++){
- for(x=0; x<8; x++){
- score1+= (s1[x ] - s2[x ])*(s1[x ] - s2[x ]);
- }
- if(y+1<h){
- for(x=0; x<7; x++){
- score2+= FFABS( s1[x ] - s1[x +stride]
- - s1[x+1] + s1[x+1+stride])
- -FFABS( s2[x ] - s2[x +stride]
- - s2[x+1] + s2[x+1+stride]);
- }
- }
- s1+= stride;
- s2+= stride;
- }
- if(c) return score1 + FFABS(score2)*c->avctx->nsse_weight;
- else return score1 + FFABS(score2)*8;
- }
- static int try_8x8basis_c(int16_t rem[64], int16_t weight[64], int16_t basis[64], int scale){
- int i;
- unsigned int sum=0;
- for(i=0; i<8*8; i++){
- int b= rem[i] + ((basis[i]*scale + (1<<(BASIS_SHIFT - RECON_SHIFT-1)))>>(BASIS_SHIFT - RECON_SHIFT));
- int w= weight[i];
- b>>= RECON_SHIFT;
- assert(-512<b && b<512);
- sum += (w*b)*(w*b)>>4;
- }
- return sum>>2;
- }
- static void add_8x8basis_c(int16_t rem[64], int16_t basis[64], int scale){
- int i;
- for(i=0; i<8*8; i++){
- rem[i] += (basis[i]*scale + (1<<(BASIS_SHIFT - RECON_SHIFT-1)))>>(BASIS_SHIFT - RECON_SHIFT);
- }
- }
- /**
- * permutes an 8x8 block.
- * @param block the block which will be permuted according to the given permutation vector
- * @param permutation the permutation vector
- * @param last the last non zero coefficient in scantable order, used to speed the permutation up
- * @param scantable the used scantable, this is only used to speed the permutation up, the block is not
- * (inverse) permutated to scantable order!
- */
- void ff_block_permute(DCTELEM *block, uint8_t *permutation, const uint8_t *scantable, int last)
- {
- int i;
- DCTELEM temp[64];
- if(last<=0) return;
- //if(permutation[1]==1) return; //FIXME it is ok but not clean and might fail for some permutations
- for(i=0; i<=last; i++){
- const int j= scantable[i];
- temp[j]= block[j];
- block[j]=0;
- }
- for(i=0; i<=last; i++){
- const int j= scantable[i];
- const int perm_j= permutation[j];
- block[perm_j]= temp[j];
- }
- }
- static int zero_cmp(void *s, uint8_t *a, uint8_t *b, int stride, int h){
- return 0;
- }
- void ff_set_cmp(DSPContext* c, me_cmp_func *cmp, int type){
- int i;
- memset(cmp, 0, sizeof(void*)*5);
- for(i=0; i<5; i++){
- switch(type&0xFF){
- case FF_CMP_SAD:
- cmp[i]= c->sad[i];
- break;
- case FF_CMP_SATD:
- cmp[i]= c->hadamard8_diff[i];
- break;
- case FF_CMP_SSE:
- cmp[i]= c->sse[i];
- break;
- case FF_CMP_DCT:
- cmp[i]= c->dct_sad[i];
- break;
- case FF_CMP_DCT264:
- cmp[i]= c->dct264_sad[i];
- break;
- case FF_CMP_DCTMAX:
- cmp[i]= c->dct_max[i];
- break;
- case FF_CMP_PSNR:
- cmp[i]= c->quant_psnr[i];
- break;
- case FF_CMP_BIT:
- cmp[i]= c->bit[i];
- break;
- case FF_CMP_RD:
- cmp[i]= c->rd[i];
- break;
- case FF_CMP_VSAD:
- cmp[i]= c->vsad[i];
- break;
- case FF_CMP_VSSE:
- cmp[i]= c->vsse[i];
- break;
- case FF_CMP_ZERO:
- cmp[i]= zero_cmp;
- break;
- case FF_CMP_NSSE:
- cmp[i]= c->nsse[i];
- break;
- #ifdef CONFIG_SNOW_ENCODER
- case FF_CMP_W53:
- cmp[i]= c->w53[i];
- break;
- case FF_CMP_W97:
- cmp[i]= c->w97[i];
- break;
- #endif
- default:
- av_log(NULL, AV_LOG_ERROR,"internal error in cmp function selectionn");
- }
- }
- }
- static void clear_block_c(DCTELEM *block)
- {
- memset(block, 0, sizeof(DCTELEM)*64);
- }
- /**
- * memset(blocks, 0, sizeof(DCTELEM)*6*64)
- */
- static void clear_blocks_c(DCTELEM *blocks)
- {
- memset(blocks, 0, sizeof(DCTELEM)*6*64);
- }
- static void add_bytes_c(uint8_t *dst, uint8_t *src, int w){
- long i;
- for(i=0; i<=w-(long)sizeof(long); i+=(long)sizeof(long)){
- long a = *(long*)(src+i);
- long b = *(long*)(dst+i);
- *(long*)(dst+i) = ((a&pb_7f) + (b&pb_7f)) ^ ((a^b)&pb_80);
- }
- for(; i<w; i++)
- dst[i+0] += src[i+0];
- }
- static void add_bytes_l2_c(uint8_t *dst, uint8_t *src1, uint8_t *src2, int w){
- long i;
- for(i=0; i<=w-(long)sizeof(long); i+=(long)sizeof(long)){
- long a = *(long*)(src1+i);
- long b = *(long*)(src2+i);
- *(long*)(dst+i) = ((a&pb_7f) + (b&pb_7f)) ^ ((a^b)&pb_80);
- }
- for(; i<w; i++)
- dst[i] = src1[i]+src2[i];
- }
- static void diff_bytes_c(uint8_t *dst, uint8_t *src1, uint8_t *src2, int w){
- long i;
- #ifndef HAVE_FAST_UNALIGNED
- if((long)src2 & (sizeof(long)-1)){
- for(i=0; i+7<w; i+=8){
- dst[i+0] = src1[i+0]-src2[i+0];
- dst[i+1] = src1[i+1]-src2[i+1];
- dst[i+2] = src1[i+2]-src2[i+2];
- dst[i+3] = src1[i+3]-src2[i+3];
- dst[i+4] = src1[i+4]-src2[i+4];
- dst[i+5] = src1[i+5]-src2[i+5];
- dst[i+6] = src1[i+6]-src2[i+6];
- dst[i+7] = src1[i+7]-src2[i+7];
- }
- }else
- #endif
- for(i=0; i<=w-(long)sizeof(long); i+=(long)sizeof(long)){
- long a = *(long*)(src1+i);
- long b = *(long*)(src2+i);
- *(long*)(dst+i) = ((a|pb_80) - (b&pb_7f)) ^ ((a^b^pb_80)&pb_80);
- }
- for(; i<w; i++)
- dst[i+0] = src1[i+0]-src2[i+0];
- }
- static void sub_hfyu_median_prediction_c(uint8_t *dst, uint8_t *src1, uint8_t *src2, int w, int *left, int *left_top){
