MrBesen/ffmpeg
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Merge commit 'da0c8664b4dc906696803685f7e53ade68594ab8'

Browse Source 
* commit 'da0c8664b4dc906696803685f7e53ade68594ab8':
  mpegvideo: Move various temporary buffers to a separate context

Conflicts:
	libavcodec/mpegvideo.c
	libavcodec/mpegvideo_enc.c
	libavcodec/mpegvideo_motion.c
	libavcodec/rv34.c
	libavcodec/vc1_mc.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>
This commit is contained in:
Michael Niedermayer 2015-06-08 20:58:13 +02:00
commit db8ae37a78
10 changed files with 107 additions and 101 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
libavcodec/intrax8.c
View File

@ -303,7 +303,7 @@ static int x8_setup_spatial_predictor(IntraX8Context * const w, const int chroma
int sum;
int quant;
w->dsp.setup_spatial_compensation(s->dest[chroma], s->edge_emu_buffer,
w->dsp.setup_spatial_compensation(s->dest[chroma], s->sc.edge_emu_buffer,
s->current_picture.f->linesize[chroma>0],
&range, &sum, w->edges);
if(chroma){
@ -639,7 +639,7 @@ static int x8_decode_intra_mb(IntraX8Context* const w, const int chroma){
if(w->flat_dc){
dsp_x8_put_solidcolor(w->predicted_dc, s->dest[chroma], s->current_picture.f->linesize[!!chroma]);
}else{
w->dsp.spatial_compensation[w->orient]( s->edge_emu_buffer,
w->dsp.spatial_compensation[w->orient]( s->sc.edge_emu_buffer,
s->dest[chroma],
s->current_picture.f->linesize[!!chroma] );
}

