Merge remote-tracking branch 'qatar/master'

* qatar/master:
  ARM: allow runtime masking of CPU features
  dsputil: remove unused functions
  mov: Treat keyframe indexes as 1-origin if starting at non-zero.
  mov: Take stps entries into consideration also about key_off.
  Remove lowres video decoding

Conflicts:
	ffmpeg.c
	ffplay.c
	libavcodec/arm/vp8dsp_init_arm.c
	libavcodec/libopenjpegdec.c
	libavcodec/mjpegdec.c
	libavcodec/mpegvideo.c
	libavcodec/utils.c
	libavformat/mov.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>
This commit is contained in:
Michael Niedermayer 2012-04-22 22:26:42 +02:00
commit 92ef4be4ab
52 changed files with 219 additions and 841 deletions

View file

@ -1784,381 +1784,6 @@ void ff_print_debug_info(MpegEncContext *s, AVFrame *pict)
}
}
static inline int hpel_motion_lowres(MpegEncContext *s,
uint8_t *dest, uint8_t *src,
int field_based, int field_select,
int src_x, int src_y,
int width, int height, int stride,
int h_edge_pos, int v_edge_pos,
int w, int h, h264_chroma_mc_func *pix_op,
int motion_x, int motion_y)
{
const int lowres = s->avctx->lowres;
const int op_index = FFMIN(lowres, 2);
const int s_mask = (2 << lowres) - 1;
int emu = 0;
int sx, sy;
if (s->quarter_sample) {
motion_x /= 2;
motion_y /= 2;
}
sx = motion_x & s_mask;
sy = motion_y & s_mask;
src_x += motion_x >> lowres + 1;
src_y += motion_y >> lowres + 1;
src += src_y * stride + src_x;
if ((unsigned)src_x > FFMAX( h_edge_pos - (!!sx) - w, 0) ||
(unsigned)src_y > FFMAX((v_edge_pos >> field_based) - (!!sy) - h, 0)) {
s->dsp.emulated_edge_mc(s->edge_emu_buffer, src, 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;
emu = 1;
}
sx = (sx << 2) >> lowres;
sy = (sy << 2) >> lowres;
if (field_select)
src += s->linesize;
pix_op[op_index](dest, src, stride, h, sx, sy);
return emu;
}
/* apply one mpeg motion vector to the three components */
static av_always_inline void mpeg_motion_lowres(MpegEncContext *s,
uint8_t *dest_y,
uint8_t *dest_cb,
uint8_t *dest_cr,
int field_based,
int bottom_field,
int field_select,
uint8_t **ref_picture,
h264_chroma_mc_func *pix_op,
int motion_x, int motion_y,
int h, int mb_y)
{
uint8_t *ptr_y, *ptr_cb, *ptr_cr;
int mx, my, src_x, src_y, uvsrc_x, uvsrc_y, uvlinesize, linesize, sx, sy,
uvsx, uvsy;
const int lowres = s->avctx->lowres;
const int op_index = FFMIN(lowres-1+s->chroma_x_shift, 2);
const int block_s = 8>>lowres;
const int s_mask = (2 << lowres) - 1;
const int h_edge_pos = s->h_edge_pos >> lowres;
const int v_edge_pos = s->v_edge_pos >> lowres;
linesize = s->current_picture.f.linesize[0] << field_based;
uvlinesize = s->current_picture.f.linesize[1] << field_based;
// FIXME obviously not perfect but qpel will not work in lowres anyway
if (s->quarter_sample) {
motion_x /= 2;
motion_y /= 2;
}
if(field_based){
motion_y += (bottom_field - field_select)*((1 << lowres)-1);
}
sx = motion_x & s_mask;
sy = motion_y & s_mask;
src_x = s->mb_x * 2 * block_s + (motion_x >> lowres + 1);
src_y = (mb_y * 2 * block_s >> field_based) + (motion_y >> lowres + 1);
if (s->out_format == FMT_H263) {
uvsx = ((motion_x >> 1) & s_mask) | (sx & 1);
uvsy = ((motion_y >> 1) & s_mask) | (sy & 1);
uvsrc_x = src_x >> 1;
