H264 decoder & demuxer

Originally committed as revision 1732 to svn://svn.ffmpeg.org/ffmpeg/trunk

1 files changed, 502 insertions, 0 deletions

diff --git a/libavcodec/dsputil.c b/libavcodec/dsputil.c
index 5f4190f755..146a657dd6 100644
--- a/libavcodec/dsputil.c
+++ b/libavcodec/dsputil.c
@@ -466,6 +466,14 @@ CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_xy2_c, OPNAME ## _no_rnd_pixels_xy2_c,
 #else // 64 bit variant
 
 #define PIXOP2(OPNAME, OP) \
+static void OPNAME ## _pixels4_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
+    int i;\
+    for(i=0; i<h; i++){\
+        OP(*((uint32_t*)(block  )), LD32(pixels  ));\
+        pixels+=line_size;\
+        block +=line_size;\
+    }\
+}\
 static void OPNAME ## _pixels8_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
     int i;\
     for(i=0; i<h; i++){\
@@ -507,6 +515,17 @@ static inline void OPNAME ## _pixels8_l2(uint8_t *dst, const uint8_t *src1, cons
     }\
 }\
 \
+static inline void OPNAME ## _pixels4_l2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
+                                                int src_stride1, int src_stride2, int h){\
+    int i;\
+    for(i=0; i<h; i++){\
+        uint32_t a,b;\
+        a= LD32(&src1[i*src_stride1  ]);\
+        b= LD32(&src2[i*src_stride2  ]);\
+        OP(*((uint32_t*)&dst[i*dst_stride  ]), (a|b) - (((a^b)&0xFEFEFEFEUL)>>1));\
+    }\
+}\
+\
 static inline void OPNAME ## _pixels16_l2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
                                                 int src_stride1, int src_stride2, int h){\
     OPNAME ## _pixels8_l2(dst  , src1  , src2  , dst_stride, src_stride1, src_stride2, h);\
@@ -800,6 +819,113 @@ static void gmc_c(uint8_t *dst, uint8_t *src, int stride, int h, int ox, int oy,
         oy += dyy;
     }
 }
+#define H264_CHROMA_MC(OPNAME, OP)\
+static void OPNAME ## h264_chroma_mc2_c(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int stride, int h, int x, int y){\
+    const int A=(8-x)*(8-y);\
+    const int B=(  x)*(8-y);\
+    const int C=(8-x)*(  y);\
+    const int D=(  x)*(  y);\
+    int i;\
+    \
+    assert(x<8 && y<8 && x>=0 && y>=0);\
+\
+    for(i=0; i<h; i++)\
+    {\
+        OP(dst[0], (A*src[0] + B*src[1] + C*src[stride+0] + D*src[stride+1]));\
+        OP(dst[1], (A*src[1] + B*src[2] + C*src[stride+1] + D*src[stride+2]));\
+        dst+= stride;\
+        src+= stride;\
+    }\
+}\
+\
+static void OPNAME ## h264_chroma_mc4_c(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int stride, int h, int x, int y){\
+    const int A=(8-x)*(8-y);\
+    const int B=(  x)*(8-y);\
+    const int C=(8-x)*(  y);\
+    const int D=(  x)*(  y);\
+    int i;\
+    \
+    assert(x<8 && y<8 && x>=0 && y>=0);\
+\
+    for(i=0; i<h; i++)\
+    {\
+        OP(dst[0], (A*src[0] + B*src[1] + C*src[stride+0] + D*src[stride+1]));\
+        OP(dst[1], (A*src[1] + B*src[2] + C*src[stride+1] + D*src[stride+2]));\
+        OP(dst[2], (A*src[2] + B*src[3] + C*src[stride+2] + D*src[stride+3]));\
+        OP(dst[3], (A*src[3] + B*src[4] + C*src[stride+3] + D*src[stride+4]));\
+        dst+= stride;\
+        src+= stride;\
+    }\
+}\
+\
+static void OPNAME ## h264_chroma_mc8_c(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int stride, int h, int x, int y){\
+    const int A=(8-x)*(8-y);\
+    const int B=(  x)*(8-y);\
+    const int C=(8-x)*(  y);\
+    const int D=(  x)*(  y);\
+    int i;\
+    \
+    assert(x<8 && y<8 && x>=0 && y>=0);\
+\
+    for(i=0; i<h; i++)\
+    {\
+        OP(dst[0], (A*src[0] + B*src[1] + C*src[stride+0] + D*src[stride+1]));\
+        OP(dst[1], (A*src[1] + B*src[2] + C*src[stride+1] + D*src[stride+2]));\
+        OP(dst[2], (A*src[2] + B*src[3] + C*src[stride+2] + D*src[stride+3]));\
+        OP(dst[3], (A*src[3] + B*src[4] + C*src[stride+3] + D*src[stride+4]));\
+        OP(dst[4], (A*src[4] + B*src[5] + C*src[stride+4] + D*src[stride+5]));\
+        OP(dst[5], (A*src[5] + B*src[6] + C*src[stride+5] + D*src[stride+6]));\
+        OP(dst[6], (A*src[6] + B*src[7] + C*src[stride+6] + D*src[stride+7]));\
+        OP(dst[7], (A*src[7] + B*src[8] + C*src[stride+7] + D*src[stride+8]));\
+        dst+= stride;\
+        src+= stride;\
+    }\
+}
+
+#define op_avg(a, b) a = (((a)+(((b) + 32)>>6)+1)>>1)
+#define op_put(a, b) a = (((b) + 32)>>6)
+
+H264_CHROMA_MC(put_       , op_put)
+H264_CHROMA_MC(avg_       , op_avg)
+#undef op_avg
+#undef op_put
+
+static inline void copy_block4(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
+{
+    int i;
+    for(i=0; i<h; i++)
+    {
+        ST32(dst   , LD32(src   ));
+        dst+=dstStride;
+        src+=srcStride;
+    }
+}
+
+static inline void copy_block8(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
+{
+    int i;
+    for(i=0; i<h; i++)
+    {
+        ST32(dst   , LD32(src   ));
+        ST32(dst+4 , LD32(src+4 ));
+        dst+=dstStride;
+        src+=srcStride;
+    }
+}
+
+static inline void copy_block16(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
+{
+    int i;
+    for(i=0; i<h; i++)
+    {
+        ST32(dst   , LD32(src   ));
+        ST32(dst+4 , LD32(src+4 ));
+        ST32(dst+8 , LD32(src+8 ));
+        ST32(dst+12, LD32(src+12));
+        dst+=dstStride;
+        src+=srcStride;
+    }
+}
 