- int i;
- uint8_t l, lt;
- l= *left;
- lt= *left_top;
- for(i=0; i<w; i++){
- const int pred= mid_pred(l, src1[i], (l + src1[i] - lt)&0xFF);
- lt= src1[i];
- l= src2[i];
- dst[i]= l - pred;
- }
- *left= l;
- *left_top= lt;
- }
- #define BUTTERFLY2(o1,o2,i1,i2)
- o1= (i1)+(i2);
- o2= (i1)-(i2);
- #define BUTTERFLY1(x,y)
- {
- int a,b;
- a= x;
- b= y;
- x= a+b;
- y= a-b;
- }
- #define BUTTERFLYA(x,y) (FFABS((x)+(y)) + FFABS((x)-(y)))
- static int hadamard8_diff8x8_c(/*MpegEncContext*/ void *s, uint8_t *dst, uint8_t *src, int stride, int h){
- int i;
- int temp[64];
- int sum=0;
- assert(h==8);
- for(i=0; i<8; i++){
- //FIXME try pointer walks
- BUTTERFLY2(temp[8*i+0], temp[8*i+1], src[stride*i+0]-dst[stride*i+0],src[stride*i+1]-dst[stride*i+1]);
- BUTTERFLY2(temp[8*i+2], temp[8*i+3], src[stride*i+2]-dst[stride*i+2],src[stride*i+3]-dst[stride*i+3]);
- BUTTERFLY2(temp[8*i+4], temp[8*i+5], src[stride*i+4]-dst[stride*i+4],src[stride*i+5]-dst[stride*i+5]);
- BUTTERFLY2(temp[8*i+6], temp[8*i+7], src[stride*i+6]-dst[stride*i+6],src[stride*i+7]-dst[stride*i+7]);
- BUTTERFLY1(temp[8*i+0], temp[8*i+2]);
- BUTTERFLY1(temp[8*i+1], temp[8*i+3]);
- BUTTERFLY1(temp[8*i+4], temp[8*i+6]);
- BUTTERFLY1(temp[8*i+5], temp[8*i+7]);
- BUTTERFLY1(temp[8*i+0], temp[8*i+4]);
- BUTTERFLY1(temp[8*i+1], temp[8*i+5]);
- BUTTERFLY1(temp[8*i+2], temp[8*i+6]);
- BUTTERFLY1(temp[8*i+3], temp[8*i+7]);
- }
- for(i=0; i<8; i++){
- BUTTERFLY1(temp[8*0+i], temp[8*1+i]);
- BUTTERFLY1(temp[8*2+i], temp[8*3+i]);
- BUTTERFLY1(temp[8*4+i], temp[8*5+i]);
- BUTTERFLY1(temp[8*6+i], temp[8*7+i]);
- BUTTERFLY1(temp[8*0+i], temp[8*2+i]);
- BUTTERFLY1(temp[8*1+i], temp[8*3+i]);
- BUTTERFLY1(temp[8*4+i], temp[8*6+i]);
- BUTTERFLY1(temp[8*5+i], temp[8*7+i]);
- sum +=
- BUTTERFLYA(temp[8*0+i], temp[8*4+i])
- +BUTTERFLYA(temp[8*1+i], temp[8*5+i])
- +BUTTERFLYA(temp[8*2+i], temp[8*6+i])
- +BUTTERFLYA(temp[8*3+i], temp[8*7+i]);
- }
- #if 0
- static int maxi=0;
- if(sum>maxi){
- maxi=sum;
- printf("MAX:%dn", maxi);
- }
- #endif
- return sum;
- }
- static int hadamard8_intra8x8_c(/*MpegEncContext*/ void *s, uint8_t *src, uint8_t *dummy, int stride, int h){
- int i;
- int temp[64];
- int sum=0;
- assert(h==8);
- for(i=0; i<8; i++){
- //FIXME try pointer walks
- BUTTERFLY2(temp[8*i+0], temp[8*i+1], src[stride*i+0],src[stride*i+1]);
- BUTTERFLY2(temp[8*i+2], temp[8*i+3], src[stride*i+2],src[stride*i+3]);
- BUTTERFLY2(temp[8*i+4], temp[8*i+5], src[stride*i+4],src[stride*i+5]);
- BUTTERFLY2(temp[8*i+6], temp[8*i+7], src[stride*i+6],src[stride*i+7]);
- BUTTERFLY1(temp[8*i+0], temp[8*i+2]);
- BUTTERFLY1(temp[8*i+1], temp[8*i+3]);
- BUTTERFLY1(temp[8*i+4], temp[8*i+6]);
- BUTTERFLY1(temp[8*i+5], temp[8*i+7]);
- BUTTERFLY1(temp[8*i+0], temp[8*i+4]);
- BUTTERFLY1(temp[8*i+1], temp[8*i+5]);
- BUTTERFLY1(temp[8*i+2], temp[8*i+6]);
- BUTTERFLY1(temp[8*i+3], temp[8*i+7]);
- }
- for(i=0; i<8; i++){
- BUTTERFLY1(temp[8*0+i], temp[8*1+i]);
- BUTTERFLY1(temp[8*2+i], temp[8*3+i]);
- BUTTERFLY1(temp[8*4+i], temp[8*5+i]);
- BUTTERFLY1(temp[8*6+i], temp[8*7+i]);
- BUTTERFLY1(temp[8*0+i], temp[8*2+i]);
- BUTTERFLY1(temp[8*1+i], temp[8*3+i]);
- BUTTERFLY1(temp[8*4+i], temp[8*6+i]);
- BUTTERFLY1(temp[8*5+i], temp[8*7+i]);
- sum +=
- BUTTERFLYA(temp[8*0+i], temp[8*4+i])
- +BUTTERFLYA(temp[8*1+i], temp[8*5+i])
- +BUTTERFLYA(temp[8*2+i], temp[8*6+i])
- +BUTTERFLYA(temp[8*3+i], temp[8*7+i]);
- }
- sum -= FFABS(temp[8*0] + temp[8*4]); // -mean
- return sum;
- }
- static int dct_sad8x8_c(/*MpegEncContext*/ void *c, uint8_t *src1, uint8_t *src2, int stride, int h){
- MpegEncContext * const s= (MpegEncContext *)c;
- DECLARE_ALIGNED_16(uint64_t, aligned_temp[sizeof(DCTELEM)*64/8]);
- DCTELEM * const temp= (DCTELEM*)aligned_temp;
- assert(h==8);
- s->dsp.diff_pixels(temp, src1, src2, stride);
- s->dsp.fdct(temp);
- return s->dsp.sum_abs_dctelem(temp);
- }
- #ifdef CONFIG_GPL
- #define DCT8_1D {
- const int s07 = SRC(0) + SRC(7);
- const int s16 = SRC(1) + SRC(6);
- const int s25 = SRC(2) + SRC(5);
- const int s34 = SRC(3) + SRC(4);
- const int a0 = s07 + s34;
- const int a1 = s16 + s25;
- const int a2 = s07 - s34;
- const int a3 = s16 - s25;
- const int d07 = SRC(0) - SRC(7);
- const int d16 = SRC(1) - SRC(6);
- const int d25 = SRC(2) - SRC(5);
- const int d34 = SRC(3) - SRC(4);
- const int a4 = d16 + d25 + (d07 + (d07>>1));
- const int a5 = d07 - d34 - (d25 + (d25>>1));
- const int a6 = d07 + d34 - (d16 + (d16>>1));
- const int a7 = d16 - d25 + (d34 + (d34>>1));
- DST(0, a0 + a1 ) ;
- DST(1, a4 + (a7>>2)) ;
- DST(2, a2 + (a3>>1)) ;
- DST(3, a5 + (a6>>2)) ;
- DST(4, a0 - a1 ) ;
- DST(5, a6 - (a5>>2)) ;
- DST(6, (a2>>1) - a3 ) ;
- DST(7, (a4>>2) - a7 ) ;
- }
