* VC-1 and WMV3 decoder - DSP functions
* Copyright (c) 2006 Konstantin Shishkov
*
- * This file is part of FFmpeg.
+ * This file is part of Libav.
*
- * FFmpeg is free software; you can redistribute it and/or
+ * Libav 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 2.1 of the License, or (at your option) any later version.
*
- * FFmpeg is distributed in the hope that it will be useful,
+ * Libav 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 FFmpeg; if not, write to the Free Software
+ * License along with Libav; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
/**
-* @file libavcodec/vc1dsp.c
+* @file
* VC-1 and WMV3 decoder
*
*/
-#include "dsputil.h"
+#include "vc1dsp.h"
/** Apply overlap transform to horizontal edge
}
}
+static void vc1_v_s_overlap_c(DCTELEM *top, DCTELEM *bottom)
+{
+ int i;
+ int a, b, c, d;
+ int d1, d2;
+ int rnd1 = 4, rnd2 = 3;
+ for(i = 0; i < 8; i++) {
+ a = top[48];
+ b = top[56];
+ c = bottom[0];
+ d = bottom[8];
+ d1 = a - d;
+ d2 = a - d + b - c;
+
+ top[48] = ((a << 3) - d1 + rnd1) >> 3;
+ top[56] = ((b << 3) - d2 + rnd2) >> 3;
+ bottom[0] = ((c << 3) + d2 + rnd1) >> 3;
+ bottom[8] = ((d << 3) + d1 + rnd2) >> 3;
+
+ bottom++;
+ top++;
+ rnd2 = 7 - rnd2;
+ rnd1 = 7 - rnd1;
+ }
+}
+
+static void vc1_h_s_overlap_c(DCTELEM *left, DCTELEM *right)
+{
+ int i;
+ int a, b, c, d;
+ int d1, d2;
+ int rnd1 = 4, rnd2 = 3;
+ for(i = 0; i < 8; i++) {
+ a = left[6];
+ b = left[7];
+ c = right[0];
+ d = right[1];
+ d1 = a - d;
+ d2 = a - d + b - c;
+
+ left[6] = ((a << 3) - d1 + rnd1) >> 3;
+ left[7] = ((b << 3) - d2 + rnd2) >> 3;
+ right[0] = ((c << 3) + d2 + rnd1) >> 3;
+ right[1] = ((d << 3) + d1 + rnd2) >> 3;
+
+ right += 8;
+ left += 8;
+ rnd2 = 7 - rnd2;
+ rnd1 = 7 - rnd1;
+ }
+}
+
/**
* VC-1 in-loop deblocking filter for one line
* @param src source block type
/** Do inverse transform on 8x8 block
*/
+static void vc1_inv_trans_8x8_dc_c(uint8_t *dest, int linesize, DCTELEM *block)
+{
+ int i;
+ int dc = block[0];
+ const uint8_t *cm;
+ dc = (3 * dc + 1) >> 1;
+ dc = (3 * dc + 16) >> 5;
+ cm = ff_cropTbl + MAX_NEG_CROP + dc;
+ for(i = 0; i < 8; i++){
+ dest[0] = cm[dest[0]];
+ dest[1] = cm[dest[1]];
+ dest[2] = cm[dest[2]];
+ dest[3] = cm[dest[3]];
+ dest[4] = cm[dest[4]];
+ dest[5] = cm[dest[5]];
+ dest[6] = cm[dest[6]];
+ dest[7] = cm[dest[7]];
+ dest += linesize;
+ }
+}
+
static void vc1_inv_trans_8x8_c(DCTELEM block[64])
{
int i;
register int t1,t2,t3,t4,t5,t6,t7,t8;
- DCTELEM *src, *dst;
+ DCTELEM *src, *dst, temp[64];
src = block;
- dst = block;
+ dst = temp;
for(i = 0; i < 8; i++){
- 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];
+ t1 = 12 * (src[ 0] + src[32]) + 4;
+ t2 = 12 * (src[ 0] - src[32]) + 4;
+ t3 = 16 * src[16] + 6 * src[48];
+ t4 = 6 * src[16] - 16 * src[48];
t5 = t1 + t3;
t6 = t2 + t4;
t7 = t2 - t4;
t8 = t1 - t3;
- t1 = 16 * src[1] + 15 * src[3] + 9 * src[5] + 4 * src[7];
- t2 = 15 * src[1] - 4 * src[3] - 16 * src[5] - 9 * src[7];
- 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];