 static inline void copy_block17(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
 {
@@ -1327,6 +1453,368 @@ QPEL_MC(0, avg_       , _       , op_avg)
 #undef op_put
 #undef op_put_no_rnd
 
+#if 1
+#define H264_LOWPASS(OPNAME, OP, OP2) \
+static void OPNAME ## h264_qpel4_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+    const int h=4;\
+    uint8_t *cm = 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]));\
+        OP(dst[1], (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]));\
+        OP(dst[2], (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5]));\
+        OP(dst[3], (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6]));\
+        dst+=dstStride;\
+        src+=srcStride;\
+    }\
+}\
+\
+static void OPNAME ## h264_qpel4_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+    const int w=4;\
+    uint8_t *cm = 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];\
+        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));\
+        dst++;\
+        src++;\
+    }\
+}\
+\
+static void OPNAME ## h264_qpel4_hv_lowpass(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){\
+    const int h=4;\
+    const int w=4;\
+    uint8_t *cm = 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+=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];\
+        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));\
+        dst++;\
+        tmp++;\
+    }\
+}\
+\
+static void OPNAME ## h264_qpel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+    const int h=8;\
+    uint8_t *cm = 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 ]));\
+        OP(dst[1], (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4 ]));\
+        OP(dst[2], (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5 ]));\
+        OP(dst[3], (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6 ]));\
+        OP(dst[4], (src[4]+src[5])*20 - (src[3 ]+src[6])*5 + (src[2 ]+src[7 ]));\
+        OP(dst[5], (src[5]+src[6])*20 - (src[4 ]+src[7])*5 + (src[3 ]+src[8 ]));\
+        OP(dst[6], (src[6]+src[7])*20 - (src[5 ]+src[8])*5 + (src[4 ]+src[9 ]));\
+        OP(dst[7], (src[7]+src[8])*20 - (src[6 ]+src[9])*5 + (src[5 ]+src[10]));\
+        dst+=dstStride;\
+        src+=srcStride;\
+    }\
+}\
+\
+static void OPNAME ## h264_qpel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+    const int w=8;\
+    uint8_t *cm = 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 = 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;\
+    tmp += 8*tmpStride;\
+    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 op_avg2(a, b) a = (((a)*w1+cm[((b) + 16)>>5]*w2 + o + 64)>>7)
+#define op_put(a, b)  a = cm[((b) + 16)>>5]
+#define op2_avg(a, b)  a = (((a)+cm[((b) + 512)>>10]+1)>>1)
+#define op2_put(a, b)  a = cm[((b) + 512)>>10]
+
+H264_LOWPASS(put_       , op_put, op2_put)
+H264_LOWPASS(avg_       , op_avg, op2_avg)
+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
+
 static void wmv2_mspel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){
     uint8_t *cm = cropTbl + MAX_NEG_CROP;
     int i;
@@ -2107,7 +2595,21 @@ void dsputil_init(DSPContext* c, AVCodecContext *avctx)
 
     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(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_mspel_pixels_tab[0]= put_mspel8_mc00_c;
     c->put_mspel_pixels_tab[1]= put_mspel8_mc10_c;