67
libavcodec/mpegvideo.c
View File

@ -364,6 +364,7 @@ av_cold void ff_mpv_idct_init(MpegEncContext *s)
static int frame_size_alloc(MpegEncContext *s, int linesize)
{
int alloc_size = FFALIGN(FFABS(linesize) + 64, 32);
ScratchpadContext *sc = &s->sc;
if (s->avctx->hwaccel || s->avctx->codec->capabilities & CODEC_CAP_HWACCEL_VDPAU)
return 0;
@ -379,19 +380,19 @@ static int frame_size_alloc(MpegEncContext *s, int linesize)
// at uvlinesize. It supports only YUV420 so 24x24 is enough
// linesize * interlaced * MBsize
// we also use this buffer for encoding in encode_mb_internal() needig an additional 32 lines
FF_ALLOCZ_ARRAY_OR_GOTO(s->avctx, s->edge_emu_buffer, alloc_size, 4 * 68,
FF_ALLOCZ_ARRAY_OR_GOTO(s->avctx, s->sc.edge_emu_buffer, alloc_size, 4 * 68,
fail);
FF_ALLOCZ_ARRAY_OR_GOTO(s->avctx, s->me.scratchpad, alloc_size, 4 * 16 * 2,
fail)
s->me.temp = s->me.scratchpad;
s->rd_scratchpad = s->me.scratchpad;
s->b_scratchpad = s->me.scratchpad;
s->obmc_scratchpad = s->me.scratchpad + 16;
s->me.temp = s->me.scratchpad;
sc->rd_scratchpad = s->me.scratchpad;
sc->b_scratchpad = s->me.scratchpad;
sc->obmc_scratchpad = s->me.scratchpad + 16;
return 0;
fail:
av_freep(&s->edge_emu_buffer);
av_freep(&sc->edge_emu_buffer);
return AVERROR(ENOMEM);
}
@ -466,7 +467,7 @@ static int alloc_frame_buffer(MpegEncContext *s, Picture *pic)
return -1;
}
if (!s->edge_emu_buffer &&
if (!s->sc.edge_emu_buffer &&
(ret = frame_size_alloc(s, pic->f->linesize[0])) < 0) {
av_log(s->avctx, AV_LOG_ERROR,
"get_buffer() failed to allocate context scratch buffers.\n");
@ -740,12 +741,12 @@ static int init_duplicate_context(MpegEncContext *s)
if (s->mb_height & 1)
yc_size += 2*s->b8_stride + 2*s->mb_stride;
s->edge_emu_buffer =
s->sc.edge_emu_buffer =
s->me.scratchpad =
s->me.temp =
s->rd_scratchpad =
s->b_scratchpad =
s->obmc_scratchpad = NULL;
s->sc.rd_scratchpad =
s->sc.b_scratchpad =
s->sc.obmc_scratchpad = NULL;
if (s->encoding) {
FF_ALLOCZ_OR_GOTO(s->avctx, s->me.map,
@ -787,12 +788,12 @@ static void free_duplicate_context(MpegEncContext *s)
if (!s)
return;
av_freep(&s->edge_emu_buffer);
av_freep(&s->sc.edge_emu_buffer);
av_freep(&s->me.scratchpad);
s->me.temp =
s->rd_scratchpad =
s->b_scratchpad =
s->obmc_scratchpad = NULL;
s->sc.rd_scratchpad =
s->sc.b_scratchpad =
s->sc.obmc_scratchpad = NULL;
av_freep(&s->dct_error_sum);
av_freep(&s->me.map);
@ -805,12 +806,12 @@ static void free_duplicate_context(MpegEncContext *s)
static void backup_duplicate_context(MpegEncContext *bak, MpegEncContext *src)
{
#define COPY(a) bak->a = src->a
COPY(edge_emu_buffer);
COPY(sc.edge_emu_buffer);
COPY(me.scratchpad);
COPY(me.temp);
COPY(rd_scratchpad);
COPY(b_scratchpad);
COPY(obmc_scratchpad);
COPY(sc.rd_scratchpad);
COPY(sc.b_scratchpad);
COPY(sc.obmc_scratchpad);
COPY(me.map);
COPY(me.score_map);
COPY(blocks);
@ -845,7 +846,7 @@ int ff_update_duplicate_context(MpegEncContext *dst, MpegEncContext *src)
// exchange uv
FFSWAP(void *, dst->pblocks[4], dst->pblocks[5]);
}
if (!dst->edge_emu_buffer &&
if (!dst->sc.edge_emu_buffer &&
(ret = frame_size_alloc(dst, dst->linesize)) < 0) {
av_log(dst->avctx, AV_LOG_ERROR, "failed to allocate context "
"scratch buffers.\n");
@ -975,7 +976,7 @@ do {\
}
// linesize dependend scratch buffer allocation
if (!s->edge_emu_buffer)
if (!s->sc.edge_emu_buffer)
if (s1->linesize) {
if (frame_size_alloc(s, s1->linesize) < 0) {
av_log(s->avctx, AV_LOG_ERROR, "Failed to allocate context "
@ -2393,12 +2394,12 @@ static inline int hpel_motion_lowres(MpegEncContext *s,
if ((unsigned)src_x > FFMAX( h_edge_pos - (!!sx) - w, 0) ||
(unsigned)src_y > FFMAX((v_edge_pos >> field_based) - (!!sy) - h, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, src,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, src,
s->linesize, s->linesize,
w + 1, (h + 1) << field_based,
src_x, src_y << field_based,
h_edge_pos, v_edge_pos);
src = s->edge_emu_buffer;
src = s->sc.edge_emu_buffer;
emu = 1;
}
@ -2495,14 +2496,14 @@ static av_always_inline void mpeg_motion_lowres(MpegEncContext *s,
if ((unsigned) src_x > FFMAX( h_edge_pos - (!!sx) - 2 * block_s, 0) || uvsrc_y<0 ||
(unsigned) src_y > FFMAX((v_edge_pos >> field_based) - (!!sy) - h, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr_y,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, ptr_y,
linesize >> field_based, linesize >> field_based,
17, 17 + field_based,
src_x, src_y << field_based, h_edge_pos,
v_edge_pos);
ptr_y = s->edge_emu_buffer;
ptr_y = s->sc.edge_emu_buffer;
if (!CONFIG_GRAY || !(s->avctx->flags & CODEC_FLAG_GRAY)) {
uint8_t *ubuf = s->edge_emu_buffer + 18 * s->linesize;
uint8_t *ubuf = s->sc.edge_emu_buffer + 18 * s->linesize;
uint8_t *vbuf =ubuf + 9 * s->uvlinesize;
s->vdsp.emulated_edge_mc(ubuf, ptr_cb,
uvlinesize >> field_based, uvlinesize >> field_based,
@ -2583,11 +2584,11 @@ static inline void chroma_4mv_motion_lowres(MpegEncContext *s,
ptr = ref_picture[1] + offset;
if ((unsigned) src_x > FFMAX(h_edge_pos - (!!sx) - block_s, 0) ||
(unsigned) src_y > FFMAX(v_edge_pos - (!!sy) - block_s, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, ptr,
s->uvlinesize, s->uvlinesize,
9, 9,
src_x, src_y, h_edge_pos, v_edge_pos);
ptr = s->edge_emu_buffer;
ptr = s->sc.edge_emu_buffer;
emu = 1;
}
sx = (sx << 2) >> lowres;
@ -2596,11 +2597,11 @@ static inline void chroma_4mv_motion_lowres(MpegEncContext *s,
ptr = ref_picture[2] + offset;
if (emu) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, ptr,
s->uvlinesize, s->uvlinesize,
9, 9,
src_x, src_y, h_edge_pos, v_edge_pos);
ptr = s->edge_emu_buffer;
ptr = s->sc.edge_emu_buffer;
}
pix_op[op_index](dest_cr, ptr, s->uvlinesize, block_s, sx, sy);
}
@ -2932,9 +2933,9 @@ void mpv_decode_mb_internal(MpegEncContext *s, int16_t block[12][64],
dest_cb= s->dest[1];
dest_cr= s->dest[2];
}else{
dest_y = s->b_scratchpad;
dest_cb= s->b_scratchpad+16*linesize;
dest_cr= s->b_scratchpad+32*linesize;
dest_y = s->sc.b_scratchpad;
dest_cb= s->sc.b_scratchpad+16*linesize;
dest_cr= s->sc.b_scratchpad+32*linesize;
}
if (!s->mb_intra) {