+ t1 = 16 * src[ 8] + 15 * src[24] + 9 * src[40] + 4 * src[56];
+ t2 = 15 * src[ 8] - 4 * src[24] - 16 * src[40] - 9 * src[56];
+ 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) >> 3;
dst[1] = (t6 + t2) >> 3;
dst[6] = (t6 - t2) >> 3;
dst[7] = (t5 - t1) >> 3;
- src += 8;
+ src += 1;
dst += 8;
}
- src = block;
+ src = temp;
dst = block;
for(i = 0; i < 8; i++){
t1 = 12 * (src[ 0] + src[32]) + 64;
/** Do inverse transform on 8x4 part of block
*/
+static void vc1_inv_trans_8x4_dc_c(uint8_t *dest, int linesize, DCTELEM *block)
+{
+ int i;
+ int dc = block[0];
+ const uint8_t *cm;
+ dc = ( 3 * dc + 1) >> 1;
+ dc = (17 * dc + 64) >> 7;
+ cm = ff_cropTbl + MAX_NEG_CROP + dc;
+ for(i = 0; i < 4; i++){
+ dest[0] = cm[dest[0]];
+ dest[1] = cm[dest[1]];
+ dest[2] = cm[dest[2]];
+ dest[3] = cm[dest[3]];
+ dest[4] = cm[dest[4]];
+ dest[5] = cm[dest[5]];
+ dest[6] = cm[dest[6]];
+ dest[7] = cm[dest[7]];
+ dest += linesize;
+ }
+}
+
static void vc1_inv_trans_8x4_c(uint8_t *dest, int linesize, DCTELEM *block)
{
int i;
/** Do inverse transform on 4x8 parts of block
*/
+static void vc1_inv_trans_4x8_dc_c(uint8_t *dest, int linesize, DCTELEM *block)
+{
+ int i;
+ int dc = block[0];
+ const uint8_t *cm;
+ dc = (17 * dc + 4) >> 3;
+ dc = (12 * dc + 64) >> 7;
+ cm = ff_cropTbl + MAX_NEG_CROP + dc;
+ for(i = 0; i < 8; i++){
+ dest[0] = cm[dest[0]];
+ dest[1] = cm[dest[1]];
+ dest[2] = cm[dest[2]];
+ dest[3] = cm[dest[3]];
+ dest += linesize;
+ }
+}
+
static void vc1_inv_trans_4x8_c(uint8_t *dest, int linesize, DCTELEM *block)
{
int i;
/** Do inverse transform on 4x4 part of block
*/
+static void vc1_inv_trans_4x4_dc_c(uint8_t *dest, int linesize, DCTELEM *block)
+{
+ int i;
+ int dc = block[0];
+ const uint8_t *cm;
+ dc = (17 * dc + 4) >> 3;
+ dc = (17 * dc + 64) >> 7;
+ cm = ff_cropTbl + MAX_NEG_CROP + dc;
+ for(i = 0; i < 4; i++){
+ dest[0] = cm[dest[0]];
+ dest[1] = cm[dest[1]];
+ dest[2] = cm[dest[2]];
+ dest[3] = cm[dest[3]];
+ dest += linesize;
+ }
+}
+
static void vc1_inv_trans_4x4_c(uint8_t *dest, int linesize, DCTELEM *block)
{
int i;
return 0; /* should not occur */ \
}
-VC1_MSPEL_FILTER_16B(ver, uint8_t);
-VC1_MSPEL_FILTER_16B(hor, int16_t);
+VC1_MSPEL_FILTER_16B(ver, uint8_t)
+VC1_MSPEL_FILTER_16B(hor, int16_t)
/** Filter used to interpolate fractional pel values
/* pixel functions - really are entry points to vc1_mspel_mc */
-/* 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);
-void ff_avg_vc1_mspel_mc00_c(uint8_t *dst, const uint8_t *src, int stride, int 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) { \
put_vc1_mspel_mc(dst, src, stride, a, b, rnd); \
PUT_VC1_MSPEL(2, 3)
PUT_VC1_MSPEL(3, 3)
-void ff_vc1dsp_init(DSPContext* dsp, AVCodecContext *avctx) {
+static void put_no_rnd_vc1_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++)
+ {
+ dst[0] = (A*src[0] + B*src[1] + C*src[stride+0] + D*src[stride+1] + 32 - 4) >> 6;
+ dst[1] = (A*src[1] + B*src[2] + C*src[stride+1] + D*src[stride+2] + 32 - 4) >> 6;
+ dst[2] = (A*src[2] + B*src[3] + C*src[stride+2] + D*src[stride+3] + 32 - 4) >> 6;