- static int dct264_sad8x8_c(/*MpegEncContext*/ void *c, uint8_t *src1, uint8_t *src2, int stride, int h){
- MpegEncContext * const s= (MpegEncContext *)c;
- DCTELEM dct[8][8];
- int i;
- int sum=0;
- s->dsp.diff_pixels(dct[0], src1, src2, stride);
- #define SRC(x) dct[i][x]
- #define DST(x,v) dct[i][x]= v
- for( i = 0; i < 8; i++ )
- DCT8_1D
- #undef SRC
- #undef DST
- #define SRC(x) dct[x][i]
- #define DST(x,v) sum += FFABS(v)
- for( i = 0; i < 8; i++ )
- DCT8_1D
- #undef SRC
- #undef DST
- return sum;
- }
- #endif
- static int dct_max8x8_c(/*MpegEncContext*/ void *c, uint8_t *src1, uint8_t *src2, int stride, int h){
- MpegEncContext * const s= (MpegEncContext *)c;
- DECLARE_ALIGNED_8(uint64_t, aligned_temp[sizeof(DCTELEM)*64/8]);
- DCTELEM * const temp= (DCTELEM*)aligned_temp;
- int sum=0, i;
- assert(h==8);
- s->dsp.diff_pixels(temp, src1, src2, stride);
- s->dsp.fdct(temp);
- for(i=0; i<64; i++)
- sum= FFMAX(sum, FFABS(temp[i]));
- return sum;
- }
- static int quant_psnr8x8_c(/*MpegEncContext*/ void *c, uint8_t *src1, uint8_t *src2, int stride, int h){
- MpegEncContext * const s= (MpegEncContext *)c;
- DECLARE_ALIGNED_8 (uint64_t, aligned_temp[sizeof(DCTELEM)*64*2/8]);
- DCTELEM * const temp= (DCTELEM*)aligned_temp;
- DCTELEM * const bak = ((DCTELEM*)aligned_temp)+64;
- int sum=0, i;
- assert(h==8);
- s->mb_intra=0;
- s->dsp.diff_pixels(temp, src1, src2, stride);
- memcpy(bak, temp, 64*sizeof(DCTELEM));
- s->block_last_index[0/*FIXME*/]= s->fast_dct_quantize(s, temp, 0/*FIXME*/, s->qscale, &i);
- s->dct_unquantize_inter(s, temp, 0, s->qscale);
- ff_simple_idct(temp); //FIXME
- for(i=0; i<64; i++)
- sum+= (temp[i]-bak[i])*(temp[i]-bak[i]);
- return sum;
- }
- static int rd8x8_c(/*MpegEncContext*/ void *c, uint8_t *src1, uint8_t *src2, int stride, int h){
- MpegEncContext * const s= (MpegEncContext *)c;
- const uint8_t *scantable= s->intra_scantable.permutated;
- DECLARE_ALIGNED_8 (uint64_t, aligned_temp[sizeof(DCTELEM)*64/8]);
- //DECLARE_ALIGNED_8 (uint64_t, aligned_bak[stride]);
- uint64_t* aligned_bak = (uint64_t*)av_malloc(stride*sizeof(uint64_t));
- DCTELEM * const temp= (DCTELEM*)aligned_temp;
- uint8_t * const bak= (uint8_t*)aligned_bak;
- int i, last, run, bits, level, distortion, start_i;
- const int esc_length= s->ac_esc_length;
- uint8_t * length;
- uint8_t * last_length;
- assert(h==8);
- for(i=0; i<8; i++){
- ((uint32_t*)(bak + i*stride))[0]= ((uint32_t*)(src2 + i*stride))[0];
- ((uint32_t*)(bak + i*stride))[1]= ((uint32_t*)(src2 + i*stride))[1];
- }
- s->dsp.diff_pixels(temp, src1, src2, stride);
- s->block_last_index[0/*FIXME*/]= last= s->fast_dct_quantize(s, temp, 0/*FIXME*/, s->qscale, &i);
- bits=0;
- if (s->mb_intra) {
- start_i = 1;
- length = s->intra_ac_vlc_length;
- last_length= s->intra_ac_vlc_last_length;
- bits+= s->luma_dc_vlc_length[temp[0] + 256]; //FIXME chroma
- } else {
- start_i = 0;
- length = s->inter_ac_vlc_length;
- last_length= s->inter_ac_vlc_last_length;
- }
- if(last>=start_i){
- run=0;
- for(i=start_i; i<last; i++){
- int j= scantable[i];
- level= temp[j];
- if(level){
- level+=64;
- if((level&(~127)) == 0){
- bits+= length[UNI_AC_ENC_INDEX(run, level)];
- }else
- bits+= esc_length;
- run=0;
- }else
- run++;
- }
- i= scantable[last];
- level= temp[i] + 64;
- assert(level - 64);
- if((level&(~127)) == 0){
- bits+= last_length[UNI_AC_ENC_INDEX(run, level)];
- }else
- bits+= esc_length;
- }
- if(last>=0){
- if(s->mb_intra)
- s->dct_unquantize_intra(s, temp, 0, s->qscale);
- else
- s->dct_unquantize_inter(s, temp, 0, s->qscale);
- }
- s->dsp.idct_add(bak, stride, temp);
- distortion= s->dsp.sse[1](NULL, bak, src1, stride, 8);
- av_free(aligned_bak);
- return distortion + ((bits*s->qscale*s->qscale*109 + 64)>>7);
- }
- static int bit8x8_c(/*MpegEncContext*/ void *c, uint8_t *src1, uint8_t *src2, int stride, int h){
- MpegEncContext * const s= (MpegEncContext *)c;
- const uint8_t *scantable= s->intra_scantable.permutated;
- DECLARE_ALIGNED_8 (uint64_t, aligned_temp[sizeof(DCTELEM)*64/8]);
- DCTELEM * const temp= (DCTELEM*)aligned_temp;
- int i, last, run, bits, level, start_i;
- const int esc_length= s->ac_esc_length;
- uint8_t * length;
- uint8_t * last_length;
- assert(h==8);
- s->dsp.diff_pixels(temp, src1, src2, stride);
- s->block_last_index[0/*FIXME*/]= last= s->fast_dct_quantize(s, temp, 0/*FIXME*/, s->qscale, &i);
- bits=0;
- if (s->mb_intra) {
- start_i = 1;
- length = s->intra_ac_vlc_length;
- last_length= s->intra_ac_vlc_last_length;
- bits+= s->luma_dc_vlc_length[temp[0] + 256]; //FIXME chroma
- } else {
- start_i = 0;
- length = s->inter_ac_vlc_length;
- last_length= s->inter_ac_vlc_last_length;
- }
- if(last>=start_i){
- run=0;
- for(i=start_i; i<last; i++){
- int j= scantable[i];
- level= temp[j];
- if(level){