13
libavcodec/mpegvideo.h
View File

@ -143,6 +143,13 @@ typedef struct Picture{
uint64_t error[AV_NUM_DATA_POINTERS];
} Picture;
typedef struct ScratchpadContext {
uint8_t *edge_emu_buffer; ///< temporary buffer for if MVs point to out-of-frame data
uint8_t *rd_scratchpad; ///< scratchpad for rate distortion mb decision
uint8_t *obmc_scratchpad;
uint8_t *b_scratchpad; ///< scratchpad used for writing into write only buffers
} ScratchpadContext;
/**
* MpegEncContext.
*/
@ -266,10 +273,8 @@ typedef struct MpegEncContext {
uint8_t *mbintra_table; ///< used to avoid setting {ac, dc, cbp}-pred stuff to zero on inter MB decoding
uint8_t *cbp_table; ///< used to store cbp, ac_pred for partitioned decoding
uint8_t *pred_dir_table; ///< used to store pred_dir for partitioned decoding
uint8_t *edge_emu_buffer; ///< temporary buffer for if MVs point to out-of-frame data
uint8_t *rd_scratchpad; ///< scratchpad for rate distortion mb decision
uint8_t *obmc_scratchpad;
uint8_t *b_scratchpad; ///< scratchpad used for writing into write only buffers
ScratchpadContext sc;
int qscale; ///< QP
int chroma_qscale; ///< chroma QP