+ dst[3] = (A*src[3] + B*src[4] + C*src[stride+3] + D*src[stride+4] + 32 - 4) >> 6;
+ dst[4] = (A*src[4] + B*src[5] + C*src[stride+4] + D*src[stride+5] + 32 - 4) >> 6;
+ dst[5] = (A*src[5] + B*src[6] + C*src[stride+5] + D*src[stride+6] + 32 - 4) >> 6;
+ dst[6] = (A*src[6] + B*src[7] + C*src[stride+6] + D*src[stride+7] + 32 - 4) >> 6;
+ dst[7] = (A*src[7] + B*src[8] + C*src[stride+7] + D*src[stride+8] + 32 - 4) >> 6;
+ dst+= stride;
+ src+= stride;
+ }
+}
+
+static void put_no_rnd_vc1_chroma_mc4_c(uint8_t *dst, uint8_t *src, 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++)
+ {
+ dst[0] = (A*src[0] + B*src[1] + C*src[stride+0] + D*src[stride+1] + 32 - 4) >> 6;
+ dst[1] = (A*src[1] + B*src[2] + C*src[stride+1] + D*src[stride+2] + 32 - 4) >> 6;
+ dst[2] = (A*src[2] + B*src[3] + C*src[stride+2] + D*src[stride+3] + 32 - 4) >> 6;
+ dst[3] = (A*src[3] + B*src[4] + C*src[stride+3] + D*src[stride+4] + 32 - 4) >> 6;
+ dst+= stride;
+ src+= stride;
+ }
+}
+
+#define avg2(a,b) ((a+b+1)>>1)
+static void avg_no_rnd_vc1_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++)
+ {
+ dst[0] = avg2(dst[0], ((A*src[0] + B*src[1] + C*src[stride+0] + D*src[stride+1] + 32 - 4) >> 6));
+ dst[1] = avg2(dst[1], ((A*src[1] + B*src[2] + C*src[stride+1] + D*src[stride+2] + 32 - 4) >> 6));
+ dst[2] = avg2(dst[2], ((A*src[2] + B*src[3] + C*src[stride+2] + D*src[stride+3] + 32 - 4) >> 6));
+ dst[3] = avg2(dst[3], ((A*src[3] + B*src[4] + C*src[stride+3] + D*src[stride+4] + 32 - 4) >> 6));
+ dst[4] = avg2(dst[4], ((A*src[4] + B*src[5] + C*src[stride+4] + D*src[stride+5] + 32 - 4) >> 6));
+ dst[5] = avg2(dst[5], ((A*src[5] + B*src[6] + C*src[stride+5] + D*src[stride+6] + 32 - 4) >> 6));
+ dst[6] = avg2(dst[6], ((A*src[6] + B*src[7] + C*src[stride+6] + D*src[stride+7] + 32 - 4) >> 6));
+ dst[7] = avg2(dst[7], ((A*src[7] + B*src[8] + C*src[stride+7] + D*src[stride+8] + 32 - 4) >> 6));
+ dst+= stride;
+ src+= stride;
+ }
+}
+
+#if CONFIG_WMV3IMAGE_DECODER || CONFIG_VC1IMAGE_DECODER
+
+static void sprite_h_c(uint8_t *dst, const uint8_t *src, int offset, int advance, int count)
+{
+ while (count--) {
+ int a = src[(offset >> 16) ];
+ int b = src[(offset >> 16) + 1];
+ *dst++ = a + ((b - a) * (offset&0xFFFF) >> 16);
+ offset += advance;
+ }
+}
+
+static av_always_inline void sprite_v_template(uint8_t *dst, const uint8_t *src1a, const uint8_t *src1b, int offset1,
+ int two_sprites, const uint8_t *src2a, const uint8_t *src2b, int offset2,
+ int alpha, int scaled, int width)
+{
+ int a1, b1, a2, b2;
+ while (width--) {
+ a1 = *src1a++;
+ if (scaled) {
+ b1 = *src1b++;
+ a1 = a1 + ((b1 - a1) * offset1 >> 16);
+ }
+ if (two_sprites) {
+ a2 = *src2a++;
+ if (scaled > 1) {
+ b2 = *src2b++;
+ a2 = a2 + ((b2 - a2) * offset2 >> 16);
+ }
+ a1 = a1 + ((a2 - a1) * alpha >> 16);
+ }
+ *dst++ = a1;
+ }
+}
+
+static void sprite_v_single_c(uint8_t *dst, const uint8_t *src1a, const uint8_t *src1b, int offset, int width)
+{
+ sprite_v_template(dst, src1a, src1b, offset, 0, NULL, NULL, 0, 0, 1, width);