- level+=64;
- if((level&(~127)) == 0){
- bits+= length[UNI_AC_ENC_INDEX(run, level)];
- }else
- bits+= esc_length;
- run=0;
- }else
- run++;
- }
- i= scantable[last];
- level= temp[i] + 64;
- assert(level - 64);
- if((level&(~127)) == 0){
- bits+= last_length[UNI_AC_ENC_INDEX(run, level)];
- }else
- bits+= esc_length;
- }
- return bits;
- }
- static int vsad_intra16_c(/*MpegEncContext*/ void *c, uint8_t *s, uint8_t *dummy, int stride, int h){
- int score=0;
- int x,y;
- for(y=1; y<h; y++){
- for(x=0; x<16; x+=4){
- score+= FFABS(s[x ] - s[x +stride]) + FFABS(s[x+1] - s[x+1+stride])
- +FFABS(s[x+2] - s[x+2+stride]) + FFABS(s[x+3] - s[x+3+stride]);
- }
- s+= stride;
- }
- return score;
- }
- static int vsad16_c(/*MpegEncContext*/ void *c, uint8_t *s1, uint8_t *s2, int stride, int h){
- int score=0;
- int x,y;
- for(y=1; y<h; y++){
- for(x=0; x<16; x++){
- score+= FFABS(s1[x ] - s2[x ] - s1[x +stride] + s2[x +stride]);
- }
- s1+= stride;
- s2+= stride;
- }
- return score;
- }
- #define SQ(a) ((a)*(a))
- static int vsse_intra16_c(/*MpegEncContext*/ void *c, uint8_t *s, uint8_t *dummy, int stride, int h){
- int score=0;
- int x,y;
- for(y=1; y<h; y++){
- for(x=0; x<16; x+=4){
- score+= SQ(s[x ] - s[x +stride]) + SQ(s[x+1] - s[x+1+stride])
- +SQ(s[x+2] - s[x+2+stride]) + SQ(s[x+3] - s[x+3+stride]);
- }
- s+= stride;
- }
- return score;
- }
- static int vsse16_c(/*MpegEncContext*/ void *c, uint8_t *s1, uint8_t *s2, int stride, int h){
- int score=0;
- int x,y;
- for(y=1; y<h; y++){
- for(x=0; x<16; x++){
- score+= SQ(s1[x ] - s2[x ] - s1[x +stride] + s2[x +stride]);
- }
- s1+= stride;
- s2+= stride;
- }
- return score;
- }
- static int ssd_int8_vs_int16_c(const int8_t *pix1, const int16_t *pix2,
- int size){
- int score=0;
- int i;
- for(i=0; i<size; i++)
- score += (pix1[i]-pix2[i])*(pix1[i]-pix2[i]);
- return score;
- }
- WRAPPER8_16_SQ(hadamard8_diff8x8_c, hadamard8_diff16_c)
- WRAPPER8_16_SQ(hadamard8_intra8x8_c, hadamard8_intra16_c)
- WRAPPER8_16_SQ(dct_sad8x8_c, dct_sad16_c)
- #ifdef CONFIG_GPL
- WRAPPER8_16_SQ(dct264_sad8x8_c, dct264_sad16_c)
- #endif
- WRAPPER8_16_SQ(dct_max8x8_c, dct_max16_c)
- WRAPPER8_16_SQ(quant_psnr8x8_c, quant_psnr16_c)
- WRAPPER8_16_SQ(rd8x8_c, rd16_c)
- WRAPPER8_16_SQ(bit8x8_c, bit16_c)
- static void vector_fmul_c(float *dst, const float *src, int len){
- int i;
- for(i=0; i<len; i++)
- dst[i] *= src[i];
- }
- static void vector_fmul_reverse_c(float *dst, const float *src0, const float *src1, int len){
- int i;
- src1 += len-1;
- for(i=0; i<len; i++)
- dst[i] = src0[i] * src1[-i];
- }
- void ff_vector_fmul_add_add_c(float *dst, const float *src0, const float *src1, const float *src2, int src3, int len, int step){
- int i;
- for(i=0; i<len; i++)
- dst[i*step] = src0[i] * src1[i] + src2[i] + src3;
- }
- void ff_vector_fmul_window_c(float *dst, const float *src0, const float *src1, const float *win, float add_bias, int len){
- int i,j;
- dst += len;
- win += len;
- src0+= len;
- for(i=-len, j=len-1; i<0; i++, j--) {
- float s0 = src0[i];
- float s1 = src1[j];
- float wi = win[i];
- float wj = win[j];
- dst[i] = s0*wj - s1*wi + add_bias;
- dst[j] = s0*wi + s1*wj + add_bias;
- }
- }
- static void int32_to_float_fmul_scalar_c(float *dst, const int *src, float mul, int len){
- int i;
- for(i=0; i<len; i++)
- dst[i] = src[i] * mul;
- }
- static av_always_inline int float_to_int16_one(const float *src){
- int_fast32_t tmp = *(const int32_t*)src;
- if(tmp & 0xf0000){
- tmp = (0x43c0ffff - tmp)>>31;
- // is this faster on some gcc/cpu combinations?
- // if(tmp > 0x43c0ffff) tmp = 0xFFFF;
- // else tmp = 0;
- }
- return tmp - 0x8000;
- }
- void ff_float_to_int16_c(int16_t *dst, const float *src, long len){
- int i;
- for(i=0; i<len; i++)
- dst[i] = float_to_int16_one(src+i);
- }
- void ff_float_to_int16_interleave_c(int16_t *dst, const float **src, long len, int channels){
- int i,j,c;
- if(channels==2){
- for(i=0; i<len; i++){
- dst[2*i] = float_to_int16_one(src[0]+i);
- dst[2*i+1] = float_to_int16_one(src[1]+i);
- }
- }else{
- for(c=0; c<channels; c++)
- for(i=0, j=c; i<len; i++, j+=channels)
- dst[j] = float_to_int16_one(src[c]+i);
- }
- }
- static void add_int16_c(int16_t * v1, int16_t * v2, int order)
- {
- while (order--)
- *v1++ += *v2++;
- }
- static void sub_int16_c(int16_t * v1, int16_t * v2, int order)
- {
- while (order--)
- *v1++ -= *v2++;
- }
- static int32_t scalarproduct_int16_c(int16_t * v1, int16_t * v2, int order, int shift)
- {
- int res = 0;
- while (order--)
- res += (*v1++ * *v2++) >> shift;
- return res;
- }
- #define W0 2048
- #define W1 2841 /* 2048*sqrt (2)*cos (1*pi/16) */
- #define W2 2676 /* 2048*sqrt (2)*cos (2*pi/16) */
- #define W3 2408 /* 2048*sqrt (2)*cos (3*pi/16) */