14
libavcodec/mpegvideo_enc.c
View File

@ -2100,7 +2100,7 @@ static av_always_inline void encode_mb_internal(MpegEncContext *s,
(mb_y * mb_block_height * wrap_c) + mb_x * mb_block_width;
if((mb_x * 16 + 16 > s->width || mb_y * 16 + 16 > s->height) && s->codec_id != AV_CODEC_ID_AMV){
uint8_t *ebuf = s->edge_emu_buffer + 36 * wrap_y;
uint8_t *ebuf = s->sc.edge_emu_buffer + 36 * wrap_y;
int cw = (s->width + s->chroma_x_shift) >> s->chroma_x_shift;
int ch = (s->height + s->chroma_y_shift) >> s->chroma_y_shift;
s->vdsp.emulated_edge_mc(ebuf, ptr_y,
@ -2512,9 +2512,9 @@ static inline void encode_mb_hq(MpegEncContext *s, MpegEncContext *backup, MpegE
if(*next_block){
memcpy(dest_backup, s->dest, sizeof(s->dest));
s->dest[0] = s->rd_scratchpad;
s->dest[1] = s->rd_scratchpad + 16*s->linesize;
s->dest[2] = s->rd_scratchpad + 16*s->linesize + 8;
s->dest[0] = s->sc.rd_scratchpad;
s->dest[1] = s->sc.rd_scratchpad + 16*s->linesize;
s->dest[2] = s->sc.rd_scratchpad + 16*s->linesize + 8;
av_assert0(s->linesize >= 32); //FIXME
}
@ -3213,9 +3213,9 @@ static int encode_thread(AVCodecContext *c, void *arg){
ff_h263_update_motion_val(s);
if(next_block==0){ //FIXME 16 vs linesize16
s->hdsp.put_pixels_tab[0][0](s->dest[0], s->rd_scratchpad , s->linesize ,16);
s->hdsp.put_pixels_tab[1][0](s->dest[1], s->rd_scratchpad + 16*s->linesize , s->uvlinesize, 8);
s->hdsp.put_pixels_tab[1][0](s->dest[2], s->rd_scratchpad + 16*s->linesize + 8, s->uvlinesize, 8);
s->hdsp.put_pixels_tab[0][0](s->dest[0], s->sc.rd_scratchpad , s->linesize ,16);
s->hdsp.put_pixels_tab[1][0](s->dest[1], s->sc.rd_scratchpad + 16*s->linesize , s->uvlinesize, 8);
s->hdsp.put_pixels_tab[1][0](s->dest[2], s->sc.rd_scratchpad + 16*s->linesize + 8, s->uvlinesize, 8);
}
if(s->avctx->mb_decision == FF_MB_DECISION_BITS)

