Mercurial > libavcodec.hg

comparison h264.c @ 5638:4a26dc4ca11d libavcodec

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Move H.264 intra prediction functions into their own context
author kostya
date Wed, 05 Sep 2007 05:30:08 +0000
parents 550e37ec2204
children d2598034f2a9
comparison
equal deleted inserted replaced
5637:473cada682a1 5638:4a26dc4ca11d
1679 } 1679 }
1680 1680
1681 return last_non_zero; 1681 return last_non_zero;
1682 } 1682 }
1683 1683
1684 static void pred4x4_vertical_c(uint8_t *src, uint8_t *topright, int stride){
1685 const uint32_t a= ((uint32_t*)(src-stride))[0];
1686 ((uint32_t*)(src+0*stride))[0]= a;
1687 ((uint32_t*)(src+1*stride))[0]= a;
1688 ((uint32_t*)(src+2*stride))[0]= a;
1689 ((uint32_t*)(src+3*stride))[0]= a;
1690 }
1691
1692 static void pred4x4_horizontal_c(uint8_t *src, uint8_t *topright, int stride){
1693 ((uint32_t*)(src+0*stride))[0]= src[-1+0*stride]*0x01010101;
1694 ((uint32_t*)(src+1*stride))[0]= src[-1+1*stride]*0x01010101;
1695 ((uint32_t*)(src+2*stride))[0]= src[-1+2*stride]*0x01010101;
1696 ((uint32_t*)(src+3*stride))[0]= src[-1+3*stride]*0x01010101;
1697 }
1698
1699 static void pred4x4_dc_c(uint8_t *src, uint8_t *topright, int stride){
1700 const int dc= ( src[-stride] + src[1-stride] + src[2-stride] + src[3-stride]
1701 + src[-1+0*stride] + src[-1+1*stride] + src[-1+2*stride] + src[-1+3*stride] + 4) >>3;
1702
1703 ((uint32_t*)(src+0*stride))[0]=
1704 ((uint32_t*)(src+1*stride))[0]=
1705 ((uint32_t*)(src+2*stride))[0]=
1706 ((uint32_t*)(src+3*stride))[0]= dc* 0x01010101;
1707 }
1708
1709 static void pred4x4_left_dc_c(uint8_t *src, uint8_t *topright, int stride){
1710 const int dc= ( src[-1+0*stride] + src[-1+1*stride] + src[-1+2*stride] + src[-1+3*stride] + 2) >>2;
1711
1712 ((uint32_t*)(src+0*stride))[0]=
1713 ((uint32_t*)(src+1*stride))[0]=
1714 ((uint32_t*)(src+2*stride))[0]=
1715 ((uint32_t*)(src+3*stride))[0]= dc* 0x01010101;
1716 }
1717
1718 static void pred4x4_top_dc_c(uint8_t *src, uint8_t *topright, int stride){
1719 const int dc= ( src[-stride] + src[1-stride] + src[2-stride] + src[3-stride] + 2) >>2;
1720
1721 ((uint32_t*)(src+0*stride))[0]=
1722 ((uint32_t*)(src+1*stride))[0]=
1723 ((uint32_t*)(src+2*stride))[0]=
1724 ((uint32_t*)(src+3*stride))[0]= dc* 0x01010101;
1725 }
1726
1727 static void pred4x4_128_dc_c(uint8_t *src, uint8_t *topright, int stride){
1728 ((uint32_t*)(src+0*stride))[0]=
1729 ((uint32_t*)(src+1*stride))[0]=
1730 ((uint32_t*)(src+2*stride))[0]=
1731 ((uint32_t*)(src+3*stride))[0]= 128U*0x01010101U;
1732 }
1733
1734
1735 #define LOAD_TOP_RIGHT_EDGE\
1736 const int av_unused t4= topright[0];\
1737 const int av_unused t5= topright[1];\
1738 const int av_unused t6= topright[2];\
1739 const int av_unused t7= topright[3];\
1740
1741 #define LOAD_LEFT_EDGE\
1742 const int av_unused l0= src[-1+0*stride];\
1743 const int av_unused l1= src[-1+1*stride];\
1744 const int av_unused l2= src[-1+2*stride];\
1745 const int av_unused l3= src[-1+3*stride];\
1746
1747 #define LOAD_TOP_EDGE\
1748 const int av_unused t0= src[ 0-1*stride];\
1749 const int av_unused t1= src[ 1-1*stride];\
1750 const int av_unused t2= src[ 2-1*stride];\
1751 const int av_unused t3= src[ 3-1*stride];\
1752
1753 static void pred4x4_down_right_c(uint8_t *src, uint8_t *topright, int stride){
1754 const int lt= src[-1-1*stride];
1755 LOAD_TOP_EDGE
1756 LOAD_LEFT_EDGE
1757
1758 src[0+3*stride]=(l3 + 2*l2 + l1 + 2)>>2;
1759 src[0+2*stride]=
1760 src[1+3*stride]=(l2 + 2*l1 + l0 + 2)>>2;
1761 src[0+1*stride]=
1762 src[1+2*stride]=
1763 src[2+3*stride]=(l1 + 2*l0 + lt + 2)>>2;
1764 src[0+0*stride]=
1765 src[1+1*stride]=
1766 src[2+2*stride]=
1767 src[3+3*stride]=(l0 + 2*lt + t0 + 2)>>2;
1768 src[1+0*stride]=
1769 src[2+1*stride]=
1770 src[3+2*stride]=(lt + 2*t0 + t1 + 2)>>2;
1771 src[2+0*stride]=
1772 src[3+1*stride]=(t0 + 2*t1 + t2 + 2)>>2;
1773 src[3+0*stride]=(t1 + 2*t2 + t3 + 2)>>2;
1774 }
1775
1776 static void pred4x4_down_left_c(uint8_t *src, uint8_t *topright, int stride){
1777 LOAD_TOP_EDGE
1778 LOAD_TOP_RIGHT_EDGE
1779 // LOAD_LEFT_EDGE
1780