- #define W4 2048 /* 2048*sqrt (2)*cos (4*pi/16) */
- #define W5 1609 /* 2048*sqrt (2)*cos (5*pi/16) */
- #define W6 1108 /* 2048*sqrt (2)*cos (6*pi/16) */
- #define W7 565 /* 2048*sqrt (2)*cos (7*pi/16) */
- static void wmv2_idct_row(short * b)
- {
- int s1,s2;
- int a0,a1,a2,a3,a4,a5,a6,a7;
- /*step 1*/
- a1 = W1*b[1]+W7*b[7];
- a7 = W7*b[1]-W1*b[7];
- a5 = W5*b[5]+W3*b[3];
- a3 = W3*b[5]-W5*b[3];
- a2 = W2*b[2]+W6*b[6];
- a6 = W6*b[2]-W2*b[6];
- a0 = W0*b[0]+W0*b[4];
- a4 = W0*b[0]-W0*b[4];
- /*step 2*/
- s1 = (181*(a1-a5+a7-a3)+128)>>8;//1,3,5,7,
- s2 = (181*(a1-a5-a7+a3)+128)>>8;
- /*step 3*/
- b[0] = (a0+a2+a1+a5 + (1<<7))>>8;
- b[1] = (a4+a6 +s1 + (1<<7))>>8;
- b[2] = (a4-a6 +s2 + (1<<7))>>8;
- b[3] = (a0-a2+a7+a3 + (1<<7))>>8;
- b[4] = (a0-a2-a7-a3 + (1<<7))>>8;
- b[5] = (a4-a6 -s2 + (1<<7))>>8;
- b[6] = (a4+a6 -s1 + (1<<7))>>8;
- b[7] = (a0+a2-a1-a5 + (1<<7))>>8;
- }
- static void wmv2_idct_col(short * b)
- {
- int s1,s2;
- int a0,a1,a2,a3,a4,a5,a6,a7;
- /*step 1, with extended precision*/
- a1 = (W1*b[8*1]+W7*b[8*7] + 4)>>3;
- a7 = (W7*b[8*1]-W1*b[8*7] + 4)>>3;
- a5 = (W5*b[8*5]+W3*b[8*3] + 4)>>3;
- a3 = (W3*b[8*5]-W5*b[8*3] + 4)>>3;
- a2 = (W2*b[8*2]+W6*b[8*6] + 4)>>3;
- a6 = (W6*b[8*2]-W2*b[8*6] + 4)>>3;
- a0 = (W0*b[8*0]+W0*b[8*4] )>>3;
- a4 = (W0*b[8*0]-W0*b[8*4] )>>3;
- /*step 2*/
- s1 = (181*(a1-a5+a7-a3)+128)>>8;
- s2 = (181*(a1-a5-a7+a3)+128)>>8;
- /*step 3*/
- b[8*0] = (a0+a2+a1+a5 + (1<<13))>>14;
- b[8*1] = (a4+a6 +s1 + (1<<13))>>14;
- b[8*2] = (a4-a6 +s2 + (1<<13))>>14;
- b[8*3] = (a0-a2+a7+a3 + (1<<13))>>14;
- b[8*4] = (a0-a2-a7-a3 + (1<<13))>>14;
- b[8*5] = (a4-a6 -s2 + (1<<13))>>14;
- b[8*6] = (a4+a6 -s1 + (1<<13))>>14;
- b[8*7] = (a0+a2-a1-a5 + (1<<13))>>14;
- }
- void ff_wmv2_idct_c(short * block){
- int i;
- for(i=0;i<64;i+=8){
- wmv2_idct_row(block+i);
- }
- for(i=0;i<8;i++){
- wmv2_idct_col(block+i);
- }
- }
- /* XXX: those functions should be suppressed ASAP when all IDCTs are
- converted */
- /*
- static void ff_wmv2_idct_put_c(uint8_t *dest, int line_size, DCTELEM *block)
- {
- ff_wmv2_idct_c(block);
- put_pixels_clamped_c(block, dest, line_size);
- }
- static void ff_wmv2_idct_add_c(uint8_t *dest, int line_size, DCTELEM *block)
- {
- ff_wmv2_idct_c(block);
- add_pixels_clamped_c(block, dest, line_size);
- }
- */
- static void ff_jref_idct_put(uint8_t *dest, int line_size, DCTELEM *block)
- {
- j_rev_dct (block);
- put_pixels_clamped_c(block, dest, line_size);
- }
- static void ff_jref_idct_add(uint8_t *dest, int line_size, DCTELEM *block)
- {
- j_rev_dct (block);
- add_pixels_clamped_c(block, dest, line_size);
- }
- static void ff_jref_idct4_put(uint8_t *dest, int line_size, DCTELEM *block)
- {
- j_rev_dct4 (block);
- put_pixels_clamped4_c(block, dest, line_size);
- }
- static void ff_jref_idct4_add(uint8_t *dest, int line_size, DCTELEM *block)
- {
- j_rev_dct4 (block);
- add_pixels_clamped4_c(block, dest, line_size);
- }
- static void ff_jref_idct2_put(uint8_t *dest, int line_size, DCTELEM *block)
- {
- j_rev_dct2 (block);
- put_pixels_clamped2_c(block, dest, line_size);
- }
- static void ff_jref_idct2_add(uint8_t *dest, int line_size, DCTELEM *block)
- {
- j_rev_dct2 (block);
- add_pixels_clamped2_c(block, dest, line_size);
- }
- static void ff_jref_idct1_put(uint8_t *dest, int line_size, DCTELEM *block)
- {
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
- dest[0] = cm[(block[0] + 4)>>3];
- }
- static void ff_jref_idct1_add(uint8_t *dest, int line_size, DCTELEM *block)
- {
- uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
- dest[0] = cm[dest[0] + ((block[0] + 4)>>3)];
- }
- static void just_return(void *mem av_unused, int stride av_unused, int h av_unused) { return; }
- /* init static data */
- void dsputil_static_init(void)
- {
- int i;
- for(i=0;i<256;i++) ff_cropTbl[i + MAX_NEG_CROP] = i;
- for(i=0;i<MAX_NEG_CROP;i++) {
- ff_cropTbl[i] = 0;
- ff_cropTbl[i + MAX_NEG_CROP + 256] = 255;
- }
- for(i=0;i<512;i++) {
- ff_squareTbl[i] = (i - 256) * (i - 256);
- }
- for(i=0; i<64; i++) inv_zigzag_direct16[ff_zigzag_direct[i]]= i+1;
- }
- int ff_check_alignment(void){
- static int did_fail=0;
- DECLARE_ALIGNED_16(int, aligned);
- if((long)&aligned & 15){
- if(!did_fail){
- #if defined(HAVE_MMX) || defined(HAVE_ALTIVEC)
- av_log(NULL, AV_LOG_ERROR,
- "Compiler did not align stack variables. Libavcodec has been miscompiledn"
- "and may be very slow or crash. This is not a bug in libavcodec,n"
- "but in the compiler. You may try recompiling using gcc >= 4.2.n"
- "Do not report crashes to FFmpeg developers.n");
- #endif
- did_fail=1;
- }
- return -1;
- }
- return 0;
- }
- void dsputil_init(DSPContext* c, AVCodecContext *avctx)
- {
- int i;
- ff_check_alignment();
- #ifdef CONFIG_ENCODERS