42
libavcodec/mpegvideo_motion.c
View File

@ -64,12 +64,12 @@ static void gmc1_motion(MpegEncContext *s,
if ((unsigned)src_x >= FFMAX(s->h_edge_pos - 17, 0) ||
(unsigned)src_y >= FFMAX(s->v_edge_pos - 17, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, ptr,
linesize, linesize,
17, 17,
src_x, src_y,
s->h_edge_pos, s->v_edge_pos);
ptr = s->edge_emu_buffer;
ptr = s->sc.edge_emu_buffer;
}
if ((motion_x | motion_y) & 7) {
@ -108,12 +108,12 @@ static void gmc1_motion(MpegEncContext *s,
ptr = ref_picture[1] + offset;
if ((unsigned)src_x >= FFMAX((s->h_edge_pos >> 1) - 9, 0) ||
(unsigned)src_y >= FFMAX((s->v_edge_pos >> 1) - 9, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, ptr,
uvlinesize, uvlinesize,
9, 9,
src_x, src_y,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
ptr = s->edge_emu_buffer;
ptr = s->sc.edge_emu_buffer;
emu = 1;
}
s->mdsp.gmc1(dest_cb, ptr, uvlinesize, 8,
@ -121,12 +121,12 @@ static void gmc1_motion(MpegEncContext *s,
ptr = ref_picture[2] + offset;
if (emu) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, ptr,
uvlinesize, uvlinesize,
9, 9,
src_x, src_y,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
ptr = s->edge_emu_buffer;
ptr = s->sc.edge_emu_buffer;
}
s->mdsp.gmc1(dest_cr, ptr, uvlinesize, 8,
motion_x & 15, motion_y & 15, 128 - s->no_rounding);
@ -213,12 +213,12 @@ static inline int hpel_motion(MpegEncContext *s,
if ((unsigned)src_x >= FFMAX(s->h_edge_pos - (motion_x & 1) - 7, 0) ||
(unsigned)src_y >= FFMAX(s->v_edge_pos - (motion_y & 1) - 7, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, src,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, src,
s->linesize, s->linesize,
9, 9,
src_x, src_y,
s->h_edge_pos, s->v_edge_pos);
src = s->edge_emu_buffer;
src = s->sc.edge_emu_buffer;
emu = 1;
}
pix_op[dxy](dest, src, s->linesize, 8);
@ -318,14 +318,14 @@ void mpeg_motion_internal(MpegEncContext *s,
return;
}
src_y = (unsigned)src_y << field_based;
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr_y,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, ptr_y,
s->linesize, s->linesize,
17, 17 + field_based,
src_x, src_y,
s->h_edge_pos, s->v_edge_pos);
ptr_y = s->edge_emu_buffer;
ptr_y = s->sc.edge_emu_buffer;
if (!CONFIG_GRAY || !(s->avctx->flags & CODEC_FLAG_GRAY)) {
uint8_t *ubuf = s->edge_emu_buffer + 18 * s->linesize;
uint8_t *ubuf = s->sc.edge_emu_buffer + 18 * s->linesize;
uint8_t *vbuf = ubuf + 9 * s->uvlinesize;
uvsrc_y = (unsigned)uvsrc_y << field_based;
s->vdsp.emulated_edge_mc(ubuf, ptr_cb,
@ -478,7 +478,7 @@ static inline void obmc_motion(MpegEncContext *s,
if (i && mv[i][0] == mv[MID][0] && mv[i][1] == mv[MID][1]) {
ptr[i] = ptr[MID];
} else {
ptr[i] = s->obmc_scratchpad + 8 * (i & 1) +
ptr[i] = s->sc.obmc_scratchpad + 8 * (i & 1) +
s->linesize * 8 * (i >> 1);
hpel_motion(s, ptr[i], src, src_x, src_y, pix_op,
mv[i][0], mv[i][1]);
@ -541,14 +541,14 @@ static inline void qpel_motion(MpegEncContext *s,
if ((unsigned)src_x >= FFMAX(s->h_edge_pos - (motion_x & 3) - 15 , 0) ||
(unsigned)src_y >= FFMAX( v_edge_pos - (motion_y & 3) - h + 1, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr_y,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, ptr_y,
s->linesize, s->linesize,
17, 17 + field_based,
src_x, src_y << field_based,
s->h_edge_pos, s->v_edge_pos);
ptr_y = s->edge_emu_buffer;
ptr_y = s->sc.edge_emu_buffer;
if (!CONFIG_GRAY || !(s->avctx->flags & CODEC_FLAG_GRAY)) {
uint8_t *ubuf = s->edge_emu_buffer + 18 * s->linesize;
uint8_t *ubuf = s->sc.edge_emu_buffer + 18 * s->linesize;
uint8_t *vbuf = ubuf + 9 * s->uvlinesize;
s->vdsp.emulated_edge_mc(ubuf, ptr_cb,
s->uvlinesize, s->uvlinesize,
@ -625,22 +625,22 @@ static void chroma_4mv_motion(MpegEncContext *s,
ptr = ref_picture[1] + offset;
if ((unsigned)src_x >= FFMAX((s->h_edge_pos >> 1) - (dxy & 1) - 7, 0) ||
(unsigned)src_y >= FFMAX((s->v_edge_pos >> 1) - (dxy >> 1) - 7, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, ptr,
s->uvlinesize, s->uvlinesize,
9, 9, src_x, src_y,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
ptr = s->edge_emu_buffer;
ptr = s->sc.edge_emu_buffer;
emu = 1;
}
pix_op[dxy](dest_cb, ptr, s->uvlinesize, 8);
ptr = ref_picture[2] + offset;
if (emu) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, ptr,
s->uvlinesize, s->uvlinesize,
9, 9, src_x, src_y,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
ptr = s->edge_emu_buffer;
ptr = s->sc.edge_emu_buffer;
}
pix_op[dxy](dest_cr, ptr, s->uvlinesize, 8);
}
@ -783,13 +783,13 @@ static inline void apply_8x8(MpegEncContext *s,
ptr = ref_picture[0] + (src_y * s->linesize) + (src_x);
if ((unsigned)src_x >= FFMAX(s->h_edge_pos - (motion_x & 3) - 7, 0) ||
(unsigned)src_y >= FFMAX(s->v_edge_pos - (motion_y & 3) - 7, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, ptr,
s->linesize, s->linesize,
9, 9,
src_x, src_y,
s->h_edge_pos,
s->v_edge_pos);
ptr = s->edge_emu_buffer;
ptr = s->sc.edge_emu_buffer;
}
dest = dest_y + ((i & 1) * 8) + (i >> 1) * 8 * s->linesize;
qpix_op[1][dxy](dest, ptr, s->linesize);