1781 src[0+0*stride]=(t0 + t2 + 2*t1 + 2)>>2;
1782 src[1+0*stride]=
1783 src[0+1*stride]=(t1 + t3 + 2*t2 + 2)>>2;
1784 src[2+0*stride]=
1785 src[1+1*stride]=
1786 src[0+2*stride]=(t2 + t4 + 2*t3 + 2)>>2;
1787 src[3+0*stride]=
1788 src[2+1*stride]=
1789 src[1+2*stride]=
1790 src[0+3*stride]=(t3 + t5 + 2*t4 + 2)>>2;
1791 src[3+1*stride]=
1792 src[2+2*stride]=
1793 src[1+3*stride]=(t4 + t6 + 2*t5 + 2)>>2;
1794 src[3+2*stride]=
1795 src[2+3*stride]=(t5 + t7 + 2*t6 + 2)>>2;
1796 src[3+3*stride]=(t6 + 3*t7 + 2)>>2;
1797 }
1798
1799 static void pred4x4_vertical_right_c(uint8_t *src, uint8_t *topright, int stride){
1800 const int lt= src[-1-1*stride];
1801 LOAD_TOP_EDGE
1802 LOAD_LEFT_EDGE
1803
1804 src[0+0*stride]=
1805 src[1+2*stride]=(lt + t0 + 1)>>1;
1806 src[1+0*stride]=
1807 src[2+2*stride]=(t0 + t1 + 1)>>1;
1808 src[2+0*stride]=
1809 src[3+2*stride]=(t1 + t2 + 1)>>1;
1810 src[3+0*stride]=(t2 + t3 + 1)>>1;
1811 src[0+1*stride]=
1812 src[1+3*stride]=(l0 + 2*lt + t0 + 2)>>2;
1813 src[1+1*stride]=
1814 src[2+3*stride]=(lt + 2*t0 + t1 + 2)>>2;
1815 src[2+1*stride]=
1816 src[3+3*stride]=(t0 + 2*t1 + t2 + 2)>>2;
1817 src[3+1*stride]=(t1 + 2*t2 + t3 + 2)>>2;
1818 src[0+2*stride]=(lt + 2*l0 + l1 + 2)>>2;
1819 src[0+3*stride]=(l0 + 2*l1 + l2 + 2)>>2;
1820 }
1821
1822 static void pred4x4_vertical_left_c(uint8_t *src, uint8_t *topright, int stride){
1823 LOAD_TOP_EDGE
1824 LOAD_TOP_RIGHT_EDGE
1825
1826 src[0+0*stride]=(t0 + t1 + 1)>>1;
1827 src[1+0*stride]=
1828 src[0+2*stride]=(t1 + t2 + 1)>>1;
1829 src[2+0*stride]=
1830 src[1+2*stride]=(t2 + t3 + 1)>>1;
1831 src[3+0*stride]=
1832 src[2+2*stride]=(t3 + t4+ 1)>>1;
1833 src[3+2*stride]=(t4 + t5+ 1)>>1;
1834 src[0+1*stride]=(t0 + 2*t1 + t2 + 2)>>2;
1835 src[1+1*stride]=
1836 src[0+3*stride]=(t1 + 2*t2 + t3 + 2)>>2;
1837 src[2+1*stride]=
1838 src[1+3*stride]=(t2 + 2*t3 + t4 + 2)>>2;
1839 src[3+1*stride]=
1840 src[2+3*stride]=(t3 + 2*t4 + t5 + 2)>>2;
1841 src[3+3*stride]=(t4 + 2*t5 + t6 + 2)>>2;
1842 }
1843
1844 static void pred4x4_horizontal_up_c(uint8_t *src, uint8_t *topright, int stride){
1845 LOAD_LEFT_EDGE
1846
1847 src[0+0*stride]=(l0 + l1 + 1)>>1;
1848 src[1+0*stride]=(l0 + 2*l1 + l2 + 2)>>2;
1849 src[2+0*stride]=
1850 src[0+1*stride]=(l1 + l2 + 1)>>1;
1851 src[3+0*stride]=
1852 src[1+1*stride]=(l1 + 2*l2 + l3 + 2)>>2;
1853 src[2+1*stride]=
1854 src[0+2*stride]=(l2 + l3 + 1)>>1;
1855 src[3+1*stride]=
1856 src[1+2*stride]=(l2 + 2*l3 + l3 + 2)>>2;
1857 src[3+2*stride]=
1858 src[1+3*stride]=
1859 src[0+3*stride]=
1860 src[2+2*stride]=
1861 src[2+3*stride]=
1862 src[3+3*stride]=l3;
1863 }
1864
1865 static void pred4x4_horizontal_down_c(uint8_t *src, uint8_t *topright, int stride){
1866 const int lt= src[-1-1*stride];
1867 LOAD_TOP_EDGE
1868 LOAD_LEFT_EDGE
1869
1870 src[0+0*stride]=
1871 src[2+1*stride]=(lt + l0 + 1)>>1;
1872 src[1+0*stride]=
1873 src[3+1*stride]=(l0 + 2*lt + t0 + 2)>>2;
1874 src[2+0*stride]=(lt + 2*t0 + t1 + 2)>>2;
1875 src[3+0*stride]=(t0 + 2*t1 + t2 + 2)>>2;
1876 src[0+1*stride]=
1877 src[2+2*stride]=(l0 + l1 + 1)>>1;
1878 src[1+1*stride]=
1879 src[3+2*stride]=(lt + 2*l0 + l1 + 2)>>2;
1880 src[0+2*stride]=
1881 src[2+3*stride]=(l1 + l2+ 1)>>1;
1882 src[1+2*stride]=
1883 src[3+3*stride]=(l0 + 2*l1 + l2 + 2)>>2;
1884 src[0+3*stride]=(l2 + l3 + 1)>>1;
1885 src[1+3*stride]=(l1 + 2*l2 + l3 + 2)>>2;
1886 }
1887
1888 void ff_pred16x16_vertical_c(uint8_t *src, int stride){
1889 int i;
1890 const uint32_t a= ((uint32_t*)(src-stride))[0];
1891 const uint32_t b= ((uint32_t*)(src-stride))[1];
1892 const uint32_t c= ((uint32_t*)(src-stride))[2];
1893 const uint32_t d= ((uint32_t*)(src-stride))[3];
1894
1895 for(i=0; i<16; i++){
1896 ((uint32_t*)(src+i*stride))[0]= a;
1897 ((uint32_t*)(src+i*stride))[1]= b;
1898 ((uint32_t*)(src+i*stride))[2]= c;
1899 ((uint32_t*)(src+i*stride))[3]= d;
1900 }
1901 }
1902
1903 void ff_pred16x16_horizontal_c(uint8_t *src, int stride){
1904 int i;
1905
1906 for(i=0; i<16; i++){
1907 ((uint32_t*)(src+i*stride))[0]=
1908 ((uint32_t*)(src+i*stride))[1]=
1909 ((uint32_t*)(src+i*stride))[2]=
1910 ((uint32_t*)(src+i*stride))[3]= src[-1+i*stride]*0x01010101;
1911 }