+}
+
+static void sprite_v_double_noscale_c(uint8_t *dst, const uint8_t *src1a, const uint8_t *src2a, int alpha, int width)
+{
+ sprite_v_template(dst, src1a, NULL, 0, 1, src2a, NULL, 0, alpha, 0, width);
+}
+
+static void sprite_v_double_onescale_c(uint8_t *dst, const uint8_t *src1a, const uint8_t *src1b, int offset1,
+ const uint8_t *src2a, int alpha, int width)
+{
+ sprite_v_template(dst, src1a, src1b, offset1, 1, src2a, NULL, 0, alpha, 1, width);
+}
+
+static void sprite_v_double_twoscale_c(uint8_t *dst, const uint8_t *src1a, const uint8_t *src1b, int offset1,
+ const uint8_t *src2a, const uint8_t *src2b, int offset2,
+ int alpha, int width)
+{
+ sprite_v_template(dst, src1a, src1b, offset1, 1, src2a, src2b, offset2, alpha, 2, width);
+}
+
+#endif
+
+av_cold void ff_vc1dsp_init(VC1DSPContext* dsp) {
dsp->vc1_inv_trans_8x8 = vc1_inv_trans_8x8_c;
dsp->vc1_inv_trans_4x8 = vc1_inv_trans_4x8_c;
dsp->vc1_inv_trans_8x4 = vc1_inv_trans_8x4_c;
dsp->vc1_inv_trans_4x4 = vc1_inv_trans_4x4_c;
+ dsp->vc1_inv_trans_8x8_dc = vc1_inv_trans_8x8_dc_c;
+ dsp->vc1_inv_trans_4x8_dc = vc1_inv_trans_4x8_dc_c;
+ dsp->vc1_inv_trans_8x4_dc = vc1_inv_trans_8x4_dc_c;
+ dsp->vc1_inv_trans_4x4_dc = vc1_inv_trans_4x4_dc_c;
dsp->vc1_h_overlap = vc1_h_overlap_c;
dsp->vc1_v_overlap = vc1_v_overlap_c;
+ dsp->vc1_h_s_overlap = vc1_h_s_overlap_c;
+ dsp->vc1_v_s_overlap = vc1_v_s_overlap_c;
dsp->vc1_v_loop_filter4 = vc1_v_loop_filter4_c;
dsp->vc1_h_loop_filter4 = vc1_h_loop_filter4_c;
dsp->vc1_v_loop_filter8 = vc1_v_loop_filter8_c;
dsp->vc1_v_loop_filter16 = vc1_v_loop_filter16_c;
dsp->vc1_h_loop_filter16 = vc1_h_loop_filter16_c;
- dsp->put_vc1_mspel_pixels_tab[ 0] = ff_put_vc1_mspel_mc00_c;
+ dsp->put_vc1_mspel_pixels_tab[ 0] = ff_put_pixels8x8_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[14] = put_vc1_mspel_mc23_c;
dsp->put_vc1_mspel_pixels_tab[15] = put_vc1_mspel_mc33_c;
- dsp->avg_vc1_mspel_pixels_tab[ 0] = ff_avg_vc1_mspel_mc00_c;
+ dsp->avg_vc1_mspel_pixels_tab[ 0] = ff_avg_pixels8x8_c;
dsp->avg_vc1_mspel_pixels_tab[ 1] = avg_vc1_mspel_mc10_c;
dsp->avg_vc1_mspel_pixels_tab[ 2] = avg_vc1_mspel_mc20_c;
dsp->avg_vc1_mspel_pixels_tab[ 3] = avg_vc1_mspel_mc30_c;
dsp->avg_vc1_mspel_pixels_tab[13] = avg_vc1_mspel_mc13_c;
dsp->avg_vc1_mspel_pixels_tab[14] = avg_vc1_mspel_mc23_c;
dsp->avg_vc1_mspel_pixels_tab[15] = avg_vc1_mspel_mc33_c;
+
+ dsp->put_no_rnd_vc1_chroma_pixels_tab[0]= put_no_rnd_vc1_chroma_mc8_c;
+ dsp->avg_no_rnd_vc1_chroma_pixels_tab[0]= avg_no_rnd_vc1_chroma_mc8_c;
+ dsp->put_no_rnd_vc1_chroma_pixels_tab[1] = put_no_rnd_vc1_chroma_mc4_c;
+
+#if CONFIG_WMV3IMAGE_DECODER || CONFIG_VC1IMAGE_DECODER
+ dsp->sprite_h = sprite_h_c;
+ dsp->sprite_v_single = sprite_v_single_c;
+ dsp->sprite_v_double_noscale = sprite_v_double_noscale_c;
+ dsp->sprite_v_double_onescale = sprite_v_double_onescale_c;
+ dsp->sprite_v_double_twoscale = sprite_v_double_twoscale_c;
+#endif
+
+ if (HAVE_ALTIVEC)
+ ff_vc1dsp_init_altivec(dsp);
+ if (HAVE_MMX)
+ ff_vc1dsp_init_mmx(dsp);
}