6
libavcodec/rv34.c
View File

@ -726,12 +726,12 @@ static inline void rv34_mc(RV34DecContext *r, const int block_type,
(unsigned)(src_x - !!lx*2) > s->h_edge_pos - !!lx*2 - (width <<3) - 4 ||
(unsigned)(src_y - !!ly*2) > s->v_edge_pos - !!ly*2 - (height<<3) - 4) {
srcY -= 2 + 2*s->linesize;
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcY,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, srcY,
s->linesize, s->linesize,
(width << 3) + 6, (height << 3) + 6,
src_x - 2, src_y - 2,
s->h_edge_pos, s->v_edge_pos);
srcY = s->edge_emu_buffer + 2 + 2*s->linesize;
srcY = s->sc.edge_emu_buffer + 2 + 2*s->linesize;
emu = 1;
}
if(!weighted){
@ -756,7 +756,7 @@ static inline void rv34_mc(RV34DecContext *r, const int block_type,
is16x16 = (block_type != RV34_MB_P_8x8) && (block_type != RV34_MB_P_16x8) && (block_type != RV34_MB_P_8x16);
qpel_mc[!is16x16][dxy](Y, srcY, s->linesize);
if (emu) {
uint8_t *uvbuf = s->edge_emu_buffer;
uint8_t *uvbuf = s->sc.edge_emu_buffer;
s->vdsp.emulated_edge_mc(uvbuf, srcU,
s->uvlinesize, s->uvlinesize,

2
libavcodec/snow.c
View File

@ -704,7 +704,7 @@ av_cold void ff_snow_common_end(SnowContext *s)
av_freep(&s->m.me.scratchpad);
av_freep(&s->m.me.map);
av_freep(&s->m.me.score_map);
av_freep(&s->m.obmc_scratchpad);
av_freep(&s->m.sc.obmc_scratchpad);
av_freep(&s->block);
av_freep(&s->scratchbuf);

8
libavcodec/snowenc.c
View File

@ -75,8 +75,8 @@ static av_cold int encode_init(AVCodecContext *avctx)
s->m.me.scratchpad= av_mallocz_array((avctx->width+64), 2*16*2*sizeof(uint8_t));
s->m.me.map = av_mallocz(ME_MAP_SIZE*sizeof(uint32_t));
s->m.me.score_map = av_mallocz(ME_MAP_SIZE*sizeof(uint32_t));
s->m.obmc_scratchpad= av_mallocz(MB_SIZE*MB_SIZE*12*sizeof(uint32_t));
if (!s->m.me.scratchpad || !s->m.me.map || !s->m.me.score_map || !s->m.obmc_scratchpad)
s->m.sc.obmc_scratchpad= av_mallocz(MB_SIZE*MB_SIZE*12*sizeof(uint32_t));
if (!s->m.me.scratchpad || !s->m.me.map || !s->m.me.score_map || !s->m.sc.obmc_scratchpad)
return AVERROR(ENOMEM);
ff_h263_encode_init(&s->m); //mv_penalty
@ -501,7 +501,7 @@ static int get_dc(SnowContext *s, int mb_x, int mb_y, int plane_index){
const int obmc_stride= plane_index ? (2*block_size)>>s->chroma_h_shift : 2*block_size;
const int ref_stride= s->current_picture->linesize[plane_index];
uint8_t *src= s-> input_picture->data[plane_index];
IDWTELEM *dst= (IDWTELEM*)s->m.obmc_scratchpad + plane_index*block_size*block_size*4; //FIXME change to unsigned
IDWTELEM *dst= (IDWTELEM*)s->m.sc.obmc_scratchpad + plane_index*block_size*block_size*4; //FIXME change to unsigned
const int b_stride = s->b_width << s->block_max_depth;
const int w= p->width;
const int h= p->height;
@ -596,7 +596,7 @@ static int get_block_rd(SnowContext *s, int mb_x, int mb_y, int plane_index, uin
const int ref_stride= s->current_picture->linesize[plane_index];
uint8_t *dst= s->current_picture->data[plane_index];
uint8_t *src= s-> input_picture->data[plane_index];
IDWTELEM *pred= (IDWTELEM*)s->m.obmc_scratchpad + plane_index*block_size*block_size*4;
IDWTELEM *pred= (IDWTELEM*)s->m.sc.obmc_scratchpad + plane_index*block_size*block_size*4;
uint8_t *cur = s->scratchbuf;
uint8_t *tmp = s->emu_edge_buffer;
const int b_stride = s->b_width << s->block_max_depth;