1912 }
1913
1914 void ff_pred16x16_dc_c(uint8_t *src, int stride){
1915 int i, dc=0;
1916
1917 for(i=0;i<16; i++){
1918 dc+= src[-1+i*stride];
1919 }
1920
1921 for(i=0;i<16; i++){
1922 dc+= src[i-stride];
1923 }
1924
1925 dc= 0x01010101*((dc + 16)>>5);
1926
1927 for(i=0; i<16; i++){
1928 ((uint32_t*)(src+i*stride))[0]=
1929 ((uint32_t*)(src+i*stride))[1]=
1930 ((uint32_t*)(src+i*stride))[2]=
1931 ((uint32_t*)(src+i*stride))[3]= dc;
1932 }
1933 }
1934
1935 void ff_pred16x16_left_dc_c(uint8_t *src, int stride){
1936 int i, dc=0;
1937
1938 for(i=0;i<16; i++){
1939 dc+= src[-1+i*stride];
1940 }
1941
1942 dc= 0x01010101*((dc + 8)>>4);
1943
1944 for(i=0; i<16; i++){
1945 ((uint32_t*)(src+i*stride))[0]=
1946 ((uint32_t*)(src+i*stride))[1]=
1947 ((uint32_t*)(src+i*stride))[2]=
1948 ((uint32_t*)(src+i*stride))[3]= dc;
1949 }
1950 }
1951
1952 void ff_pred16x16_top_dc_c(uint8_t *src, int stride){
1953 int i, dc=0;
1954
1955 for(i=0;i<16; i++){
1956 dc+= src[i-stride];
1957 }
1958 dc= 0x01010101*((dc + 8)>>4);
1959
1960 for(i=0; i<16; i++){
1961 ((uint32_t*)(src+i*stride))[0]=
1962 ((uint32_t*)(src+i*stride))[1]=
1963 ((uint32_t*)(src+i*stride))[2]=
1964 ((uint32_t*)(src+i*stride))[3]= dc;
1965 }
1966 }
1967
1968 void ff_pred16x16_128_dc_c(uint8_t *src, int stride){
1969 int i;
1970
1971 for(i=0; i<16; i++){
1972 ((uint32_t*)(src+i*stride))[0]=
1973 ((uint32_t*)(src+i*stride))[1]=
1974 ((uint32_t*)(src+i*stride))[2]=
1975 ((uint32_t*)(src+i*stride))[3]= 0x01010101U*128U;
1976 }
1977 }
1978
1979 static inline void pred16x16_plane_compat_c(uint8_t *src, int stride, const int svq3){
1980 int i, j, k;
1981 int a;
1982 uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
1983 const uint8_t * const src0 = src+7-stride;
1984 const uint8_t *src1 = src+8*stride-1;
1985 const uint8_t *src2 = src1-2*stride; // == src+6*stride-1;
1986 int H = src0[1] - src0[-1];
1987 int V = src1[0] - src2[ 0];
1988 for(k=2; k<=8; ++k) {
1989 src1 += stride; src2 -= stride;
1990 H += k*(src0[k] - src0[-k]);
1991 V += k*(src1[0] - src2[ 0]);
1992 }
1993 if(svq3){
1994 H = ( 5*(H/4) ) / 16;
1995 V = ( 5*(V/4) ) / 16;
1996
1997 /* required for 100% accuracy */
1998 i = H; H = V; V = i;
1999 }else{
2000 H = ( 5*H+32 ) >> 6;
2001 V = ( 5*V+32 ) >> 6;
2002 }
2003
2004 a = 16*(src1[0] + src2[16] + 1) - 7*(V+H);
2005 for(j=16; j>0; --j) {
2006 int b = a;
2007 a += V;
2008 for(i=-16; i<0; i+=4) {
2009 src[16+i] = cm[ (b ) >> 5 ];
2010 src[17+i] = cm[ (b+ H) >> 5 ];
2011 src[18+i] = cm[ (b+2*H) >> 5 ];
2012 src[19+i] = cm[ (b+3*H) >> 5 ];
2013 b += 4*H;
2014 }
2015 src += stride;
2016 }
2017 }
2018
2019 void ff_pred16x16_plane_c(uint8_t *src, int stride){
2020 pred16x16_plane_compat_c(src, stride, 0);
2021 }
2022
2023 void ff_pred8x8_vertical_c(uint8_t *src, int stride){
2024 int i;
2025 const uint32_t a= ((uint32_t*)(src-stride))[0];
2026 const uint32_t b= ((uint32_t*)(src-stride))[1];
2027
2028 for(i=0; i<8; i++){
2029 ((uint32_t*)(src+i*stride))[0]= a;
2030 ((uint32_t*)(src+i*stride))[1]= b;
2031 }
2032 }
2033
2034 void ff_pred8x8_horizontal_c(uint8_t *src, int stride){
2035 int i;
2036
2037 for(i=0; i<8; i++){
2038 ((uint32_t*)(src+i*stride))[0]=
2039 ((uint32_t*)(src+i*stride))[1]= src[-1+i*stride]*0x01010101;
2040 }
2041 }
2042
2043 void ff_pred8x8_128_dc_c(uint8_t *src, int stride){
2044 int i;
2045
2046 for(i=0; i<8; i++){
2047 ((uint32_t*)(src+i*stride))[0]=
2048 ((uint32_t*)(src+i*stride))[1]= 0x01010101U*128U;
2049 }
2050 }
2051
2052 void ff_pred8x8_left_dc_c(uint8_t *src, int stride){
2053 int i;
2054 int dc0, dc2;
2055
2056 dc0=dc2=0;
2057 for(i=0;i<4; i++){
2058 dc0+= src[-1+i*stride];
2059 dc2+= src[-1+(i+4)*stride];
2060 }
2061 dc0= 0x01010101*((dc0 + 2)>>2);
2062 dc2= 0x01010101*((dc2 + 2)>>2);
2063
2064 for(i=0; i<4; i++){
2065 ((uint32_t*)(src+i*stride))[0]=
2066 ((uint32_t*)(src+i*stride))[1]= dc0;
2067 }
2068 for(i=4; i<8; i++){
2069 ((uint32_t*)(src+i*stride))[0]=
2070 ((uint32_t*)(src+i*stride))[1]= dc2;
2071 }
2072 }
2073
2074 void ff_pred8x8_top_dc_c(uint8_t *src, int stride){
2075 int i;
2076 int dc0, dc1;
2077
2078 dc0=dc1=0;
