#include "faandct.h"
#include "faanidct.h"
#include "mathops.h"
-#include "h263.h"
-#include "snow.h"
-
-/* snow.c */
-void ff_spatial_dwt(int *buffer, int width, int height, int stride, int type, int decomposition_count);
-
-/* vorbis.c */
-void vorbis_inverse_coupling(float *mag, float *ang, int blocksize);
-
-/* ac3dec.c */
-void ff_ac3_downmix_c(float (*samples)[256], float (*matrix)[2], int out_ch, int in_ch, int len);
-
-/* flacenc.c */
-void ff_flac_compute_autocorr(const int32_t *data, int len, int lag, double *autoc);
-
-/* pngdec.c */
-void ff_add_png_paeth_prediction(uint8_t *dst, uint8_t *src, uint8_t *top, int w, int bpp);
-
-/* eaidct.c */
-void ff_ea_idct_put_c(uint8_t *dest, int linesize, DCTELEM *block);
+#include "mpegvideo.h"
+#include "config.h"
+#include "lpc.h"
+#include "ac3dec.h"
+#include "vorbis.h"
+#include "png.h"
uint8_t ff_cropTbl[256 + 2 * MAX_NEG_CROP] = {0, };
uint32_t ff_squareTbl[512] = {0, };
};
/* not permutated inverse zigzag_direct + 1 for MMX quantizer */
-DECLARE_ALIGNED_8(uint16_t, inv_zigzag_direct16[64]) = {0, };
+DECLARE_ALIGNED(16, uint16_t, inv_zigzag_direct16)[64];
const uint8_t ff_alternate_horizontal_scan[64] = {
0, 1, 2, 3, 8, 9, 16, 17,
38, 46, 54, 62, 39, 47, 55, 63,
};
-/* a*inverse[b]>>32 == a/b for all 0<=a<=65536 && 2<=b<=255 */
-const uint32_t ff_inverse[256]={
+/* a*inverse[b]>>32 == a/b for all 0<=a<=16909558 && 2<=b<=256
+ * for a>16909558, is an overestimate by less than 1 part in 1<<24 */
+const uint32_t ff_inverse[257]={
0, 4294967295U,2147483648U,1431655766, 1073741824, 858993460, 715827883, 613566757,
536870912, 477218589, 429496730, 390451573, 357913942, 330382100, 306783379, 286331154,
268435456, 252645136, 238609295, 226050911, 214748365, 204522253, 195225787, 186737709,
18512791, 18433337, 18354562, 18276457, 18199014, 18122225, 18046082, 17970575,
17895698, 17821442, 17747799, 17674763, 17602325, 17530479, 17459217, 17388532,
17318417, 17248865, 17179870, 17111424, 17043522, 16976156, 16909321, 16843010,
+ 16777216
};
/* Input permutation for the simple_idct_mmx */
return s;
}
-
-#if CONFIG_SNOW_ENCODER //dwt is in snow.c
-static inline int w_c(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int w, int h, int type){
- int s, i, j;
- const int dec_count= w==8 ? 3 : 4;
- int tmp[32*32];
- int level, ori;
- static const int scale[2][2][4][4]={
- {
- {
- // 9/7 8x8 dec=3
- {268, 239, 239, 213},
- { 0, 224, 224, 152},
- { 0, 135, 135, 110},
- },{
- // 9/7 16x16 or 32x32 dec=4
- {344, 310, 310, 280},
- { 0, 320, 320, 228},
- { 0, 175, 175, 136},
- { 0, 129, 129, 102},
- }
- },{
- {
- // 5/3 8x8 dec=3
- {275, 245, 245, 218},
- { 0, 230, 230, 156},
- { 0, 138, 138, 113},
- },{
- // 5/3 16x16 or 32x32 dec=4
- {352, 317, 317, 286},
- { 0, 328, 328, 233},
- { 0, 180, 180, 140},
- { 0, 132, 132, 105},
- }
- }
- };
-
- for (i = 0; i < h; i++) {
- for (j = 0; j < w; j+=4) {
- tmp[32*i+j+0] = (pix1[j+0] - pix2[j+0])<<4;
- tmp[32*i+j+1] = (pix1[j+1] - pix2[j+1])<<4;
- tmp[32*i+j+2] = (pix1[j+2] - pix2[j+2])<<4;
- tmp[32*i+j+3] = (pix1[j+3] - pix2[j+3])<<4;
- }
- pix1 += line_size;
- pix2 += line_size;
- }
-
- ff_spatial_dwt(tmp, w, h, 32, type, dec_count);