40
libavcodec/vc1_mc.c
View File

@ -271,25 +271,25 @@ void ff_vc1_mc_1mv(VC1Context *v, int dir)
/* for grayscale we should not try to read from unknown area */
if (CONFIG_GRAY && s->avctx->flags & CODEC_FLAG_GRAY) {
srcU = s->edge_emu_buffer + 18 * s->linesize;
srcV = s->edge_emu_buffer + 18 * s->linesize;
srcU = s->sc.edge_emu_buffer + 18 * s->linesize;
srcV = s->sc.edge_emu_buffer + 18 * s->linesize;
}
if (v->rangeredfrm || use_ic
|| s->h_edge_pos < 22 || v_edge_pos < 22
|| (unsigned)(src_x - s->mspel) > s->h_edge_pos - (mx&3) - 16 - s->mspel * 3
|| (unsigned)(src_y - 1) > v_edge_pos - (my&3) - 16 - 3) {
uint8_t *ubuf = s->edge_emu_buffer + 19 * s->linesize;
uint8_t *ubuf = s->sc.edge_emu_buffer + 19 * s->linesize;
uint8_t *vbuf = ubuf + 9 * s->uvlinesize;
const int k = 17 + s->mspel * 2;
srcY -= s->mspel * (1 + s->linesize);
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcY,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, srcY,
s->linesize, s->linesize,
k, k,
src_x - s->mspel, src_y - s->mspel,
s->h_edge_pos, v_edge_pos);
srcY = s->edge_emu_buffer;
srcY = s->sc.edge_emu_buffer;
s->vdsp.emulated_edge_mc(ubuf, srcU,
s->uvlinesize, s->uvlinesize,
8 + 1, 8 + 1,
@ -467,12 +467,12 @@ void ff_vc1_mc_4mv_luma(VC1Context *v, int n, int dir, int avg)
srcY -= s->mspel * (1 + (s->linesize << fieldmv));
/* check emulate edge stride and offset */
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcY,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, srcY,
s->linesize, s->linesize,
k, k << fieldmv,
src_x - s->mspel, src_y - (s->mspel << fieldmv),
s->h_edge_pos, v_edge_pos);
srcY = s->edge_emu_buffer;
srcY = s->sc.edge_emu_buffer;
/* if we deal with range reduction we need to scale source blocks */
if (v->rangeredfrm) {
vc1_scale_luma(srcY, k, s->linesize << fieldmv);
@ -602,16 +602,16 @@ void ff_vc1_mc_4mv_chroma(VC1Context *v, int dir)
|| s->h_edge_pos < 18 || v_edge_pos < 18
|| (unsigned)uvsrc_x > (s->h_edge_pos >> 1) - 9
|| (unsigned)uvsrc_y > (v_edge_pos >> 1) - 9) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcU,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, srcU,
s->uvlinesize, s->uvlinesize,
8 + 1, 8 + 1, uvsrc_x, uvsrc_y,
s->h_edge_pos >> 1, v_edge_pos >> 1);
s->vdsp.emulated_edge_mc(s->edge_emu_buffer + 16, srcV,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer + 16, srcV,
s->uvlinesize, s->uvlinesize,
8 + 1, 8 + 1, uvsrc_x, uvsrc_y,
s->h_edge_pos >> 1, v_edge_pos >> 1);
srcU = s->edge_emu_buffer;
srcV = s->edge_emu_buffer + 16;
srcU = s->sc.edge_emu_buffer;
srcV = s->sc.edge_emu_buffer + 16;
/* if we deal with range reduction we need to scale source blocks */
if (v->rangeredfrm) {
@ -704,16 +704,16 @@ void ff_vc1_mc_4mv_chroma4(VC1Context *v, int dir, int dir2, int avg)
|| s->h_edge_pos < 10 || v_edge_pos < (5 << fieldmv)
|| (unsigned)uvsrc_x > (s->h_edge_pos >> 1) - 5
|| (unsigned)uvsrc_y > v_edge_pos - (5 << fieldmv)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcU,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, srcU,
s->uvlinesize, s->uvlinesize,
5, (5 << fieldmv), uvsrc_x, uvsrc_y,
s->h_edge_pos >> 1, v_edge_pos);
s->vdsp.emulated_edge_mc(s->edge_emu_buffer + 16, srcV,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer + 16, srcV,
s->uvlinesize, s->uvlinesize,
5, (5 << fieldmv), uvsrc_x, uvsrc_y,
s->h_edge_pos >> 1, v_edge_pos);
srcU = s->edge_emu_buffer;
srcV = s->edge_emu_buffer + 16;
srcU = s->sc.edge_emu_buffer;
srcV = s->sc.edge_emu_buffer + 16;
/* if we deal with intensity compensation we need to scale source blocks */
if (use_ic) {
@ -802,24 +802,24 @@ void ff_vc1_interp_mc(VC1Context *v)
/* for grayscale we should not try to read from unknown area */
if (CONFIG_GRAY && s->avctx->flags & CODEC_FLAG_GRAY) {
srcU = s->edge_emu_buffer + 18 * s->linesize;
srcV = s->edge_emu_buffer + 18 * s->linesize;
srcU = s->sc.edge_emu_buffer + 18 * s->linesize;
srcV = s->sc.edge_emu_buffer + 18 * s->linesize;
}
if (v->rangeredfrm || s->h_edge_pos < 22 || v_edge_pos < 22 || use_ic
|| (unsigned)(src_x - 1) > s->h_edge_pos - (mx & 3) - 16 - 3
|| (unsigned)(src_y - 1) > v_edge_pos - (my & 3) - 16 - 3) {
uint8_t *ubuf = s->edge_emu_buffer + 19 * s->linesize;
uint8_t *ubuf = s->sc.edge_emu_buffer + 19 * s->linesize;
uint8_t *vbuf = ubuf + 9 * s->uvlinesize;
const int k = 17 + s->mspel * 2;
srcY -= s->mspel * (1 + s->linesize);
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcY,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, srcY,
s->linesize, s->linesize,
k, k,
src_x - s->mspel, src_y - s->mspel,
s->h_edge_pos, v_edge_pos);
srcY = s->edge_emu_buffer;
srcY = s->sc.edge_emu_buffer;
s->vdsp.emulated_edge_mc(ubuf, srcU,
s->uvlinesize, s->uvlinesize,
8 + 1, 8 + 1,