2079 for(i=0;i<4; i++){
2080 dc0+= src[i-stride];
2081 dc1+= src[4+i-stride];
2082 }
2083 dc0= 0x01010101*((dc0 + 2)>>2);
2084 dc1= 0x01010101*((dc1 + 2)>>2);
2085
2086 for(i=0; i<4; i++){
2087 ((uint32_t*)(src+i*stride))[0]= dc0;
2088 ((uint32_t*)(src+i*stride))[1]= dc1;
2089 }
2090 for(i=4; i<8; i++){
2091 ((uint32_t*)(src+i*stride))[0]= dc0;
2092 ((uint32_t*)(src+i*stride))[1]= dc1;
2093 }
2094 }
2095
2096
2097 void ff_pred8x8_dc_c(uint8_t *src, int stride){
2098 int i;
2099 int dc0, dc1, dc2, dc3;
2100
2101 dc0=dc1=dc2=0;
2102 for(i=0;i<4; i++){
2103 dc0+= src[-1+i*stride] + src[i-stride];
2104 dc1+= src[4+i-stride];
2105 dc2+= src[-1+(i+4)*stride];
2106 }
2107 dc3= 0x01010101*((dc1 + dc2 + 4)>>3);
2108 dc0= 0x01010101*((dc0 + 4)>>3);
2109 dc1= 0x01010101*((dc1 + 2)>>2);
2110 dc2= 0x01010101*((dc2 + 2)>>2);
2111
2112 for(i=0; i<4; i++){
2113 ((uint32_t*)(src+i*stride))[0]= dc0;
2114 ((uint32_t*)(src+i*stride))[1]= dc1;
2115 }
2116 for(i=4; i<8; i++){
2117 ((uint32_t*)(src+i*stride))[0]= dc2;
2118 ((uint32_t*)(src+i*stride))[1]= dc3;
2119 }
2120 }
2121
2122 void ff_pred8x8_plane_c(uint8_t *src, int stride){
2123 int j, k;
2124 int a;
2125 uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
2126 const uint8_t * const src0 = src+3-stride;
2127 const uint8_t *src1 = src+4*stride-1;
2128 const uint8_t *src2 = src1-2*stride; // == src+2*stride-1;
2129 int H = src0[1] - src0[-1];
2130 int V = src1[0] - src2[ 0];
2131 for(k=2; k<=4; ++k) {
2132 src1 += stride; src2 -= stride;
2133 H += k*(src0[k] - src0[-k]);
2134 V += k*(src1[0] - src2[ 0]);
2135 }
2136 H = ( 17*H+16 ) >> 5;
2137 V = ( 17*V+16 ) >> 5;
2138
2139 a = 16*(src1[0] + src2[8]+1) - 3*(V+H);
2140 for(j=8; j>0; --j) {
2141 int b = a;
2142 a += V;
2143 src[0] = cm[ (b ) >> 5 ];
2144 src[1] = cm[ (b+ H) >> 5 ];
2145 src[2] = cm[ (b+2*H) >> 5 ];
2146 src[3] = cm[ (b+3*H) >> 5 ];
2147 src[4] = cm[ (b+4*H) >> 5 ];
2148 src[5] = cm[ (b+5*H) >> 5 ];
2149 src[6] = cm[ (b+6*H) >> 5 ];
2150 src[7] = cm[ (b+7*H) >> 5 ];
2151 src += stride;
2152 }
2153 }
2154
2155 #define SRC(x,y) src[(x)+(y)*stride]
2156 #define PL(y) \
2157 const int l##y = (SRC(-1,y-1) + 2*SRC(-1,y) + SRC(-1,y+1) + 2) >> 2;
2158 #define PREDICT_8x8_LOAD_LEFT \
2159 const int l0 = ((has_topleft ? SRC(-1,-1) : SRC(-1,0)) \
2160 + 2*SRC(-1,0) + SRC(-1,1) + 2) >> 2; \
2161 PL(1) PL(2) PL(3) PL(4) PL(5) PL(6) \
2162 const int l7 av_unused = (SRC(-1,6) + 3*SRC(-1,7) + 2) >> 2
2163
2164 #define PT(x) \
2165 const int t##x = (SRC(x-1,-1) + 2*SRC(x,-1) + SRC(x+1,-1) + 2) >> 2;
2166 #define PREDICT_8x8_LOAD_TOP \
2167 const int t0 = ((has_topleft ? SRC(-1,-1) : SRC(0,-1)) \
2168 + 2*SRC(0,-1) + SRC(1,-1) + 2) >> 2; \
2169 PT(1) PT(2) PT(3) PT(4) PT(5) PT(6) \
2170 const int t7 av_unused = ((has_topright ? SRC(8,-1) : SRC(7,-1)) \
2171 + 2*SRC(7,-1) + SRC(6,-1) + 2) >> 2
2172
2173 #define PTR(x) \
2174 t##x = (SRC(x-1,-1) + 2*SRC(x,-1) + SRC(x+1,-1) + 2) >> 2;
2175 #define PREDICT_8x8_LOAD_TOPRIGHT \
2176 int t8, t9, t10, t11, t12, t13, t14, t15; \
2177 if(has_topright) { \
2178 PTR(8) PTR(9) PTR(10) PTR(11) PTR(12) PTR(13) PTR(14) \
2179 t15 = (SRC(14,-1) + 3*SRC(15,-1) + 2) >> 2; \
2180 } else t8=t9=t10=t11=t12=t13=t14=t15= SRC(7,-1);
2181
2182 #define PREDICT_8x8_LOAD_TOPLEFT \
2183 const int lt = (SRC(-1,0) + 2*SRC(-1,-1) + SRC(0,-1) + 2) >> 2
2184
2185 #define PREDICT_8x8_DC(v) \
2186 int y; \
2187 for( y = 0; y < 8; y++ ) { \
2188 ((uint32_t*)src)[0] = \
2189 ((uint32_t*)src)[1] = v; \
2190 src += stride; \
2191 }
2192
2193 static void pred8x8l_128_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride)
2194 {
2195 PREDICT_8x8_DC(0x80808080);
2196 }
2197 static void pred8x8l_left_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride)
2198 {
2199 PREDICT_8x8_LOAD_LEFT;
2200 const uint32_t dc = ((l0+l1+l2+l3+l4+l5+l6+l7+4) >> 3) * 0x01010101;
2201 PREDICT_8x8_DC(dc);
2202 }
2203 static void pred8x8l_top_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride)
2204 {
2205 PREDICT_8x8_LOAD_TOP;
2206 const uint32_t dc = ((t0+t1+t2+t3+t4+t5+t6+t7+4) >> 3) * 0x01010101;
2207 PREDICT_8x8_DC(dc);