-
- s=0;
- assert(w==h);
- for(level=0; level<dec_count; level++){
- for(ori= level ? 1 : 0; ori<4; ori++){
- int size= w>>(dec_count-level);
- int sx= (ori&1) ? size : 0;
- int stride= 32<<(dec_count-level);
- int sy= (ori&2) ? stride>>1 : 0;
-
- for(i=0; i<size; i++){
- for(j=0; j<size; j++){
- int v= tmp[sx + sy + i*stride + j] * scale[type][dec_count-3][level][ori];
- s += FFABS(v);
- }
- }
- }
- }
- assert(s>=0);
- return s>>9;
-}
-
-static int w53_8_c(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int h){
- return w_c(v, pix1, pix2, line_size, 8, h, 1);
-}
-
-static int w97_8_c(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int h){
- return w_c(v, pix1, pix2, line_size, 8, h, 0);
-}
-
-static int w53_16_c(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int h){
- return w_c(v, pix1, pix2, line_size, 16, h, 1);
-}
-
-static int w97_16_c(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int h){
- return w_c(v, pix1, pix2, line_size, 16, h, 0);
-}
-
-int w53_32_c(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int h){
- return w_c(v, pix1, pix2, line_size, 32, h, 1);
-}
-
-int w97_32_c(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int h){
- return w_c(v, pix1, pix2, line_size, 32, h, 0);
-}
-#endif
-
/* draw the edges of width 'w' of an image of size width, height */
//FIXME check that this is ok for mpeg4 interlaced
static void draw_edges_c(uint8_t *buf, int wrap, int width, int height, int w)
}
}
+static void put_pixels_nonclamped_c(const DCTELEM *block, uint8_t *restrict pixels,
+ int line_size)
+{
+ int i;
+
+ /* read the pixels */
+ for(i=0;i<8;i++) {
+ pixels[0] = block[0];
+ pixels[1] = block[1];
+ pixels[2] = block[2];
+ pixels[3] = block[3];
+ pixels[4] = block[4];
+ pixels[5] = block[5];
+ pixels[6] = block[6];
+ pixels[7] = block[7];
+
+ pixels += line_size;
+ block += 8;
+ }
+}
+
static void add_pixels_clamped_c(const DCTELEM *block, uint8_t *restrict pixels,
int line_size)
{
return sum;
}
+static void fill_block16_c(uint8_t *block, uint8_t value, int line_size, int h)
+{
+ int i;
+
+ for (i = 0; i < h; i++) {
+ memset(block, value, 16);
+ block += line_size;
+ }
+}
+
+static void fill_block8_c(uint8_t *block, uint8_t value, int line_size, int h)
+{
+ int i;
+
+ for (i = 0; i < h; i++) {
+ memset(block, value, 8);
+ block += line_size;
+ }
+}
+
+static void scale_block_c(const uint8_t src[64]/*align 8*/, uint8_t *dst/*align 8*/, int linesize)
+{
+ int i, j;
+ uint16_t *dst1 = (uint16_t *) dst;
+ uint16_t *dst2 = (uint16_t *)(dst + linesize);
+
+ for (j = 0; j < 8; j++) {
+ for (i = 0; i < 8; i++) {
+ dst1[i] = dst2[i] = src[i] * 0x0101;
+ }
+ src += 8;
+ dst1 += linesize;
+ dst2 += linesize;
+ }
+}
+
#if 0
#define PIXOP2(OPNAME, OP) \
#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;
#if 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);
}
}
#endif /* CONFIG_CAVS_DECODER */
-void ff_mlp_init(DSPContext* c, AVCodecContext *avctx);
-
-#if CONFIG_VC1_DECODER || CONFIG_WMV3_DECODER
+#if CONFIG_VC1_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) {
+void ff_put_vc1_mspel_mc00_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) {
