X-Git-Url: https://git.sesse.net/?a=blobdiff_plain;f=libavcodec%2Fvc1dsp.c;h=8a1a83411fed4da1059c836fd1f037cd9800852a;hb=ec0c5d482e6f0cc8c94980494cf6bad39419302c;hp=16fe31b90d91e7170dcbe34a808a07103c91fa1a;hpb=a70c27e813346a11e5fff3d329ecba82fb76826a;p=ffmpeg diff --git a/libavcodec/vc1dsp.c b/libavcodec/vc1dsp.c index 16fe31b90d9..8a1a83411fe 100644 --- a/libavcodec/vc1dsp.c +++ b/libavcodec/vc1dsp.c @@ -2,20 +2,21 @@ * VC-1 and WMV3 decoder - DSP functions * Copyright (c) 2006 Konstantin Shishkov * - * This library is free software; you can redistribute it and/or + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either - * version 2 of the License, or (at your option) any later version. + * version 2.1 of the License, or (at your option) any later version. * - * This library is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with this library; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - * */ /** @@ -27,43 +28,53 @@ #include "dsputil.h" -/** Apply overlap transform to vertical edge +/** Apply overlap transform to horizontal edge */ -static void vc1_v_overlap_c(uint8_t* src, int stride, int rnd) +static void vc1_v_overlap_c(uint8_t* src, int stride) { int i; int a, b, c, d; + int d1, d2; + int rnd = 1; for(i = 0; i < 8; i++) { a = src[-2*stride]; b = src[-stride]; c = src[0]; d = src[stride]; + d1 = (a - d + 3 + rnd) >> 3; + d2 = (a - d + b - c + 4 - rnd) >> 3; - src[-2*stride] = clip_uint8((7*a + d + 4 - rnd) >> 3); - src[-stride] = clip_uint8((-a + 7*b + c + d + 3 + rnd) >> 3); - src[0] = clip_uint8((a + b + 7*c - d + 4 - rnd) >> 3); - src[stride] = clip_uint8((a + 7*d + 3 + rnd) >> 3); + src[-2*stride] = a - d1; + src[-stride] = b - d2; + src[0] = c + d2; + src[stride] = d + d1; src++; + rnd = !rnd; } } -/** Apply overlap transform to horizontal edge +/** Apply overlap transform to vertical edge */ -static void vc1_h_overlap_c(uint8_t* src, int stride, int rnd) +static void vc1_h_overlap_c(uint8_t* src, int stride) { int i; int a, b, c, d; + int d1, d2; + int rnd = 1; for(i = 0; i < 8; i++) { a = src[-2]; b = src[-1]; c = src[0]; d = src[1]; + d1 = (a - d + 3 + rnd) >> 3; + d2 = (a - d + b - c + 4 - rnd) >> 3; - src[-2] = clip_uint8((7*a + d + 4 - rnd) >> 3); - src[-1] = clip_uint8((-a + 7*b + c + d + 3 + rnd) >> 3); - src[0] = clip_uint8((a + b + 7*c - d + 4 - rnd) >> 3); - src[1] = clip_uint8((a + 7*d + 3 + rnd) >> 3); + src[-2] = a - d1; + src[-1] = b - d2; + src[0] = c + d2; + src[1] = d + d1; src += stride; + rnd = !rnd; } } @@ -79,8 +90,8 @@ static void vc1_inv_trans_8x8_c(DCTELEM block[64]) src = block; dst = block; for(i = 0; i < 8; i++){ - t1 = 12 * (src[0] + src[4]); - t2 = 12 * (src[0] - src[4]); + t1 = 12 * (src[0] + src[4]) + 4; + t2 = 12 * (src[0] - src[4]) + 4; t3 = 16 * src[2] + 6 * src[6]; t4 = 6 * src[2] - 16 * src[6]; @@ -94,14 +105,14 @@ static void vc1_inv_trans_8x8_c(DCTELEM block[64]) t3 = 9 * src[1] - 16 * src[3] + 4 * src[5] + 15 * src[7]; t4 = 4 * src[1] - 9 * src[3] + 15 * src[5] - 16 * src[7]; - dst[0] = (t5 + t1 + 4) >> 3; - dst[1] = (t6 + t2 + 4) >> 3; - dst[2] = (t7 + t3 + 4) >> 3; - dst[3] = (t8 + t4 + 4) >> 3; - dst[4] = (t8 - t4 + 4) >> 3; - dst[5] = (t7 - t3 + 4) >> 3; - dst[6] = (t6 - t2 + 4) >> 3; - dst[7] = (t5 - t1 + 4) >> 3; + dst[0] = (t5 + t1) >> 3; + dst[1] = (t6 + t2) >> 3; + dst[2] = (t7 + t3) >> 3; + dst[3] = (t8 + t4) >> 3; + dst[4] = (t8 - t4) >> 3; + dst[5] = (t7 - t3) >> 3; + dst[6] = (t6 - t2) >> 3; + dst[7] = (t5 - t1) >> 3; src += 8; dst += 8; @@ -110,8 +121,8 @@ static void vc1_inv_trans_8x8_c(DCTELEM block[64]) src = block; dst = block; for(i = 0; i < 8; i++){ - t1 = 12 * (src[ 0] + src[32]); - t2 = 12 * (src[ 0] - src[32]); + t1 = 12 * (src[ 0] + src[32]) + 64; + t2 = 12 * (src[ 0] - src[32]) + 64; t3 = 16 * src[16] + 6 * src[48]; t4 = 6 * src[16] - 16 * src[48]; @@ -125,14 +136,14 @@ static void vc1_inv_trans_8x8_c(DCTELEM block[64]) t3 = 9 * src[ 8] - 16 * src[24] + 4 * src[40] + 15 * src[56]; t4 = 4 * src[ 8] - 9 * src[24] + 15 * src[40] - 16 * src[56]; - dst[ 0] = (t5 + t1 + 64) >> 7; - dst[ 8] = (t6 + t2 + 64) >> 7; - dst[16] = (t7 + t3 + 64) >> 7; - dst[24] = (t8 + t4 + 64) >> 7; - dst[32] = (t8 - t4 + 64 + 1) >> 7; - dst[40] = (t7 - t3 + 64 + 1) >> 7; - dst[48] = (t6 - t2 + 64 + 1) >> 7; - dst[56] = (t5 - t1 + 64 + 1) >> 7; + dst[ 0] = (t5 + t1) >> 7; + dst[ 8] = (t6 + t2) >> 7; + dst[16] = (t7 + t3) >> 7; + dst[24] = (t8 + t4) >> 7; + dst[32] = (t8 - t4 + 1) >> 7; + dst[40] = (t7 - t3 + 1) >> 7; + dst[48] = (t6 - t2 + 1) >> 7; + dst[56] = (t5 - t1 + 1) >> 7; src++; dst++; @@ -141,19 +152,18 @@ static void vc1_inv_trans_8x8_c(DCTELEM block[64]) /** Do inverse transform on 8x4 part of block */ -static void vc1_inv_trans_8x4_c(DCTELEM block[64], int n) +static void vc1_inv_trans_8x4_c(uint8_t *dest, int linesize, DCTELEM *block) { int i; register int t1,t2,t3,t4,t5,t6,t7,t8; DCTELEM *src, *dst; - int off; + const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP; - off = n * 32; - src = block + off; - dst = block + off; + src = block; + dst = block; for(i = 0; i < 4; i++){ - t1 = 12 * (src[0] + src[4]); - t2 = 12 * (src[0] - src[4]); + t1 = 12 * (src[0] + src[4]) + 4; + t2 = 12 * (src[0] - src[4]) + 4; t3 = 16 * src[2] + 6 * src[6]; t4 = 6 * src[2] - 16 * src[6]; @@ -167,73 +177,66 @@ static void vc1_inv_trans_8x4_c(DCTELEM block[64], int n) t3 = 9 * src[1] - 16 * src[3] + 4 * src[5] + 15 * src[7]; t4 = 4 * src[1] - 9 * src[3] + 15 * src[5] - 16 * src[7]; - dst[0] = (t5 + t1 + 4) >> 3; - dst[1] = (t6 + t2 + 4) >> 3; - dst[2] = (t7 + t3 + 4) >> 3; - dst[3] = (t8 + t4 + 4) >> 3; - dst[4] = (t8 - t4 + 4) >> 3; - dst[5] = (t7 - t3 + 4) >> 3; - dst[6] = (t6 - t2 + 4) >> 3; - dst[7] = (t5 - t1 + 4) >> 3; + dst[0] = (t5 + t1) >> 3; + dst[1] = (t6 + t2) >> 3; + dst[2] = (t7 + t3) >> 3; + dst[3] = (t8 + t4) >> 3; + dst[4] = (t8 - t4) >> 3; + dst[5] = (t7 - t3) >> 3; + dst[6] = (t6 - t2) >> 3; + dst[7] = (t5 - t1) >> 3; src += 8; dst += 8; } - src = block + off; - dst = block + off; + src = block; for(i = 0; i < 8; i++){ - t1 = 17 * (src[ 0] + src[16]); - t2 = 17 * (src[ 0] - src[16]); - t3 = 22 * src[ 8]; - t4 = 22 * src[24]; - t5 = 10 * src[ 8]; - t6 = 10 * src[24]; - - dst[ 0] = (t1 + t3 + t6 + 64) >> 7; - dst[ 8] = (t2 - t4 + t5 + 64) >> 7; - dst[16] = (t2 + t4 - t5 + 64) >> 7; - dst[24] = (t1 - t3 - t6 + 64) >> 7; + t1 = 17 * (src[ 0] + src[16]) + 64; + t2 = 17 * (src[ 0] - src[16]) + 64; + t3 = 22 * src[ 8] + 10 * src[24]; + t4 = 22 * src[24] - 10 * src[ 8]; + + dest[0*linesize] = cm[dest[0*linesize] + ((t1 + t3) >> 7)]; + dest[1*linesize] = cm[dest[1*linesize] + ((t2 - t4) >> 7)]; + dest[2*linesize] = cm[dest[2*linesize] + ((t2 + t4) >> 7)]; + dest[3*linesize] = cm[dest[3*linesize] + ((t1 - t3) >> 7)]; src ++; - dst ++; + dest++; } } /** Do inverse transform on 4x8 parts of block */ -static void vc1_inv_trans_4x8_c(DCTELEM block[64], int n) +static void vc1_inv_trans_4x8_c(uint8_t *dest, int linesize, DCTELEM *block) { int i; register int t1,t2,t3,t4,t5,t6,t7,t8; DCTELEM *src, *dst; - int off; + const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP; - off = n * 4; - src = block + off; - dst = block + off; + src = block; + dst = block; for(i = 0; i < 8; i++){ - t1 = 17 * (src[0] + src[2]); - t2 = 17 * (src[0] - src[2]); - t3 = 22 * src[1]; - t4 = 22 * src[3]; - t5 = 10 * src[1]; - t6 = 10 * src[3]; - - dst[0] = (t1 + t3 + t6 + 4) >> 3; - dst[1] = (t2 - t4 + t5 + 4) >> 3; - dst[2] = (t2 + t4 - t5 + 4) >> 3; - dst[3] = (t1 - t3 - t6 + 4) >> 3; + t1 = 17 * (src[0] + src[2]) + 4; + t2 = 17 * (src[0] - src[2]) + 4; + t3 = 22 * src[1] + 10 * src[3]; + t4 = 22 * src[3] - 10 * src[1]; + + dst[0] = (t1 + t3) >> 3; + dst[1] = (t2 - t4) >> 3; + dst[2] = (t2 + t4) >> 3; + dst[3] = (t1 - t3) >> 3; src += 8; dst += 8; } - src = block + off; - dst = block + off; + src = block; for(i = 0; i < 4; i++){ - t1 = 12 * (src[ 0] + src[32]); - t2 = 12 * (src[ 0] - src[32]); + t1 = 12 * (src[ 0] + src[32]) + 64; + t2 = 12 * (src[ 0] - src[32]) + 64; t3 = 16 * src[16] + 6 * src[48]; t4 = 6 * src[16] - 16 * src[48]; @@ -247,74 +250,88 @@ static void vc1_inv_trans_4x8_c(DCTELEM block[64], int n) t3 = 9 * src[ 8] - 16 * src[24] + 4 * src[40] + 15 * src[56]; t4 = 4 * src[ 8] - 9 * src[24] + 15 * src[40] - 16 * src[56]; - dst[ 0] = (t5 + t1 + 64) >> 7; - dst[ 8] = (t6 + t2 + 64) >> 7; - dst[16] = (t7 + t3 + 64) >> 7; - dst[24] = (t8 + t4 + 64) >> 7; - dst[32] = (t8 - t4 + 64 + 1) >> 7; - dst[40] = (t7 - t3 + 64 + 1) >> 7; - dst[48] = (t6 - t2 + 64 + 1) >> 7; - dst[56] = (t5 - t1 + 64 + 1) >> 7; + dest[0*linesize] = cm[dest[0*linesize] + ((t5 + t1) >> 7)]; + dest[1*linesize] = cm[dest[1*linesize] + ((t6 + t2) >> 7)]; + dest[2*linesize] = cm[dest[2*linesize] + ((t7 + t3) >> 7)]; + dest[3*linesize] = cm[dest[3*linesize] + ((t8 + t4) >> 7)]; + dest[4*linesize] = cm[dest[4*linesize] + ((t8 - t4 + 1) >> 7)]; + dest[5*linesize] = cm[dest[5*linesize] + ((t7 - t3 + 1) >> 7)]; + dest[6*linesize] = cm[dest[6*linesize] + ((t6 - t2 + 1) >> 7)]; + dest[7*linesize] = cm[dest[7*linesize] + ((t5 - t1 + 1) >> 7)]; - src++; - dst++; + src ++; + dest++; } } /** Do inverse transform on 4x4 part of block */ -static void vc1_inv_trans_4x4_c(DCTELEM block[64], int n) +static void vc1_inv_trans_4x4_c(uint8_t *dest, int linesize, DCTELEM *block) { int i; - register int t1,t2,t3,t4,t5,t6; + register int t1,t2,t3,t4; DCTELEM *src, *dst; - int off; + const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP; - off = (n&1) * 4 + (n&2) * 16; - src = block + off; - dst = block + off; + src = block; + dst = block; for(i = 0; i < 4; i++){ - t1 = 17 * (src[0] + src[2]); - t2 = 17 * (src[0] - src[2]); - t3 = 22 * src[1]; - t4 = 22 * src[3]; - t5 = 10 * src[1]; - t6 = 10 * src[3]; - - dst[0] = (t1 + t3 + t6 + 4) >> 3; - dst[1] = (t2 - t4 + t5 + 4) >> 3; - dst[2] = (t2 + t4 - t5 + 4) >> 3; - dst[3] = (t1 - t3 - t6 + 4) >> 3; + t1 = 17 * (src[0] + src[2]) + 4; + t2 = 17 * (src[0] - src[2]) + 4; + t3 = 22 * src[1] + 10 * src[3]; + t4 = 22 * src[3] - 10 * src[1]; + + dst[0] = (t1 + t3) >> 3; + dst[1] = (t2 - t4) >> 3; + dst[2] = (t2 + t4) >> 3; + dst[3] = (t1 - t3) >> 3; src += 8; dst += 8; } - src = block + off; - dst = block + off; + src = block; for(i = 0; i < 4; i++){ - t1 = 17 * (src[ 0] + src[16]); - t2 = 17 * (src[ 0] - src[16]); - t3 = 22 * src[ 8]; - t4 = 22 * src[24]; - t5 = 10 * src[ 8]; - t6 = 10 * src[24]; - - dst[ 0] = (t1 + t3 + t6 + 64) >> 7; - dst[ 8] = (t2 - t4 + t5 + 64) >> 7; - dst[16] = (t2 + t4 - t5 + 64) >> 7; - dst[24] = (t1 - t3 - t6 + 64) >> 7; + t1 = 17 * (src[ 0] + src[16]) + 64; + t2 = 17 * (src[ 0] - src[16]) + 64; + t3 = 22 * src[ 8] + 10 * src[24]; + t4 = 22 * src[24] - 10 * src[ 8]; + + dest[0*linesize] = cm[dest[0*linesize] + ((t1 + t3) >> 7)]; + dest[1*linesize] = cm[dest[1*linesize] + ((t2 - t4) >> 7)]; + dest[2*linesize] = cm[dest[2*linesize] + ((t2 + t4) >> 7)]; + dest[3*linesize] = cm[dest[3*linesize] + ((t1 - t3) >> 7)]; src ++; - dst ++; + dest++; } } /* motion compensation functions */ +/** Filter in case of 2 filters */ +#define VC1_MSPEL_FILTER_16B(DIR, TYPE) \ +static av_always_inline int vc1_mspel_ ## DIR ## _filter_16bits(const TYPE *src, int stride, int mode) \ +{ \ + switch(mode){ \ + case 0: /* no shift - should not occur */ \ + return 0; \ + case 1: /* 1/4 shift */ \ + return -4*src[-stride] + 53*src[0] + 18*src[stride] - 3*src[stride*2]; \ + case 2: /* 1/2 shift */ \ + return -src[-stride] + 9*src[0] + 9*src[stride] - src[stride*2]; \ + case 3: /* 3/4 shift */ \ + return -3*src[-stride] + 18*src[0] + 53*src[stride] - 4*src[stride*2]; \ + } \ + return 0; /* should not occur */ \ +} + +VC1_MSPEL_FILTER_16B(ver, uint8_t); +VC1_MSPEL_FILTER_16B(hor, int16_t); + /** Filter used to interpolate fractional pel values */ -static always_inline int vc1_mspel_filter(const uint8_t *src, int stride, int mode, int r) +static av_always_inline int vc1_mspel_filter(const uint8_t *src, int stride, int mode, int r) { switch(mode){ case 0: //no shift @@ -331,31 +348,58 @@ static always_inline int vc1_mspel_filter(const uint8_t *src, int stride, int mo /** Function used to do motion compensation with bicubic interpolation */ -static void vc1_mspel_mc(uint8_t *dst, const uint8_t *src, int stride, int mode, int rnd) +static void vc1_mspel_mc(uint8_t *dst, const uint8_t *src, int stride, int hmode, int vmode, int rnd) { - int i, j; - uint8_t tmp[8*11], *tptr; - int m, r; - - m = (mode & 3); - r = rnd; - src -= stride; - tptr = tmp; - for(j = 0; j < 11; j++) { - for(i = 0; i < 8; i++) - tptr[i] = clip_uint8(vc1_mspel_filter(src + i, 1, m, r)); - src += stride; - tptr += 8; + int i, j; + + if (vmode) { /* Horizontal filter to apply */ + int r; + + if (hmode) { /* Vertical filter to apply, output to tmp */ + static const int shift_value[] = { 0, 5, 1, 5 }; + int shift = (shift_value[hmode]+shift_value[vmode])>>1; + int16_t tmp[11*8], *tptr = tmp; + + r = (1<<(shift-1)) + rnd-1; + + src -= 1; + for(j = 0; j < 8; j++) { + for(i = 0; i < 11; i++) + tptr[i] = (vc1_mspel_ver_filter_16bits(src + i, stride, vmode)+r)>>shift; + src += stride; + tptr += 11; + } + + r = 64-rnd; + tptr = tmp+1; + for(j = 0; j < 8; j++) { + for(i = 0; i < 8; i++) + dst[i] = av_clip_uint8((vc1_mspel_hor_filter_16bits(tptr + i, 1, hmode)+r)>>7); + dst += stride; + tptr += 11; + } + + return; + } + else { /* No horizontal filter, output 8 lines to dst */ + r = 1-rnd; + + for(j = 0; j < 8; j++) { + for(i = 0; i < 8; i++) + dst[i] = av_clip_uint8(vc1_mspel_filter(src + i, stride, vmode, r)); + src += stride; + dst += stride; + } + return; + } } - r = 1 - rnd; - m = (mode >> 2) & 3; - tptr = tmp + 8; + /* Horizontal mode with no vertical mode */ for(j = 0; j < 8; j++) { for(i = 0; i < 8; i++) - dst[i] = clip_uint8(vc1_mspel_filter(tptr + i, 8, m, r)); + dst[i] = av_clip_uint8(vc1_mspel_filter(src + i, 1, hmode, rnd)); dst += stride; - tptr += 8; + src += stride; } } @@ -364,65 +408,29 @@ static void vc1_mspel_mc(uint8_t *dst, const uint8_t *src, int stride, int mode, /* this one is defined in dsputil.c */ void ff_put_vc1_mspel_mc00_c(uint8_t *dst, const uint8_t *src, int stride, int rnd); -static void ff_put_vc1_mspel_mc10_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { - vc1_mspel_mc(dst, src, stride, 0x1, rnd); -} - -static void ff_put_vc1_mspel_mc20_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { - vc1_mspel_mc(dst, src, stride, 0x2, rnd); -} - -static void ff_put_vc1_mspel_mc30_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { - vc1_mspel_mc(dst, src, stride, 0x3, rnd); -} - -static void ff_put_vc1_mspel_mc01_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { - vc1_mspel_mc(dst, src, stride, 0x4, rnd); -} - -static void ff_put_vc1_mspel_mc11_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { - vc1_mspel_mc(dst, src, stride, 0x5, rnd); -} - -static void ff_put_vc1_mspel_mc21_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { - vc1_mspel_mc(dst, src, stride, 0x6, rnd); -} - -static void ff_put_vc1_mspel_mc31_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { - vc1_mspel_mc(dst, src, stride, 0x7, rnd); +#define PUT_VC1_MSPEL(a, b)\ +static void put_vc1_mspel_mc ## a ## b ##_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { \ + vc1_mspel_mc(dst, src, stride, a, b, rnd); \ } -static void ff_put_vc1_mspel_mc02_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { - vc1_mspel_mc(dst, src, stride, 0x8, rnd); -} +PUT_VC1_MSPEL(1, 0) +PUT_VC1_MSPEL(2, 0) +PUT_VC1_MSPEL(3, 0) -static void ff_put_vc1_mspel_mc12_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { - vc1_mspel_mc(dst, src, stride, 0x9, rnd); -} +PUT_VC1_MSPEL(0, 1) +PUT_VC1_MSPEL(1, 1) +PUT_VC1_MSPEL(2, 1) +PUT_VC1_MSPEL(3, 1) -static void ff_put_vc1_mspel_mc22_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { - vc1_mspel_mc(dst, src, stride, 0xA, rnd); -} +PUT_VC1_MSPEL(0, 2) +PUT_VC1_MSPEL(1, 2) +PUT_VC1_MSPEL(2, 2) +PUT_VC1_MSPEL(3, 2) -static void ff_put_vc1_mspel_mc32_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { - vc1_mspel_mc(dst, src, stride, 0xB, rnd); -} - -static void ff_put_vc1_mspel_mc03_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { - vc1_mspel_mc(dst, src, stride, 0xC, rnd); -} - -static void ff_put_vc1_mspel_mc13_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { - vc1_mspel_mc(dst, src, stride, 0xD, rnd); -} - -static void ff_put_vc1_mspel_mc23_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { - vc1_mspel_mc(dst, src, stride, 0xE, rnd); -} - -static void ff_put_vc1_mspel_mc33_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { - vc1_mspel_mc(dst, src, stride, 0xF, rnd); -} +PUT_VC1_MSPEL(0, 3) +PUT_VC1_MSPEL(1, 3) +PUT_VC1_MSPEL(2, 3) +PUT_VC1_MSPEL(3, 3) void ff_vc1dsp_init(DSPContext* dsp, AVCodecContext *avctx) { dsp->vc1_inv_trans_8x8 = vc1_inv_trans_8x8_c; @@ -433,19 +441,19 @@ void ff_vc1dsp_init(DSPContext* dsp, AVCodecContext *avctx) { dsp->vc1_v_overlap = vc1_v_overlap_c; dsp->put_vc1_mspel_pixels_tab[ 0] = ff_put_vc1_mspel_mc00_c; - dsp->put_vc1_mspel_pixels_tab[ 1] = ff_put_vc1_mspel_mc10_c; - dsp->put_vc1_mspel_pixels_tab[ 2] = ff_put_vc1_mspel_mc20_c; - dsp->put_vc1_mspel_pixels_tab[ 3] = ff_put_vc1_mspel_mc30_c; - dsp->put_vc1_mspel_pixels_tab[ 4] = ff_put_vc1_mspel_mc01_c; - dsp->put_vc1_mspel_pixels_tab[ 5] = ff_put_vc1_mspel_mc11_c; - dsp->put_vc1_mspel_pixels_tab[ 6] = ff_put_vc1_mspel_mc21_c; - dsp->put_vc1_mspel_pixels_tab[ 7] = ff_put_vc1_mspel_mc31_c; - dsp->put_vc1_mspel_pixels_tab[ 8] = ff_put_vc1_mspel_mc02_c; - dsp->put_vc1_mspel_pixels_tab[ 9] = ff_put_vc1_mspel_mc12_c; - dsp->put_vc1_mspel_pixels_tab[10] = ff_put_vc1_mspel_mc22_c; - dsp->put_vc1_mspel_pixels_tab[11] = ff_put_vc1_mspel_mc32_c; - dsp->put_vc1_mspel_pixels_tab[12] = ff_put_vc1_mspel_mc03_c; - dsp->put_vc1_mspel_pixels_tab[13] = ff_put_vc1_mspel_mc13_c; - dsp->put_vc1_mspel_pixels_tab[14] = ff_put_vc1_mspel_mc23_c; - dsp->put_vc1_mspel_pixels_tab[15] = ff_put_vc1_mspel_mc33_c; + dsp->put_vc1_mspel_pixels_tab[ 1] = put_vc1_mspel_mc10_c; + dsp->put_vc1_mspel_pixels_tab[ 2] = put_vc1_mspel_mc20_c; + dsp->put_vc1_mspel_pixels_tab[ 3] = put_vc1_mspel_mc30_c; + dsp->put_vc1_mspel_pixels_tab[ 4] = put_vc1_mspel_mc01_c; + dsp->put_vc1_mspel_pixels_tab[ 5] = put_vc1_mspel_mc11_c; + dsp->put_vc1_mspel_pixels_tab[ 6] = put_vc1_mspel_mc21_c; + dsp->put_vc1_mspel_pixels_tab[ 7] = put_vc1_mspel_mc31_c; + dsp->put_vc1_mspel_pixels_tab[ 8] = put_vc1_mspel_mc02_c; + dsp->put_vc1_mspel_pixels_tab[ 9] = put_vc1_mspel_mc12_c; + dsp->put_vc1_mspel_pixels_tab[10] = put_vc1_mspel_mc22_c; + dsp->put_vc1_mspel_pixels_tab[11] = put_vc1_mspel_mc32_c; + dsp->put_vc1_mspel_pixels_tab[12] = put_vc1_mspel_mc03_c; + dsp->put_vc1_mspel_pixels_tab[13] = put_vc1_mspel_mc13_c; + dsp->put_vc1_mspel_pixels_tab[14] = put_vc1_mspel_mc23_c; + dsp->put_vc1_mspel_pixels_tab[15] = put_vc1_mspel_mc33_c; }