2208 }
2209 static void pred8x8l_dc_c(uint8_t *src, int has_topleft, int has_topright, int stride)
2210 {
2211 PREDICT_8x8_LOAD_LEFT;
2212 PREDICT_8x8_LOAD_TOP;
2213 const uint32_t dc = ((l0+l1+l2+l3+l4+l5+l6+l7
2214 +t0+t1+t2+t3+t4+t5+t6+t7+8) >> 4) * 0x01010101;
2215 PREDICT_8x8_DC(dc);
2216 }
2217 static void pred8x8l_horizontal_c(uint8_t *src, int has_topleft, int has_topright, int stride)
2218 {
2219 PREDICT_8x8_LOAD_LEFT;
2220 #define ROW(y) ((uint32_t*)(src+y*stride))[0] =\
2221 ((uint32_t*)(src+y*stride))[1] = 0x01010101 * l##y
2222 ROW(0); ROW(1); ROW(2); ROW(3); ROW(4); ROW(5); ROW(6); ROW(7);
2223 #undef ROW
2224 }
2225 static void pred8x8l_vertical_c(uint8_t *src, int has_topleft, int has_topright, int stride)
2226 {
2227 int y;
2228 PREDICT_8x8_LOAD_TOP;
2229 src[0] = t0;
2230 src[1] = t1;
2231 src[2] = t2;
2232 src[3] = t3;
2233 src[4] = t4;
2234 src[5] = t5;
2235 src[6] = t6;
2236 src[7] = t7;
2237 for( y = 1; y < 8; y++ )
2238 *(uint64_t*)(src+y*stride) = *(uint64_t*)src;
2239 }
2240 static void pred8x8l_down_left_c(uint8_t *src, int has_topleft, int has_topright, int stride)
2241 {
2242 PREDICT_8x8_LOAD_TOP;
2243 PREDICT_8x8_LOAD_TOPRIGHT;
2244 SRC(0,0)= (t0 + 2*t1 + t2 + 2) >> 2;
2245 SRC(0,1)=SRC(1,0)= (t1 + 2*t2 + t3 + 2) >> 2;
2246 SRC(0,2)=SRC(1,1)=SRC(2,0)= (t2 + 2*t3 + t4 + 2) >> 2;
2247 SRC(0,3)=SRC(1,2)=SRC(2,1)=SRC(3,0)= (t3 + 2*t4 + t5 + 2) >> 2;
2248 SRC(0,4)=SRC(1,3)=SRC(2,2)=SRC(3,1)=SRC(4,0)= (t4 + 2*t5 + t6 + 2) >> 2;
2249 SRC(0,5)=SRC(1,4)=SRC(2,3)=SRC(3,2)=SRC(4,1)=SRC(5,0)= (t5 + 2*t6 + t7 + 2) >> 2;
2250 SRC(0,6)=SRC(1,5)=SRC(2,4)=SRC(3,3)=SRC(4,2)=SRC(5,1)=SRC(6,0)= (t6 + 2*t7 + t8 + 2) >> 2;
2251 SRC(0,7)=SRC(1,6)=SRC(2,5)=SRC(3,4)=SRC(4,3)=SRC(5,2)=SRC(6,1)=SRC(7,0)= (t7 + 2*t8 + t9 + 2) >> 2;
2252 SRC(1,7)=SRC(2,6)=SRC(3,5)=SRC(4,4)=SRC(5,3)=SRC(6,2)=SRC(7,1)= (t8 + 2*t9 + t10 + 2) >> 2;
2253 SRC(2,7)=SRC(3,6)=SRC(4,5)=SRC(5,4)=SRC(6,3)=SRC(7,2)= (t9 + 2*t10 + t11 + 2) >> 2;
2254 SRC(3,7)=SRC(4,6)=SRC(5,5)=SRC(6,4)=SRC(7,3)= (t10 + 2*t11 + t12 + 2) >> 2;
2255 SRC(4,7)=SRC(5,6)=SRC(6,5)=SRC(7,4)= (t11 + 2*t12 + t13 + 2) >> 2;
2256 SRC(5,7)=SRC(6,6)=SRC(7,5)= (t12 + 2*t13 + t14 + 2) >> 2;
2257 SRC(6,7)=SRC(7,6)= (t13 + 2*t14 + t15 + 2) >> 2;
2258 SRC(7,7)= (t14 + 3*t15 + 2) >> 2;
2259 }
2260 static void pred8x8l_down_right_c(uint8_t *src, int has_topleft, int has_topright, int stride)
2261 {
2262 PREDICT_8x8_LOAD_TOP;
2263 PREDICT_8x8_LOAD_LEFT;
2264 PREDICT_8x8_LOAD_TOPLEFT;
2265 SRC(0,7)= (l7 + 2*l6 + l5 + 2) >> 2;
2266 SRC(0,6)=SRC(1,7)= (l6 + 2*l5 + l4 + 2) >> 2;
2267 SRC(0,5)=SRC(1,6)=SRC(2,7)= (l5 + 2*l4 + l3 + 2) >> 2;
2268 SRC(0,4)=SRC(1,5)=SRC(2,6)=SRC(3,7)= (l4 + 2*l3 + l2 + 2) >> 2;
2269 SRC(0,3)=SRC(1,4)=SRC(2,5)=SRC(3,6)=SRC(4,7)= (l3 + 2*l2 + l1 + 2) >> 2;
2270 SRC(0,2)=SRC(1,3)=SRC(2,4)=SRC(3,5)=SRC(4,6)=SRC(5,7)= (l2 + 2*l1 + l0 + 2) >> 2;
2271 SRC(0,1)=SRC(1,2)=SRC(2,3)=SRC(3,4)=SRC(4,5)=SRC(5,6)=SRC(6,7)= (l1 + 2*l0 + lt + 2) >> 2;
2272 SRC(0,0)=SRC(1,1)=SRC(2,2)=SRC(3,3)=SRC(4,4)=SRC(5,5)=SRC(6,6)=SRC(7,7)= (l0 + 2*lt + t0 + 2) >> 2;
2273 SRC(1,0)=SRC(2,1)=SRC(3,2)=SRC(4,3)=SRC(5,4)=SRC(6,5)=SRC(7,6)= (lt + 2*t0 + t1 + 2) >> 2;
2274 SRC(2,0)=SRC(3,1)=SRC(4,2)=SRC(5,3)=SRC(6,4)=SRC(7,5)= (t0 + 2*t1 + t2 + 2) >> 2;
2275 SRC(3,0)=SRC(4,1)=SRC(5,2)=SRC(6,3)=SRC(7,4)= (t1 + 2*t2 + t3 + 2) >> 2;
2276 SRC(4,0)=SRC(5,1)=SRC(6,2)=SRC(7,3)= (t2 + 2*t3 + t4 + 2) >> 2;
2277 SRC(5,0)=SRC(6,1)=SRC(7,2)= (t3 + 2*t4 + t5 + 2) >> 2;
2278 SRC(6,0)=SRC(7,1)= (t4 + 2*t5 + t6 + 2) >> 2;
2279 SRC(7,0)= (t5 + 2*t6 + t7 + 2) >> 2;
2280
2281 }
2282 static void pred8x8l_vertical_right_c(uint8_t *src, int has_topleft, int has_topright, int stride)
2283 {
2284 PREDICT_8x8_LOAD_TOP;
2285 PREDICT_8x8_LOAD_LEFT;
2286 PREDICT_8x8_LOAD_TOPLEFT;
2287 SRC(0,6)= (l5 + 2*l4 + l3 + 2) >> 2;
2288 SRC(0,7)= (l6 + 2*l5 + l4 + 2) >> 2;
2289 SRC(0,4)=SRC(1,6)= (l3 + 2*l2 + l1 + 2) >> 2;
2290 SRC(0,5)=SRC(1,7)= (l4 + 2*l3 + l2 + 2) >> 2;
2291 SRC(0,2)=SRC(1,4)=SRC(2,6)= (l1 + 2*l0 + lt + 2) >> 2;
2292 SRC(0,3)=SRC(1,5)=SRC(2,7)= (l2 + 2*l1 + l0 + 2) >> 2;
2293 SRC(0,1)=SRC(1,3)=SRC(2,5)=SRC(3,7)= (l0 + 2*lt + t0 + 2) >> 2;