put_pixels8_c(dst, src, stride, 8);
}
-void ff_avg_vc1_mspel_mc00_c(uint8_t *dst, 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) {
avg_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 CONFIG_RV30_DECODER
-void ff_rv30dsp_init(DSPContext* c, AVCodecContext *avctx);
-#endif /* CONFIG_RV30_DECODER */
+#endif /* CONFIG_VC1_DECODER */
#if CONFIG_RV40_DECODER
static void put_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride){
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){
}
static void h263_v_loop_filter_c(uint8_t *src, int stride, int qscale){
- if(CONFIG_ANY_H263) {
+ if(CONFIG_H263_DECODER || CONFIG_H263_ENCODER) {
int x;
const int strength= ff_h263_loop_filter_strength[qscale];
}
static void h263_h_loop_filter_c(uint8_t *src, int stride, int qscale){
- if(CONFIG_ANY_H263) {
+ if(CONFIG_H263_DECODER || CONFIG_H263_ENCODER) {
int y;
const int strength= ff_h263_loop_filter_strength[qscale];
}
}
-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;
case FF_CMP_NSSE:
cmp[i]= c->nsse[i];
break;
-#if CONFIG_SNOW_ENCODER
+#if CONFIG_DWT
case FF_CMP_W53:
cmp[i]= c->w53[i];
break;
dst[i+0] = src1[i+0]-src2[i+0];
}
-static void add_hfyu_median_prediction_c(uint8_t *dst, uint8_t *src1, uint8_t *diff, int w, int *left, int *left_top){
+static void add_hfyu_median_prediction_c(uint8_t *dst, const uint8_t *src1, const uint8_t *diff, int w, int *left, int *left_top){
int i;
uint8_t l, lt;
*left_top= lt;
}
-static void sub_hfyu_median_prediction_c(uint8_t *dst, uint8_t *src1, uint8_t *src2, int w, int *left, int *left_top){
+static void sub_hfyu_median_prediction_c(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int w, int *left, int *left_top){
int i;
uint8_t l, lt;
*left_top= lt;
}
+static int add_hfyu_left_prediction_c(uint8_t *dst, const uint8_t *src, int w, int acc){
+ int i;
+
+ for(i=0; i<w-1; i++){
+ acc+= src[i];
+ dst[i]= acc;
+ i++;
+ acc+= src[i];
+ dst[i]= acc;
+ }
+
+ for(; i<w; i++){
+ acc+= src[i];
+ dst[i]= acc;
+ }
+
+ return acc;
+}
+
+#if HAVE_BIGENDIAN
+#define B 3
+#define G 2
+#define R 1
+#define A 0
+#else
+#define B 0
+#define G 1
+#define R 2
+#define A 3
+#endif
+static void add_hfyu_left_prediction_bgr32_c(uint8_t *dst, const uint8_t *src, int w, int *red, int *green, int *blue, int *alpha){
+ int i;
+ int r,g,b,a;
+ r= *red;
+ g= *green;
+ b= *blue;
+ a= *alpha;
+
+ for(i=0; i<w; i++){
+ b+= src[4*i+B];
+ g+= src[4*i+G];
+ r+= src[4*i+R];
+ a+= src[4*i+A];
+
+ dst[4*i+B]= b;
+ dst[4*i+G]= g;
+ dst[4*i+R]= r;
+ dst[4*i+A]= a;
+ }
+
+ *red= r;
+ *green= g;
+ *blue= b;
+ *alpha= a;
+}
+#undef B
+#undef G
+#undef R
+#undef A
+
#define BUTTERFLY2(o1,o2,i1,i2) \
o1= (i1)+(i2);\
o2= (i1)-(i2);
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;
+ LOCAL_ALIGNED_16(DCTELEM, temp, [64]);
assert(h==8);
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;
+ LOCAL_ALIGNED_16(DCTELEM, temp, [64]);
int sum=0, i;
assert(h==8);
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;
+ LOCAL_ALIGNED_16(DCTELEM, temp, [64*2]);
+ DCTELEM * const bak = temp+64;
int sum=0, i;
assert(h==8);
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]);