uvsrc_y = src_y >> 1;
} else if (s->out_format == FMT_H261) {
// even chroma mv's are full pel in H261
mx = motion_x / 4;
my = motion_y / 4;
uvsx = (2 * mx) & s_mask;
uvsy = (2 * my) & s_mask;
uvsrc_x = s->mb_x * block_s + (mx >> lowres);
uvsrc_y = mb_y * block_s + (my >> lowres);
} else {
if(s->chroma_y_shift){
mx = motion_x / 2;
my = motion_y / 2;
uvsx = mx & s_mask;
uvsy = my & s_mask;
uvsrc_x = s->mb_x * block_s + (mx >> lowres + 1);
uvsrc_y = (mb_y * block_s >> field_based) + (my >> lowres + 1);
} else {
if(s->chroma_x_shift){
//Chroma422
mx = motion_x / 2;
uvsx = mx & s_mask;
uvsy = motion_y & s_mask;
uvsrc_y = src_y;
uvsrc_x = s->mb_x*block_s + (mx >> (lowres+1));
} else {
//Chroma444
uvsx = motion_x & s_mask;
uvsy = motion_y & s_mask;
uvsrc_x = src_x;
uvsrc_y = src_y;
}
}
}
ptr_y = ref_picture[0] + src_y * linesize + src_x;
ptr_cb = ref_picture[1] + uvsrc_y * uvlinesize + uvsrc_x;
ptr_cr = ref_picture[2] + uvsrc_y * uvlinesize + uvsrc_x;
if ((unsigned) src_x > FFMAX( h_edge_pos - (!!sx) - 2 * block_s, 0) ||
(unsigned) src_y > FFMAX((v_edge_pos >> field_based) - (!!sy) - h, 0)) {
s->dsp.emulated_edge_mc(s->edge_emu_buffer, ptr_y,
s->linesize, 17, 17 + field_based,
src_x, src_y << field_based, h_edge_pos,
v_edge_pos);
ptr_y = s->edge_emu_buffer;
if (!CONFIG_GRAY || !(s->flags & CODEC_FLAG_GRAY)) {
uint8_t *uvbuf = s->edge_emu_buffer + 18 * s->linesize;
s->dsp.emulated_edge_mc(uvbuf , ptr_cb, s->uvlinesize, 9,
9 + field_based,
uvsrc_x, uvsrc_y << field_based,
h_edge_pos >> 1, v_edge_pos >> 1);
s->dsp.emulated_edge_mc(uvbuf + 16, ptr_cr, s->uvlinesize, 9,
9 + field_based,
uvsrc_x, uvsrc_y << field_based,
h_edge_pos >> 1, v_edge_pos >> 1);
ptr_cb = uvbuf;
ptr_cr = uvbuf + 16;
}
}
// FIXME use this for field pix too instead of the obnoxious hack which changes picture.f.data
if (bottom_field) {
dest_y += s->linesize;
dest_cb += s->uvlinesize;
dest_cr += s->uvlinesize;
}
if (field_select) {
ptr_y += s->linesize;
ptr_cb += s->uvlinesize;
ptr_cr += s->uvlinesize;
}
sx = (sx << 2) >> lowres;
sy = (sy << 2) >> lowres;
pix_op[lowres - 1](dest_y, ptr_y, linesize, h, sx, sy);
if (!CONFIG_GRAY || !(s->flags & CODEC_FLAG_GRAY)) {
uvsx = (uvsx << 2) >> lowres;
uvsy = (uvsy << 2) >> lowres;
if (h >> s->chroma_y_shift) {
pix_op[op_index](dest_cb, ptr_cb, uvlinesize, h >> s->chroma_y_shift, uvsx, uvsy);
pix_op[op_index](dest_cr, ptr_cr, uvlinesize, h >> s->chroma_y_shift, uvsx, uvsy);
}
}
// FIXME h261 lowres loop filter
}
static inline void chroma_4mv_motion_lowres(MpegEncContext *s,
uint8_t *dest_cb, uint8_t *dest_cr,
uint8_t **ref_picture,
h264_chroma_mc_func * pix_op,
int mx, int my)
{
const int lowres = s->avctx->lowres;
const int op_index = FFMIN(lowres, 2);
const int block_s = 8 >> lowres;
const int s_mask = (2 << lowres) - 1;
const int h_edge_pos = s->h_edge_pos >> lowres + 1;
const int v_edge_pos = s->v_edge_pos >> lowres + 1;
int emu = 0, src_x, src_y, offset, sx, sy;
uint8_t *ptr;
if (s->quarter_sample) {
mx /= 2;
my /= 2;
}
/* In case of 8X8, we construct a single chroma motion vector
with a special rounding */
mx = ff_h263_round_chroma(mx);
my = ff_h263_round_chroma(my);
sx = mx & s_mask;
sy = my & s_mask;