- if(avctx->dct_algo==FF_DCT_FASTINT) {
- c->fdct = fdct_ifast;
- c->fdct248 = fdct_ifast248;
- }
- else if(avctx->dct_algo==FF_DCT_FAAN) {
- c->fdct = ff_faandct;
- c->fdct248 = ff_faandct248;
- }
- else {
- c->fdct = ff_jpeg_fdct_islow; //slow/accurate/default
- c->fdct248 = ff_fdct248_islow;
- }
- #endif //CONFIG_ENCODERS
- if(avctx->lowres==1){
- if(avctx->idct_algo==FF_IDCT_INT || avctx->idct_algo==FF_IDCT_AUTO || !ENABLE_H264_DECODER){
- c->idct_put= ff_jref_idct4_put;
- c->idct_add= ff_jref_idct4_add;
- }else{
- c->idct_put= ff_h264_lowres_idct_put_c;
- c->idct_add= ff_h264_lowres_idct_add_c;
- }
- c->idct = j_rev_dct4;
- c->idct_permutation_type= FF_NO_IDCT_PERM;
- }else if(avctx->lowres==2){
- c->idct_put= ff_jref_idct2_put;
- c->idct_add= ff_jref_idct2_add;
- c->idct = j_rev_dct2;
- c->idct_permutation_type= FF_NO_IDCT_PERM;
- }else if(avctx->lowres==3){
- c->idct_put= ff_jref_idct1_put;
- c->idct_add= ff_jref_idct1_add;
- c->idct = j_rev_dct1;
- c->idct_permutation_type= FF_NO_IDCT_PERM;
- }
- else
- {
- if(avctx->idct_algo==FF_IDCT_INT)
- {
- c->idct_put= ff_jref_idct_put;
- c->idct_add= ff_jref_idct_add;
- c->idct = j_rev_dct;
- c->idct_permutation_type= FF_LIBMPEG2_IDCT_PERM;
- }
- /*else if((ENABLE_VP3_DECODER || ENABLE_VP5_DECODER || ENABLE_VP6_DECODER || ENABLE_THEORA_DECODER ) && avctx->idct_algo==FF_IDCT_VP3)
- {
- c->idct_put= ff_vp3_idct_put_c;
- c->idct_add= ff_vp3_idct_add_c;
- c->idct = ff_vp3_idct_c;
- c->idct_permutation_type= FF_NO_IDCT_PERM;
- }
- else if(avctx->idct_algo==FF_IDCT_WMV2)
- {
- c->idct_put= ff_wmv2_idct_put_c;
- c->idct_add= ff_wmv2_idct_add_c;
- c->idct = ff_wmv2_idct_c;
- c->idct_permutation_type= FF_NO_IDCT_PERM;
- }
- else if(avctx->idct_algo==FF_IDCT_FAAN)
- {
- c->idct_put= ff_faanidct_put;
- c->idct_add= ff_faanidct_add;
- c->idct = ff_faanidct;
- c->idct_permutation_type= FF_NO_IDCT_PERM;
- }
- else if(ENABLE_EATGQ_DECODER && avctx->idct_algo==FF_IDCT_EA)
- {
- c->idct_put= ff_ea_idct_put_c;
- c->idct_permutation_type= FF_NO_IDCT_PERM;
- }*/
- else
- { //accurate/default
- c->idct_put= ff_simple_idct_put;
- c->idct_add= ff_simple_idct_add;
- c->idct = ff_simple_idct;
- c->idct_permutation_type= FF_NO_IDCT_PERM;
- }
- }
- if (ENABLE_H264_DECODER) {
- c->h264_idct_add= ff_h264_idct_add_c;
- c->h264_idct8_add= ff_h264_idct8_add_c;
- c->h264_idct_dc_add= ff_h264_idct_dc_add_c;
- c->h264_idct8_dc_add= ff_h264_idct8_dc_add_c;
- c->h264_idct_add16 = ff_h264_idct_add16_c;
- c->h264_idct8_add4 = ff_h264_idct8_add4_c;
- c->h264_idct_add8 = ff_h264_idct_add8_c;
- c->h264_idct_add16intra= ff_h264_idct_add16intra_c;
- }
- c->get_pixels = get_pixels_c;
- c->diff_pixels = diff_pixels_c;
- c->put_pixels_clamped = put_pixels_clamped_c;
- c->put_signed_pixels_clamped = put_signed_pixels_clamped_c;
- c->add_pixels_clamped = add_pixels_clamped_c;
- c->add_pixels8 = add_pixels8_c;
- c->add_pixels4 = add_pixels4_c;
- c->sum_abs_dctelem = sum_abs_dctelem_c;
- c->gmc1 = gmc1_c;
- c->gmc = ff_gmc_c;
- c->clear_block = clear_block_c;
- c->clear_blocks = clear_blocks_c;
- c->pix_sum = pix_sum_c;
- c->pix_norm1 = pix_norm1_c;
- /* TODO [0] 16 [1] 8 */
- c->pix_abs[0][0] = pix_abs16_c;
- c->pix_abs[0][1] = pix_abs16_x2_c;
- c->pix_abs[0][2] = pix_abs16_y2_c;
- c->pix_abs[0][3] = pix_abs16_xy2_c;
- c->pix_abs[1][0] = pix_abs8_c;
- c->pix_abs[1][1] = pix_abs8_x2_c;
- c->pix_abs[1][2] = pix_abs8_y2_c;
- c->pix_abs[1][3] = pix_abs8_xy2_c;
- #define dspfunc(PFX, IDX, NUM)
- c->PFX ## _pixels_tab[IDX][0] = PFX ## _pixels ## NUM ## _c;
- c->PFX ## _pixels_tab[IDX][1] = PFX ## _pixels ## NUM ## _x2_c;
- c->PFX ## _pixels_tab[IDX][2] = PFX ## _pixels ## NUM ## _y2_c;
- c->PFX ## _pixels_tab[IDX][3] = PFX ## _pixels ## NUM ## _xy2_c
- dspfunc(put, 0, 16);
- dspfunc(put_no_rnd, 0, 16);
- dspfunc(put, 1, 8);
- dspfunc(put_no_rnd, 1, 8);
- dspfunc(put, 2, 4);
- dspfunc(put, 3, 2);
- dspfunc(avg, 0, 16);
- dspfunc(avg_no_rnd, 0, 16);
- dspfunc(avg, 1, 8);
- dspfunc(avg_no_rnd, 1, 8);
- dspfunc(avg, 2, 4);
- dspfunc(avg, 3, 2);
- #undef dspfunc
- c->put_no_rnd_pixels_l2[0]= put_no_rnd_pixels16_l2_c;
- c->put_no_rnd_pixels_l2[1]= put_no_rnd_pixels8_l2_c;
- c->put_tpel_pixels_tab[ 0] = put_tpel_pixels_mc00_c;
- c->put_tpel_pixels_tab[ 1] = put_tpel_pixels_mc10_c;
- c->put_tpel_pixels_tab[ 2] = put_tpel_pixels_mc20_c;
- c->put_tpel_pixels_tab[ 4] = put_tpel_pixels_mc01_c;
- c->put_tpel_pixels_tab[ 5] = put_tpel_pixels_mc11_c;
- c->put_tpel_pixels_tab[ 6] = put_tpel_pixels_mc21_c;
- c->put_tpel_pixels_tab[ 8] = put_tpel_pixels_mc02_c;
- c->put_tpel_pixels_tab[ 9] = put_tpel_pixels_mc12_c;
- c->put_tpel_pixels_tab[10] = put_tpel_pixels_mc22_c;
- c->avg_tpel_pixels_tab[ 0] = avg_tpel_pixels_mc00_c;
- c->avg_tpel_pixels_tab[ 1] = avg_tpel_pixels_mc10_c;