- DCTELEM * const temp= (DCTELEM*)aligned_temp;
- uint8_t * const bak= (uint8_t*)aligned_bak;
+ LOCAL_ALIGNED_16(DCTELEM, temp, [64]);
+ LOCAL_ALIGNED_16(uint8_t, lsrc1, [64]);
+ LOCAL_ALIGNED_16(uint8_t, lsrc2, [64]);
int i, last, run, bits, level, distortion, start_i;
const int esc_length= s->ac_esc_length;
uint8_t * 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];
- }
+ copy_block8(lsrc1, src1, 8, stride, 8);
+ copy_block8(lsrc2, src2, 8, stride, 8);
- s->dsp.diff_pixels(temp, src1, src2, stride);
+ s->dsp.diff_pixels(temp, lsrc1, lsrc2, 8);
s->block_last_index[0/*FIXME*/]= last= s->fast_dct_quantize(s, temp, 0/*FIXME*/, s->qscale, &i);
s->dct_unquantize_inter(s, temp, 0, s->qscale);
}
- s->dsp.idct_add(bak, stride, temp);
+ s->dsp.idct_add(lsrc2, 8, temp);
- distortion= s->dsp.sse[1](NULL, bak, src1, stride, 8);
+ distortion= s->dsp.sse[1](NULL, lsrc2, lsrc1, 8, 8);
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;
+ LOCAL_ALIGNED_16(DCTELEM, temp, [64]);
int i, last, run, bits, level, start_i;
const int esc_length= s->ac_esc_length;
uint8_t * length;
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){
+static void vector_fmul_add_c(float *dst, const float *src0, const float *src1, const float *src2, int len){
int i;
for(i=0; i<len; i++)
- dst[i*step] = src0[i] * src1[i] + src2[i] + src3;
+ dst[i] = src0[i] * src1[i] + src2[i];
}
void ff_vector_fmul_window_c(float *dst, const float *src0, const float *src1, const float *win, float add_bias, int len){
}
}
+static void vector_fmul_scalar_c(float *dst, const float *src, float mul,
+ int len)
+{
+ int i;
+ for (i = 0; i < len; i++)
+ dst[i] = src[i] * mul;
+}
+
+static void vector_fmul_sv_scalar_2_c(float *dst, const float *src,
+ const float **sv, float mul, int len)
+{
+ int i;
+ for (i = 0; i < len; i += 2, sv++) {
+ dst[i ] = src[i ] * sv[0][0] * mul;
+ dst[i+1] = src[i+1] * sv[0][1] * mul;
+ }
+}
+
+static void vector_fmul_sv_scalar_4_c(float *dst, const float *src,
+ const float **sv, float mul, int len)
+{
+ int i;
+ for (i = 0; i < len; i += 4, sv++) {
+ dst[i ] = src[i ] * sv[0][0] * mul;
+ dst[i+1] = src[i+1] * sv[0][1] * mul;
+ dst[i+2] = src[i+2] * sv[0][2] * mul;
+ dst[i+3] = src[i+3] * sv[0][3] * mul;
+ }
+}
+
+static void sv_fmul_scalar_2_c(float *dst, const float **sv, float mul,
+ int len)
+{
+ int i;
+ for (i = 0; i < len; i += 2, sv++) {
+ dst[i ] = sv[0][0] * mul;
+ dst[i+1] = sv[0][1] * mul;
+ }
+}
+
+static void sv_fmul_scalar_4_c(float *dst, const float **sv, float mul,
+ int len)
+{
+ int i;
+ for (i = 0; i < len; i += 4, sv++) {
+ dst[i ] = sv[0][0] * mul;
+ dst[i+1] = sv[0][1] * mul;
+ dst[i+2] = sv[0][2] * mul;
+ dst[i+3] = sv[0][3] * mul;
+ }
+}
+
+static void butterflies_float_c(float *restrict v1, float *restrict v2,
+ int len)
+{
+ int i;
+ for (i = 0; i < len; i++) {
+ float t = v1[i] - v2[i];
+ v1[i] += v2[i];
+ v2[i] = t;
+ }
+}
+
+static float scalarproduct_float_c(const float *v1, const float *v2, int len)
+{
+ float p = 0.0;
+ int i;
+
+ for (i = 0; i < len; i++)
+ p += v1[i] * v2[i];
+
+ return p;
+}
+
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 inline uint32_t clipf_c_one(uint32_t a, uint32_t mini,