2294 SRC(0,0)=SRC(1,2)=SRC(2,4)=SRC(3,6)= (lt + t0 + 1) >> 1;
2295 SRC(1,1)=SRC(2,3)=SRC(3,5)=SRC(4,7)= (lt + 2*t0 + t1 + 2) >> 2;
2296 SRC(1,0)=SRC(2,2)=SRC(3,4)=SRC(4,6)= (t0 + t1 + 1) >> 1;
2297 SRC(2,1)=SRC(3,3)=SRC(4,5)=SRC(5,7)= (t0 + 2*t1 + t2 + 2) >> 2;
2298 SRC(2,0)=SRC(3,2)=SRC(4,4)=SRC(5,6)= (t1 + t2 + 1) >> 1;
2299 SRC(3,1)=SRC(4,3)=SRC(5,5)=SRC(6,7)= (t1 + 2*t2 + t3 + 2) >> 2;
2300 SRC(3,0)=SRC(4,2)=SRC(5,4)=SRC(6,6)= (t2 + t3 + 1) >> 1;
2301 SRC(4,1)=SRC(5,3)=SRC(6,5)=SRC(7,7)= (t2 + 2*t3 + t4 + 2) >> 2;
2302 SRC(4,0)=SRC(5,2)=SRC(6,4)=SRC(7,6)= (t3 + t4 + 1) >> 1;
2303 SRC(5,1)=SRC(6,3)=SRC(7,5)= (t3 + 2*t4 + t5 + 2) >> 2;
2304 SRC(5,0)=SRC(6,2)=SRC(7,4)= (t4 + t5 + 1) >> 1;
2305 SRC(6,1)=SRC(7,3)= (t4 + 2*t5 + t6 + 2) >> 2;
2306 SRC(6,0)=SRC(7,2)= (t5 + t6 + 1) >> 1;
2307 SRC(7,1)= (t5 + 2*t6 + t7 + 2) >> 2;
2308 SRC(7,0)= (t6 + t7 + 1) >> 1;
2309 }
2310 static void pred8x8l_horizontal_down_c(uint8_t *src, int has_topleft, int has_topright, int stride)
2311 {
2312 PREDICT_8x8_LOAD_TOP;
2313 PREDICT_8x8_LOAD_LEFT;
2314 PREDICT_8x8_LOAD_TOPLEFT;
2315 SRC(0,7)= (l6 + l7 + 1) >> 1;
2316 SRC(1,7)= (l5 + 2*l6 + l7 + 2) >> 2;
2317 SRC(0,6)=SRC(2,7)= (l5 + l6 + 1) >> 1;
2318 SRC(1,6)=SRC(3,7)= (l4 + 2*l5 + l6 + 2) >> 2;
2319 SRC(0,5)=SRC(2,6)=SRC(4,7)= (l4 + l5 + 1) >> 1;
2320 SRC(1,5)=SRC(3,6)=SRC(5,7)= (l3 + 2*l4 + l5 + 2) >> 2;
2321 SRC(0,4)=SRC(2,5)=SRC(4,6)=SRC(6,7)= (l3 + l4 + 1) >> 1;
2322 SRC(1,4)=SRC(3,5)=SRC(5,6)=SRC(7,7)= (l2 + 2*l3 + l4 + 2) >> 2;
2323 SRC(0,3)=SRC(2,4)=SRC(4,5)=SRC(6,6)= (l2 + l3 + 1) >> 1;
2324 SRC(1,3)=SRC(3,4)=SRC(5,5)=SRC(7,6)= (l1 + 2*l2 + l3 + 2) >> 2;
2325 SRC(0,2)=SRC(2,3)=SRC(4,4)=SRC(6,5)= (l1 + l2 + 1) >> 1;
2326 SRC(1,2)=SRC(3,3)=SRC(5,4)=SRC(7,5)= (l0 + 2*l1 + l2 + 2) >> 2;
2327 SRC(0,1)=SRC(2,2)=SRC(4,3)=SRC(6,4)= (l0 + l1 + 1) >> 1;
2328 SRC(1,1)=SRC(3,2)=SRC(5,3)=SRC(7,4)= (lt + 2*l0 + l1 + 2) >> 2;
2329 SRC(0,0)=SRC(2,1)=SRC(4,2)=SRC(6,3)= (lt + l0 + 1) >> 1;
2330 SRC(1,0)=SRC(3,1)=SRC(5,2)=SRC(7,3)= (l0 + 2*lt + t0 + 2) >> 2;
2331 SRC(2,0)=SRC(4,1)=SRC(6,2)= (t1 + 2*t0 + lt + 2) >> 2;
2332 SRC(3,0)=SRC(5,1)=SRC(7,2)= (t2 + 2*t1 + t0 + 2) >> 2;
2333 SRC(4,0)=SRC(6,1)= (t3 + 2*t2 + t1 + 2) >> 2;
2334 SRC(5,0)=SRC(7,1)= (t4 + 2*t3 + t2 + 2) >> 2;
2335 SRC(6,0)= (t5 + 2*t4 + t3 + 2) >> 2;
2336 SRC(7,0)= (t6 + 2*t5 + t4 + 2) >> 2;
2337 }
2338 static void pred8x8l_vertical_left_c(uint8_t *src, int has_topleft, int has_topright, int stride)
2339 {
2340 PREDICT_8x8_LOAD_TOP;
2341 PREDICT_8x8_LOAD_TOPRIGHT;
2342 SRC(0,0)= (t0 + t1 + 1) >> 1;
2343 SRC(0,1)= (t0 + 2*t1 + t2 + 2) >> 2;
2344 SRC(0,2)=SRC(1,0)= (t1 + t2 + 1) >> 1;
2345 SRC(0,3)=SRC(1,1)= (t1 + 2*t2 + t3 + 2) >> 2;
2346 SRC(0,4)=SRC(1,2)=SRC(2,0)= (t2 + t3 + 1) >> 1;
2347 SRC(0,5)=SRC(1,3)=SRC(2,1)= (t2 + 2*t3 + t4 + 2) >> 2;
2348 SRC(0,6)=SRC(1,4)=SRC(2,2)=SRC(3,0)= (t3 + t4 + 1) >> 1;
2349 SRC(0,7)=SRC(1,5)=SRC(2,3)=SRC(3,1)= (t3 + 2*t4 + t5 + 2) >> 2;
2350 SRC(1,6)=SRC(2,4)=SRC(3,2)=SRC(4,0)= (t4 + t5 + 1) >> 1;
2351 SRC(1,7)=SRC(2,5)=SRC(3,3)=SRC(4,1)= (t4 + 2*t5 + t6 + 2) >> 2;
2352 SRC(2,6)=SRC(3,4)=SRC(4,2)=SRC(5,0)= (t5 + t6 + 1) >> 1;
2353 SRC(2,7)=SRC(3,5)=SRC(4,3)=SRC(5,1)= (t5 + 2*t6 + t7 + 2) >> 2;
2354 SRC(3,6)=SRC(4,4)=SRC(5,2)=SRC(6,0)= (t6 + t7 + 1) >> 1;
2355 SRC(3,7)=SRC(4,5)=SRC(5,3)=SRC(6,1)= (t6 + 2*t7 + t8 + 2) >> 2;
2356 SRC(4,6)=SRC(5,4)=SRC(6,2)=SRC(7,0)= (t7 + t8 + 1) >> 1;
2357 SRC(4,7)=SRC(5,5)=SRC(6,3)=SRC(7,1)= (t7 + 2*t8 + t9 + 2) >> 2;
2358 SRC(5,6)=SRC(6,4)=SRC(7,2)= (t8 + t9 + 1) >> 1;
2359 SRC(5,7)=SRC(6,5)=SRC(7,3)= (t8 + 2*t9 + t10 + 2) >> 2;
2360 SRC(6,6)=SRC(7,4)= (t9 + t10 + 1) >> 1;
2361 SRC(6,7)=SRC(7,5)= (t9 + 2*t10 + t11 + 2) >> 2;
2362 SRC(7,6)= (t10 + t11 + 1) >> 1;
2363 SRC(7,7)= (t10 + 2*t11 + t12 + 2) >> 2;
2364 }
2365 static void pred8x8l_horizontal_up_c(uint8_t *src, int has_topleft, int has_topright, int stride)
2366 {
2367 PREDICT_8x8_LOAD_LEFT;
2368 SRC(0,0)= (l0 + l1 + 1) >> 1;
2369 SRC(1,0)= (l0 + 2*l1 + l2 + 2) >> 2;
2370 SRC(0,1)=SRC(2,0)= (l1 + l2 + 1) >> 1;
2371 SRC(1,1)=SRC(3,0)= (l1 + 2*l2 + l3 + 2) >> 2;