12
libavcodec/wmv2.c
View File

@ -127,11 +127,11 @@ void ff_mspel_motion(MpegEncContext *s, uint8_t *dest_y,
if (src_x < 1 || src_y < 1 || src_x + 17 >= s->h_edge_pos ||
src_y + h + 1 >= v_edge_pos) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr - 1 - s->linesize,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, ptr - 1 - s->linesize,
s->linesize, s->linesize, 19, 19,
src_x - 1, src_y - 1,
s->h_edge_pos, s->v_edge_pos);
ptr = s->edge_emu_buffer + 1 + s->linesize;
ptr = s->sc.edge_emu_buffer + 1 + s->linesize;
emu = 1;
}
@ -162,23 +162,23 @@ void ff_mspel_motion(MpegEncContext *s, uint8_t *dest_y,
offset = (src_y * uvlinesize) + src_x;
ptr = ref_picture[1] + offset;
if (emu) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, ptr,
s->uvlinesize, s->uvlinesize,
9, 9,
src_x, src_y,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
ptr = s->edge_emu_buffer;
ptr = s->sc.edge_emu_buffer;
}
pix_op[1][dxy](dest_cb, ptr, uvlinesize, h >> 1);
ptr = ref_picture[2] + offset;
if (emu) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
s->vdsp.emulated_edge_mc(s->sc.edge_emu_buffer, ptr,
s->uvlinesize, s->uvlinesize,
9, 9,
src_x, src_y,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
ptr = s->edge_emu_buffer;
ptr = s->sc.edge_emu_buffer;
}
pix_op[1][dxy](dest_cr, ptr, uvlinesize, h >> 1);
}