+ uint32_t maxi, uint32_t maxisign)
+{
+
+ if(a > mini) return mini;
+ else if((a^(1<<31)) > maxisign) return maxi;
+ else return a;
+}
+
+static void vector_clipf_c_opposite_sign(float *dst, const float *src, float *min, float *max, int len){
+ int i;
+ uint32_t mini = *(uint32_t*)min;
+ uint32_t maxi = *(uint32_t*)max;
+ uint32_t maxisign = maxi ^ (1<<31);
+ uint32_t *dsti = (uint32_t*)dst;
+ const uint32_t *srci = (const uint32_t*)src;
+ for(i=0; i<len; i+=8) {
+ dsti[i + 0] = clipf_c_one(srci[i + 0], mini, maxi, maxisign);
+ dsti[i + 1] = clipf_c_one(srci[i + 1], mini, maxi, maxisign);
+ dsti[i + 2] = clipf_c_one(srci[i + 2], mini, maxi, maxisign);
+ dsti[i + 3] = clipf_c_one(srci[i + 3], mini, maxi, maxisign);
+ dsti[i + 4] = clipf_c_one(srci[i + 4], mini, maxi, maxisign);
+ dsti[i + 5] = clipf_c_one(srci[i + 5], mini, maxi, maxisign);
+ dsti[i + 6] = clipf_c_one(srci[i + 6], mini, maxi, maxisign);
+ dsti[i + 7] = clipf_c_one(srci[i + 7], mini, maxi, maxisign);
+ }
+}
+static void vector_clipf_c(float *dst, const float *src, float min, float max, int len){
+ int i;
+ if(min < 0 && max > 0) {
+ vector_clipf_c_opposite_sign(dst, src, &min, &max, len);
+ } else {
+ for(i=0; i < len; i+=8) {
+ dst[i ] = av_clipf(src[i ], min, max);
+ dst[i + 1] = av_clipf(src[i + 1], min, max);
+ dst[i + 2] = av_clipf(src[i + 2], min, max);
+ dst[i + 3] = av_clipf(src[i + 3], min, max);
+ dst[i + 4] = av_clipf(src[i + 4], min, max);
+ dst[i + 5] = av_clipf(src[i + 5], min, max);
+ dst[i + 6] = av_clipf(src[i + 6], min, max);
+ dst[i + 7] = av_clipf(src[i + 7], min, max);
+ }
+ }
+}
+
static av_always_inline int float_to_int16_one(const float *src){
int_fast32_t tmp = *(const int32_t*)src;
if(tmp & 0xf0000){
}
}
-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;
return res;
}
+static int32_t scalarproduct_and_madd_int16_c(int16_t *v1, int16_t *v2, int16_t *v3, int order, int mul)
+{
+ int res = 0;
+ while (order--) {
+ res += *v1 * *v2++;
+ *v1++ += mul * *v3++;
+ }
+ 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) */
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)
+av_cold void dsputil_static_init(void)
{
int i;
int ff_check_alignment(void){
static int did_fail=0;
- DECLARE_ALIGNED_16(int, aligned);
+ DECLARE_ALIGNED(16, int, aligned);
if((intptr_t)&aligned & 15){
if(!did_fail){
return 0;
}
-void dsputil_init(DSPContext* c, AVCodecContext *avctx)
+av_cold void dsputil_init(DSPContext* c, AVCodecContext *avctx)
{
int i;
c->idct_add= ff_jref_idct_add;
c->idct = j_rev_dct;
c->idct_permutation_type= FF_LIBMPEG2_IDCT_PERM;
- }else if((CONFIG_VP3_DECODER || CONFIG_VP5_DECODER || CONFIG_VP6_DECODER || CONFIG_THEORA_DECODER ) &&
+ }else if((CONFIG_VP3_DECODER || CONFIG_VP5_DECODER || CONFIG_VP6_DECODER ) &&
avctx->idct_algo==FF_IDCT_VP3){
c->idct_put= ff_vp3_idct_put_c;
c->idct_add= ff_vp3_idct_add_c;
}else if(CONFIG_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 if(CONFIG_BINK_DECODER && avctx->idct_algo==FF_IDCT_BINK) {
+ c->idct = ff_bink_idct_c;
+ c->idct_add = ff_bink_idct_add_c;
+ c->idct_put = ff_bink_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;
}
}