- c->avg_tpel_pixels_tab[ 2] = avg_tpel_pixels_mc20_c;
- c->avg_tpel_pixels_tab[ 4] = avg_tpel_pixels_mc01_c;
- c->avg_tpel_pixels_tab[ 5] = avg_tpel_pixels_mc11_c;
- c->avg_tpel_pixels_tab[ 6] = avg_tpel_pixels_mc21_c;
- c->avg_tpel_pixels_tab[ 8] = avg_tpel_pixels_mc02_c;
- c->avg_tpel_pixels_tab[ 9] = avg_tpel_pixels_mc12_c;
- c->avg_tpel_pixels_tab[10] = avg_tpel_pixels_mc22_c;
- #define dspfunc(PFX, IDX, NUM)
- c->PFX ## _pixels_tab[IDX][ 0] = PFX ## NUM ## _mc00_c;
- c->PFX ## _pixels_tab[IDX][ 1] = PFX ## NUM ## _mc10_c;
- c->PFX ## _pixels_tab[IDX][ 2] = PFX ## NUM ## _mc20_c;
- c->PFX ## _pixels_tab[IDX][ 3] = PFX ## NUM ## _mc30_c;
- c->PFX ## _pixels_tab[IDX][ 4] = PFX ## NUM ## _mc01_c;
- c->PFX ## _pixels_tab[IDX][ 5] = PFX ## NUM ## _mc11_c;
- c->PFX ## _pixels_tab[IDX][ 6] = PFX ## NUM ## _mc21_c;
- c->PFX ## _pixels_tab[IDX][ 7] = PFX ## NUM ## _mc31_c;
- c->PFX ## _pixels_tab[IDX][ 8] = PFX ## NUM ## _mc02_c;
- c->PFX ## _pixels_tab[IDX][ 9] = PFX ## NUM ## _mc12_c;
- c->PFX ## _pixels_tab[IDX][10] = PFX ## NUM ## _mc22_c;
- c->PFX ## _pixels_tab[IDX][11] = PFX ## NUM ## _mc32_c;
- c->PFX ## _pixels_tab[IDX][12] = PFX ## NUM ## _mc03_c;
- c->PFX ## _pixels_tab[IDX][13] = PFX ## NUM ## _mc13_c;
- c->PFX ## _pixels_tab[IDX][14] = PFX ## NUM ## _mc23_c;
- c->PFX ## _pixels_tab[IDX][15] = PFX ## NUM ## _mc33_c
- dspfunc(put_qpel, 0, 16);
- dspfunc(put_no_rnd_qpel, 0, 16);
- dspfunc(avg_qpel, 0, 16);
- /* dspfunc(avg_no_rnd_qpel, 0, 16); */
- dspfunc(put_qpel, 1, 8);
- dspfunc(put_no_rnd_qpel, 1, 8);
- dspfunc(avg_qpel, 1, 8);
- /* dspfunc(avg_no_rnd_qpel, 1, 8); */
- dspfunc(put_h264_qpel, 0, 16);
- dspfunc(put_h264_qpel, 1, 8);
- dspfunc(put_h264_qpel, 2, 4);
- dspfunc(put_h264_qpel, 3, 2);
- dspfunc(avg_h264_qpel, 0, 16);
- dspfunc(avg_h264_qpel, 1, 8);
- dspfunc(avg_h264_qpel, 2, 4);
- #undef dspfunc
- c->put_h264_chroma_pixels_tab[0]= put_h264_chroma_mc8_c;
- c->put_h264_chroma_pixels_tab[1]= put_h264_chroma_mc4_c;
- c->put_h264_chroma_pixels_tab[2]= put_h264_chroma_mc2_c;
- c->avg_h264_chroma_pixels_tab[0]= avg_h264_chroma_mc8_c;
- c->avg_h264_chroma_pixels_tab[1]= avg_h264_chroma_mc4_c;
- c->avg_h264_chroma_pixels_tab[2]= avg_h264_chroma_mc2_c;
- c->put_no_rnd_h264_chroma_pixels_tab[0]= put_no_rnd_h264_chroma_mc8_c;
- c->weight_h264_pixels_tab[0]= weight_h264_pixels16x16_c;
- c->weight_h264_pixels_tab[1]= weight_h264_pixels16x8_c;
- c->weight_h264_pixels_tab[2]= weight_h264_pixels8x16_c;
- c->weight_h264_pixels_tab[3]= weight_h264_pixels8x8_c;
- c->weight_h264_pixels_tab[4]= weight_h264_pixels8x4_c;
- c->weight_h264_pixels_tab[5]= weight_h264_pixels4x8_c;
- c->weight_h264_pixels_tab[6]= weight_h264_pixels4x4_c;
- c->weight_h264_pixels_tab[7]= weight_h264_pixels4x2_c;
- c->weight_h264_pixels_tab[8]= weight_h264_pixels2x4_c;
- c->weight_h264_pixels_tab[9]= weight_h264_pixels2x2_c;
- c->biweight_h264_pixels_tab[0]= biweight_h264_pixels16x16_c;
- c->biweight_h264_pixels_tab[1]= biweight_h264_pixels16x8_c;
- c->biweight_h264_pixels_tab[2]= biweight_h264_pixels8x16_c;
- c->biweight_h264_pixels_tab[3]= biweight_h264_pixels8x8_c;
- c->biweight_h264_pixels_tab[4]= biweight_h264_pixels8x4_c;
- c->biweight_h264_pixels_tab[5]= biweight_h264_pixels4x8_c;
- c->biweight_h264_pixels_tab[6]= biweight_h264_pixels4x4_c;
- c->biweight_h264_pixels_tab[7]= biweight_h264_pixels4x2_c;
- c->biweight_h264_pixels_tab[8]= biweight_h264_pixels2x4_c;
- c->biweight_h264_pixels_tab[9]= biweight_h264_pixels2x2_c;
- c->draw_edges = draw_edges_c;
- #ifdef CONFIG_CAVS_DECODER
- ff_cavsdsp_init(c,avctx);
- #endif
- #if defined(CONFIG_VC1_DECODER) || defined(CONFIG_WMV3_DECODER)
- ff_vc1dsp_init(c,avctx);
- #endif
- #if defined(CONFIG_WMV2_DECODER) || defined(CONFIG_VC1_DECODER) || defined(CONFIG_WMV3_DECODER)
- ff_intrax8dsp_init(c,avctx);
- #endif
- #if defined(CONFIG_H264_ENCODER)
- ff_h264dspenc_init(c,avctx);
- #endif
- #if defined(CONFIG_RV30_DECODER)
- ff_rv30dsp_init(c,avctx);
- #endif
- #if defined(CONFIG_RV40_DECODER)
- ff_rv40dsp_init(c,avctx);
- c->put_rv40_qpel_pixels_tab[0][15] = put_rv40_qpel16_mc33_c;
- c->avg_rv40_qpel_pixels_tab[0][15] = avg_rv40_qpel16_mc33_c;
- c->put_rv40_qpel_pixels_tab[1][15] = put_rv40_qpel8_mc33_c;
- c->avg_rv40_qpel_pixels_tab[1][15] = avg_rv40_qpel8_mc33_c;
- #endif
- c->put_mspel_pixels_tab[0]= put_mspel8_mc00_c;
- c->put_mspel_pixels_tab[1]= put_mspel8_mc10_c;
- c->put_mspel_pixels_tab[2]= put_mspel8_mc20_c;
- c->put_mspel_pixels_tab[3]= put_mspel8_mc30_c;
- c->put_mspel_pixels_tab[4]= put_mspel8_mc02_c;
- c->put_mspel_pixels_tab[5]= put_mspel8_mc12_c;
- c->put_mspel_pixels_tab[6]= put_mspel8_mc22_c;
- c->put_mspel_pixels_tab[7]= put_mspel8_mc32_c;
- #define SET_CMP_FUNC(name)