2372 SRC(0,2)=SRC(2,1)=SRC(4,0)= (l2 + l3 + 1) >> 1;
2373 SRC(1,2)=SRC(3,1)=SRC(5,0)= (l2 + 2*l3 + l4 + 2) >> 2;
2374 SRC(0,3)=SRC(2,2)=SRC(4,1)=SRC(6,0)= (l3 + l4 + 1) >> 1;
2375 SRC(1,3)=SRC(3,2)=SRC(5,1)=SRC(7,0)= (l3 + 2*l4 + l5 + 2) >> 2;
2376 SRC(0,4)=SRC(2,3)=SRC(4,2)=SRC(6,1)= (l4 + l5 + 1) >> 1;
2377 SRC(1,4)=SRC(3,3)=SRC(5,2)=SRC(7,1)= (l4 + 2*l5 + l6 + 2) >> 2;
2378 SRC(0,5)=SRC(2,4)=SRC(4,3)=SRC(6,2)= (l5 + l6 + 1) >> 1;
2379 SRC(1,5)=SRC(3,4)=SRC(5,3)=SRC(7,2)= (l5 + 2*l6 + l7 + 2) >> 2;
2380 SRC(0,6)=SRC(2,5)=SRC(4,4)=SRC(6,3)= (l6 + l7 + 1) >> 1;
2381 SRC(1,6)=SRC(3,5)=SRC(5,4)=SRC(7,3)= (l6 + 3*l7 + 2) >> 2;
2382 SRC(0,7)=SRC(1,7)=SRC(2,6)=SRC(2,7)=SRC(3,6)=
2383 SRC(3,7)=SRC(4,5)=SRC(4,6)=SRC(4,7)=SRC(5,5)=
2384 SRC(5,6)=SRC(5,7)=SRC(6,4)=SRC(6,5)=SRC(6,6)=
2385 SRC(6,7)=SRC(7,4)=SRC(7,5)=SRC(7,6)=SRC(7,7)= l7;
2386 }
2387 #undef PREDICT_8x8_LOAD_LEFT
2388 #undef PREDICT_8x8_LOAD_TOP
2389 #undef PREDICT_8x8_LOAD_TOPLEFT
2390 #undef PREDICT_8x8_LOAD_TOPRIGHT
2391 #undef PREDICT_8x8_DC
2392 #undef PTR
2393 #undef PT
2394 #undef PL
2395 #undef SRC
2396
2397 static inline void mc_dir_part(H264Context *h, Picture *pic, int n, int square, int chroma_height, int delta, int list, 1684 static inline void mc_dir_part(H264Context *h, Picture *pic, int n, int square, int chroma_height, int delta, int list,
2398 uint8_t *dest_y, uint8_t *dest_cb, uint8_t *dest_cr, 1685 uint8_t *dest_y, uint8_t *dest_cb, uint8_t *dest_cr,
2399 int src_x_offset, int src_y_offset, 1686 int src_x_offset, int src_y_offset,
2400 qpel_mc_func *qpix_op, h264_chroma_mc_func chroma_op){ 1687 qpel_mc_func *qpix_op, h264_chroma_mc_func chroma_op){
2401 MpegEncContext * const s = &h->s; 1688 MpegEncContext * const s = &h->s;
2714 &run_len [6][0], 1, 1, 2001 &run_len [6][0], 1, 1,
2715 &run_bits[6][0], 1, 1, 1); 2002 &run_bits[6][0], 1, 1, 1);
2716 } 2003 }
2717 } 2004 }
2718 2005
2719 /**
2720 * Sets the intra prediction function pointers.
2721 */
2722 static void init_pred_ptrs(H264Context *h){
2723 // MpegEncContext * const s = &h->s;
2724
2725 h->pred4x4[VERT_PRED ]= pred4x4_vertical_c;
2726 h->pred4x4[HOR_PRED ]= pred4x4_horizontal_c;
2727 h->pred4x4[DC_PRED ]= pred4x4_dc_c;
2728 h->pred4x4[DIAG_DOWN_LEFT_PRED ]= pred4x4_down_left_c;
2729 h->pred4x4[DIAG_DOWN_RIGHT_PRED]= pred4x4_down_right_c;
2730 h->pred4x4[VERT_RIGHT_PRED ]= pred4x4_vertical_right_c;
2731 h->pred4x4[HOR_DOWN_PRED ]= pred4x4_horizontal_down_c;
2732 h->pred4x4[VERT_LEFT_PRED ]= pred4x4_vertical_left_c;
2733 h->pred4x4[HOR_UP_PRED ]= pred4x4_horizontal_up_c;
2734 h->pred4x4[LEFT_DC_PRED ]= pred4x4_left_dc_c;
2735 h->pred4x4[TOP_DC_PRED ]= pred4x4_top_dc_c;
2736 h->pred4x4[DC_128_PRED ]= pred4x4_128_dc_c;
2737
2738 h->pred8x8l[VERT_PRED ]= pred8x8l_vertical_c;
2739 h->pred8x8l[HOR_PRED ]= pred8x8l_horizontal_c;
2740 h->pred8x8l[DC_PRED ]= pred8x8l_dc_c;
2741 h->pred8x8l[DIAG_DOWN_LEFT_PRED ]= pred8x8l_down_left_c;
2742 h->pred8x8l[DIAG_DOWN_RIGHT_PRED]= pred8x8l_down_right_c;
2743 h->pred8x8l[VERT_RIGHT_PRED ]= pred8x8l_vertical_right_c;
2744 h->pred8x8l[HOR_DOWN_PRED ]= pred8x8l_horizontal_down_c;
2745 h->pred8x8l[VERT_LEFT_PRED ]= pred8x8l_vertical_left_c;
2746 h->pred8x8l[HOR_UP_PRED ]= pred8x8l_horizontal_up_c;
2747 h->pred8x8l[LEFT_DC_PRED ]= pred8x8l_left_dc_c;
2748 h->pred8x8l[TOP_DC_PRED ]= pred8x8l_top_dc_c;
2749 h->pred8x8l[DC_128_PRED ]= pred8x8l_128_dc_c;
2750
2751 h->pred8x8[DC_PRED8x8 ]= ff_pred8x8_dc_c;
2752 h->pred8x8[VERT_PRED8x8 ]= ff_pred8x8_vertical_c;
2753 h->pred8x8[HOR_PRED8x8 ]= ff_pred8x8_horizontal_c;
2754 h->pred8x8[PLANE_PRED8x8 ]= ff_pred8x8_plane_c;
2755 h->pred8x8[LEFT_DC_PRED8x8]= ff_pred8x8_left_dc_c;
2756 h->pred8x8[TOP_DC_PRED8x8 ]= ff_pred8x8_top_dc_c;
2757 h->pred8x8[DC_128_PRED8x8 ]= ff_pred8x8_128_dc_c;
2758
2759 h->pred16x16[DC_PRED8x8 ]= ff_pred16x16_dc_c;
2760 h->pred16x16[VERT_PRED8x8 ]= ff_pred16x16_vertical_c;
2761 h->pred16x16[HOR_PRED8x8 ]= ff_pred16x16_horizontal_c;
2762 h->pred16x16[PLANE_PRED8x8 ]= ff_pred16x16_plane_c;
2763 h->pred16x16[LEFT_DC_PRED8x8]= ff_pred16x16_left_dc_c;
2764 h->pred16x16[TOP_DC_PRED8x8 ]= ff_pred16x16_top_dc_c;