- if (CONFIG_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->put_pixels_nonclamped = put_pixels_nonclamped_c;
c->add_pixels_clamped = add_pixels_clamped_c;
c->add_pixels8 = add_pixels8_c;
c->add_pixels4 = add_pixels4_c;
c->pix_sum = pix_sum_c;
c->pix_norm1 = pix_norm1_c;
+ c->fill_block_tab[0] = fill_block16_c;
+ c->fill_block_tab[1] = fill_block8_c;
+ c->scale_block = scale_block_c;
+
/* TODO [0] 16 [1] 8 */
c->pix_abs[0][0] = pix_abs16_c;
c->pix_abs[0][1] = pix_abs16_x2_c;
c->put_no_rnd_vc1_chroma_pixels_tab[0]= put_no_rnd_vc1_chroma_mc8_c;
c->avg_no_rnd_vc1_chroma_pixels_tab[0]= avg_no_rnd_vc1_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;
#if CONFIG_CAVS_DECODER
#if CONFIG_MLP_DECODER || CONFIG_TRUEHD_DECODER
ff_mlp_init(c, avctx);
#endif
-#if CONFIG_VC1_DECODER || CONFIG_WMV3_DECODER
+#if CONFIG_VC1_DECODER
ff_vc1dsp_init(c,avctx);
#endif
-#if CONFIG_WMV2_DECODER || CONFIG_VC1_DECODER || CONFIG_WMV3_DECODER
+#if CONFIG_WMV2_DECODER || CONFIG_VC1_DECODER
ff_intrax8dsp_init(c,avctx);
#endif
#if CONFIG_RV30_DECODER
c->vsse[5]= vsse_intra8_c;
c->nsse[0]= nsse16_c;
c->nsse[1]= nsse8_c;
-#if 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;
+#if CONFIG_DWT
+ ff_dsputil_init_dwt(c);
#endif
c->ssd_int8_vs_int16 = ssd_int8_vs_int16_c;
c->diff_bytes= diff_bytes_c;
c->add_hfyu_median_prediction= add_hfyu_median_prediction_c;
c->sub_hfyu_median_prediction= sub_hfyu_median_prediction_c;
+ c->add_hfyu_left_prediction = add_hfyu_left_prediction_c;
+ c->add_hfyu_left_prediction_bgr32 = add_hfyu_left_prediction_bgr32_c;
c->bswap_buf= bswap_buf;
#if 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 (CONFIG_ANY_H263) {
+ if (CONFIG_H263_DECODER || CONFIG_H263_ENCODER) {
c->h263_h_loop_filter= h263_h_loop_filter_c;
c->h263_v_loop_filter= h263_v_loop_filter_c;
}
- if (CONFIG_VP3_DECODER || CONFIG_THEORA_DECODER) {
+ if (CONFIG_VP3_DECODER) {
c->vp3_h_loop_filter= ff_vp3_h_loop_filter_c;
c->vp3_v_loop_filter= ff_vp3_v_loop_filter_c;
}
c->try_8x8basis= try_8x8basis_c;
c->add_8x8basis= add_8x8basis_c;
-#if 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
-
#if CONFIG_VORBIS_DECODER
c->vorbis_inverse_coupling = vorbis_inverse_coupling;
#endif
#if CONFIG_AC3_DECODER
c->ac3_downmix = ff_ac3_downmix_c;
#endif
-#if CONFIG_FLAC_ENCODER
- c->flac_compute_autocorr = ff_flac_compute_autocorr;
+#if CONFIG_LPC
+ c->lpc_compute_autocorr = ff_lpc_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_add = vector_fmul_add_c;
c->vector_fmul_window = ff_vector_fmul_window_c;
c->int32_to_float_fmul_scalar = int32_to_float_fmul_scalar_c;
+ c->vector_clipf = vector_clipf_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->scalarproduct_and_madd_int16 = scalarproduct_and_madd_int16_c;
+ c->scalarproduct_float = scalarproduct_float_c;
+ c->butterflies_float = butterflies_float_c;
+ c->vector_fmul_scalar = vector_fmul_scalar_c;
+
+ c->vector_fmul_sv_scalar[0] = vector_fmul_sv_scalar_2_c;
+ c->vector_fmul_sv_scalar[1] = vector_fmul_sv_scalar_4_c;
+
+ c->sv_fmul_scalar[0] = sv_fmul_scalar_2_c;
+ c->sv_fmul_scalar[1] = sv_fmul_scalar_4_c;
c->shrink[0]= ff_img_copy_plane;
c->shrink[1]= ff_shrink22;
projects / ffmpeg / blobdiff