- c->name[0]= name ## 16_c;
- c->name[1]= name ## 8x8_c;
- SET_CMP_FUNC(hadamard8_diff)
- c->hadamard8_diff[4]= hadamard8_intra16_c;
- SET_CMP_FUNC(dct_sad)
- SET_CMP_FUNC(dct_max)
- #ifdef CONFIG_GPL
- SET_CMP_FUNC(dct264_sad)
- #endif
- c->sad[0]= pix_abs16_c;
- c->sad[1]= pix_abs8_c;
- c->sse[0]= sse16_c;
- c->sse[1]= sse8_c;
- c->sse[2]= sse4_c;
- SET_CMP_FUNC(quant_psnr)
- SET_CMP_FUNC(rd)
- SET_CMP_FUNC(bit)
- c->vsad[0]= vsad16_c;
- c->vsad[4]= vsad_intra16_c;
- c->vsse[0]= vsse16_c;
- c->vsse[4]= vsse_intra16_c;
- c->nsse[0]= nsse16_c;
- c->nsse[1]= nsse8_c;
- #ifdef CONFIG_SNOW_ENCODER
- c->w53[0]= w53_16_c;
- c->w53[1]= w53_8_c;
- c->w97[0]= w97_16_c;
- c->w97[1]= w97_8_c;
- #endif
- c->ssd_int8_vs_int16 = ssd_int8_vs_int16_c;
- c->add_bytes= add_bytes_c;
- c->add_bytes_l2= add_bytes_l2_c;
- c->diff_bytes= diff_bytes_c;
- c->sub_hfyu_median_prediction= sub_hfyu_median_prediction_c;
- c->bswap_buf= bswap_buf;
- #ifdef CONFIG_PNG_DECODER
- c->add_png_paeth_prediction= ff_add_png_paeth_prediction;
- #endif
- c->h264_v_loop_filter_luma= h264_v_loop_filter_luma_c;
- c->h264_h_loop_filter_luma= h264_h_loop_filter_luma_c;
- c->h264_v_loop_filter_luma_intra= h264_v_loop_filter_luma_intra_c;
- c->h264_h_loop_filter_luma_intra= h264_h_loop_filter_luma_intra_c;
- c->h264_v_loop_filter_chroma= h264_v_loop_filter_chroma_c;
- c->h264_h_loop_filter_chroma= h264_h_loop_filter_chroma_c;
- c->h264_v_loop_filter_chroma_intra= h264_v_loop_filter_chroma_intra_c;
- c->h264_h_loop_filter_chroma_intra= h264_h_loop_filter_chroma_intra_c;
- c->h264_loop_filter_strength= NULL;
- /*if (ENABLE_ANY_H263) {
- c->h263_h_loop_filter= h263_h_loop_filter_c;
- c->h263_v_loop_filter= h263_v_loop_filter_c;
- }
- if (ENABLE_VP3_DECODER || ENABLE_THEORA_DECODER) {
- c->vp3_h_loop_filter= ff_vp3_h_loop_filter_c;
- c->vp3_v_loop_filter= ff_vp3_v_loop_filter_c;
- }*/
- c->h261_loop_filter= h261_loop_filter_c;
- c->try_8x8basis= try_8x8basis_c;
- c->add_8x8basis= add_8x8basis_c;
- #ifdef CONFIG_SNOW_DECODER
- c->vertical_compose97i = ff_snow_vertical_compose97i;
- c->horizontal_compose97i = ff_snow_horizontal_compose97i;
- c->inner_add_yblock = ff_snow_inner_add_yblock;
- #endif
- #ifdef CONFIG_VORBIS_DECODER
- c->vorbis_inverse_coupling = vorbis_inverse_coupling;
- #endif
- #ifdef CONFIG_AC3_DECODER
- c->ac3_downmix = ff_ac3_downmix_c;
- #endif
- #ifdef CONFIG_FLAC_ENCODER
- c->flac_compute_autocorr = ff_flac_compute_autocorr;
- #endif
- c->vector_fmul = vector_fmul_c;
- c->vector_fmul_reverse = vector_fmul_reverse_c;
- c->vector_fmul_add_add = ff_vector_fmul_add_add_c;
- c->vector_fmul_window = ff_vector_fmul_window_c;
- c->int32_to_float_fmul_scalar = int32_to_float_fmul_scalar_c;
- c->float_to_int16 = ff_float_to_int16_c;
- c->float_to_int16_interleave = ff_float_to_int16_interleave_c;
- c->add_int16 = add_int16_c;
- c->sub_int16 = sub_int16_c;
- c->scalarproduct_int16 = scalarproduct_int16_c;
- c->shrink[0]= ff_img_copy_plane;
- c->shrink[1]= ff_shrink22;
- c->shrink[2]= ff_shrink44;
- c->shrink[3]= ff_shrink88;
- c->prefetch= just_return;
- memset(c->put_2tap_qpel_pixels_tab, 0, sizeof(c->put_2tap_qpel_pixels_tab));
- memset(c->avg_2tap_qpel_pixels_tab, 0, sizeof(c->avg_2tap_qpel_pixels_tab));
- /*if (ENABLE_MMX) dsputil_init_mmx (c, avctx);
- if (ENABLE_ARM) dsputil_init_arm (c, avctx);
- if (ENABLE_MLIB) dsputil_init_mlib (c, avctx);
- if (ENABLE_VIS) dsputil_init_vis (c, avctx);
- if (ENABLE_ALPHA) dsputil_init_alpha (c, avctx);
- if (ENABLE_PPC) dsputil_init_ppc (c, avctx);
- if (ENABLE_MMI) dsputil_init_mmi (c, avctx);
- if (ENABLE_SH4) dsputil_init_sh4 (c, avctx);
- if (ENABLE_BFIN) dsputil_init_bfin (c, avctx);*/
- for(i=0; i<64; i++){
- if(!c->put_2tap_qpel_pixels_tab[0][i])
- c->put_2tap_qpel_pixels_tab[0][i]= c->put_h264_qpel_pixels_tab[0][i];
- if(!c->avg_2tap_qpel_pixels_tab[0][i])
- c->avg_2tap_qpel_pixels_tab[0][i]= c->avg_h264_qpel_pixels_tab[0][i];
- }
- switch(c->idct_permutation_type){
- case FF_NO_IDCT_PERM:
- for(i=0; i<64; i++)
- c->idct_permutation[i]= i;
- break;
- case FF_LIBMPEG2_IDCT_PERM:
- for(i=0; i<64; i++)
- c->idct_permutation[i]= (i & 0x38) | ((i & 6) >> 1) | ((i & 1) << 2);
- break;
- case FF_SIMPLE_IDCT_PERM:
- for(i=0; i<64; i++)
- c->idct_permutation[i]= simple_mmx_permutation[i];
- break;
- case FF_TRANSPOSE_IDCT_PERM:
- for(i=0; i<64; i++)
- c->idct_permutation[i]= ((i&7)<<3) | (i>>3);
- break;
- case FF_PARTTRANS_IDCT_PERM:
- for(i=0; i<64; i++)
- c->idct_permutation[i]= (i&0x24) | ((i&3)<<3) | ((i>>3)&3);
- break;
- case FF_SSE2_IDCT_PERM:
- for(i=0; i<64; i++)
- c->idct_permutation[i]= (i&0x38) | idct_sse2_row_perm[i&7];
- break;
- default:
- av_log(avctx, AV_LOG_ERROR, "Internal error, IDCT permutation not setn");
- }
- }