2765 h->pred16x16[DC_128_PRED8x8 ]= ff_pred16x16_128_dc_c;
2766 }
2767
2768 static void free_tables(H264Context *h){ 2006 static void free_tables(H264Context *h){
2769 int i; 2007 int i;
2770 av_freep(&h->intra4x4_pred_mode); 2008 av_freep(&h->intra4x4_pred_mode);
2771 av_freep(&h->chroma_pred_mode_table); 2009 av_freep(&h->chroma_pred_mode_table);
2772 av_freep(&h->cbp_table); 2010 av_freep(&h->cbp_table);
2912 2150
2913 s->width = s->avctx->width; 2151 s->width = s->avctx->width;
2914 s->height = s->avctx->height; 2152 s->height = s->avctx->height;
2915 s->codec_id= s->avctx->codec->id; 2153 s->codec_id= s->avctx->codec->id;
2916 2154
2917 init_pred_ptrs(h); 2155 ff_h264_pred_init(&h->hpc, s->codec_id);
2918 2156
2919 h->dequant_coeff_pps= -1; 2157 h->dequant_coeff_pps= -1;
2920 s->unrestricted_mv=1; 2158 s->unrestricted_mv=1;
2921 s->decode=1; //FIXME 2159 s->decode=1; //FIXME
2922 2160
3268 if(IS_INTRA(mb_type)){ 2506 if(IS_INTRA(mb_type)){
3269 if(h->deblocking_filter && (simple || !FRAME_MBAFF)) 2507 if(h->deblocking_filter && (simple || !FRAME_MBAFF))
3270 xchg_mb_border(h, dest_y, dest_cb, dest_cr, linesize, uvlinesize, 1, simple); 2508 xchg_mb_border(h, dest_y, dest_cb, dest_cr, linesize, uvlinesize, 1, simple);
3271 2509
3272 if(simple || !ENABLE_GRAY || !(s->flags&CODEC_FLAG_GRAY)){ 2510 if(simple || !ENABLE_GRAY || !(s->flags&CODEC_FLAG_GRAY)){
3273 h->pred8x8[ h->chroma_pred_mode ](dest_cb, uvlinesize); 2511 h->hpc.pred8x8[ h->chroma_pred_mode ](dest_cb, uvlinesize);
3274 h->pred8x8[ h->chroma_pred_mode ](dest_cr, uvlinesize); 2512 h->hpc.pred8x8[ h->chroma_pred_mode ](dest_cr, uvlinesize);
3275 } 2513 }
3276 2514
3277 if(IS_INTRA4x4(mb_type)){ 2515 if(IS_INTRA4x4(mb_type)){
3278 if(simple || !s->encoding){ 2516 if(simple || !s->encoding){
3279 if(IS_8x8DCT(mb_type)){ 2517 if(IS_8x8DCT(mb_type)){
3280 for(i=0; i<16; i+=4){ 2518 for(i=0; i<16; i+=4){
3281 uint8_t * const ptr= dest_y + block_offset[i]; 2519 uint8_t * const ptr= dest_y + block_offset[i];
3282 const int dir= h->intra4x4_pred_mode_cache[ scan8[i] ]; 2520 const int dir= h->intra4x4_pred_mode_cache[ scan8[i] ];
3283 const int nnz = h->non_zero_count_cache[ scan8[i] ]; 2521 const int nnz = h->non_zero_count_cache[ scan8[i] ];
3284 h->pred8x8l[ dir ](ptr, (h->topleft_samples_available<<i)&0x8000, 2522 h->hpc.pred8x8l[ dir ](ptr, (h->topleft_samples_available<<i)&0x8000,
3285 (h->topright_samples_available<<i)&0x4000, linesize); 2523 (h->topright_samples_available<<i)&0x4000, linesize);
3286 if(nnz){ 2524 if(nnz){
3287 if(nnz == 1 && h->mb[i*16]) 2525 if(nnz == 1 && h->mb[i*16])
3288 idct_dc_add(ptr, h->mb + i*16, linesize); 2526 idct_dc_add(ptr, h->mb + i*16, linesize);
3289 else 2527 else
3306 }else 2544 }else
3307 topright= ptr + 4 - linesize; 2545 topright= ptr + 4 - linesize;
3308 }else 2546 }else
3309 topright= NULL; 2547 topright= NULL;
3310 2548
3311 h->pred4x4[ dir ](ptr, topright, linesize); 2549 h->hpc.pred4x4[ dir ](ptr, topright, linesize);
3312 nnz = h->non_zero_count_cache[ scan8[i] ]; 2550 nnz = h->non_zero_count_cache[ scan8[i] ];
3313 if(nnz){ 2551 if(nnz){
3314 if(is_h264){ 2552 if(is_h264){
3315 if(nnz == 1 && h->mb[i*16]) 2553 if(nnz == 1 && h->mb[i*16])
3316 idct_dc_add(ptr, h->mb + i*16, linesize); 2554 idct_dc_add(ptr, h->mb + i*16, linesize);
3320 svq3_add_idct_c(ptr, h->mb + i*16, linesize, s->qscale, 0); 2558 svq3_add_idct_c(ptr, h->mb + i*16, linesize, s->qscale, 0);
3321 } 2559 }
3322 } 2560 }
3323 } 2561 }
3324 }else{ 2562 }else{
3325 h->pred16x16[ h->intra16x16_pred_mode ](dest_y , linesize); 2563 h->hpc.pred16x16[ h->intra16x16_pred_mode ](dest_y , linesize);
3326 if(is_h264){ 2564 if(is_h264){
3327 if(!transform_bypass) 2565 if(!transform_bypass)
3328 h264_luma_dc_dequant_idct_c(h->mb, s->qscale, h->dequant4_coeff[IS_INTRA(mb_type) ? 0:3][s->qscale][0]); 2566 h264_luma_dc_dequant_idct_c(h->mb, s->qscale, h->dequant4_coeff[IS_INTRA(mb_type) ? 0:3][s->qscale][0]);
3329 }else 2567 }else
3330 svq3_luma_dc_dequant_idct_c(h->mb, s->qscale); 2568 svq3_luma_dc_dequant_idct_c(h->mb, s->qscale);