src_x = s->mb_x * block_s + (mx >> lowres + 1);
src_y = s->mb_y * block_s + (my >> lowres + 1);
offset = src_y * s->uvlinesize + src_x;
ptr = ref_picture[1] + offset;
if (s->flags & CODEC_FLAG_EMU_EDGE) {
if ((unsigned) src_x > FFMAX(h_edge_pos - (!!sx) - block_s, 0) ||
(unsigned) src_y > FFMAX(v_edge_pos - (!!sy) - block_s, 0)) {
s->dsp.emulated_edge_mc(s->edge_emu_buffer, ptr, s->uvlinesize,
9, 9, src_x, src_y, h_edge_pos, v_edge_pos);
ptr = s->edge_emu_buffer;
emu = 1;
}
}
sx = (sx << 2) >> lowres;
sy = (sy << 2) >> lowres;
pix_op[op_index](dest_cb, ptr, s->uvlinesize, block_s, sx, sy);
ptr = ref_picture[2] + offset;
if (emu) {
s->dsp.emulated_edge_mc(s->edge_emu_buffer, ptr, s->uvlinesize, 9, 9,
src_x, src_y, h_edge_pos, v_edge_pos);
ptr = s->edge_emu_buffer;
}
pix_op[op_index](dest_cr, ptr, s->uvlinesize, block_s, sx, sy);
}
/**
* motion compensation of a single macroblock
* @param s context
* @param dest_y luma destination pointer
* @param dest_cb chroma cb/u destination pointer
* @param dest_cr chroma cr/v destination pointer
* @param dir direction (0->forward, 1->backward)
* @param ref_picture array[3] of pointers to the 3 planes of the reference picture
* @param pix_op halfpel motion compensation function (average or put normally)
* the motion vectors are taken from s->mv and the MV type from s->mv_type
*/
static inline void MPV_motion_lowres(MpegEncContext *s,
uint8_t *dest_y, uint8_t *dest_cb,
uint8_t *dest_cr,
int dir, uint8_t **ref_picture,
h264_chroma_mc_func *pix_op)
{
int mx, my;
int mb_x, mb_y, i;
const int lowres = s->avctx->lowres;
const int block_s = 8 >>lowres;
mb_x = s->mb_x;
mb_y = s->mb_y;
switch (s->mv_type) {
case MV_TYPE_16X16:
mpeg_motion_lowres(s, dest_y, dest_cb, dest_cr,
0, 0, 0,
ref_picture, pix_op,
s->mv[dir][0][0], s->mv[dir][0][1],
2 * block_s, mb_y);
break;
case MV_TYPE_8X8:
mx = 0;
my = 0;
for (i = 0; i < 4; i++) {
hpel_motion_lowres(s, dest_y + ((i & 1) + (i >> 1) *
s->linesize) * block_s,
ref_picture[0], 0, 0,
(2 * mb_x + (i & 1)) * block_s,
(2 * mb_y + (i >> 1)) * block_s,
s->width, s->height, s->linesize,
s->h_edge_pos >> lowres, s->v_edge_pos >> lowres,
block_s, block_s, pix_op,
s->mv[dir][i][0], s->mv[dir][i][1]);
mx += s->mv[dir][i][0];
my += s->mv[dir][i][1];
}
if (!CONFIG_GRAY || !(s->flags & CODEC_FLAG_GRAY))
chroma_4mv_motion_lowres(s, dest_cb, dest_cr, ref_picture,
pix_op, mx, my);
break;
case MV_TYPE_FIELD:
if (s->picture_structure == PICT_FRAME) {
/* top field */
mpeg_motion_lowres(s, dest_y, dest_cb, dest_cr,
1, 0, s->field_select[dir][0],
ref_picture, pix_op,
s->mv[dir][0][0], s->mv[dir][0][1],
block_s, mb_y);
/* bottom field */
mpeg_motion_lowres(s, dest_y, dest_cb, dest_cr,
1, 1, s->field_select[dir][1],
ref_picture, pix_op,
s->mv[dir][1][0], s->mv[dir][1][1],
block_s, mb_y);
} else {
if (s->picture_structure != s->field_select[dir][0] + 1 &&
s->pict_type != AV_PICTURE_TYPE_B && !s->first_field) {
ref_picture = s->current_picture_ptr->f.data;
}
mpeg_motion_lowres(s, dest_y, dest_cb, dest_cr,
0, 0, s->field_select[dir][0],
ref_picture, pix_op,
s->mv[dir][0][0],
s->mv[dir][0][1], 2 * block_s, mb_y >> 1);
}
break;
case MV_TYPE_16X8:
for (i = 0; i < 2; i++) {
uint8_t **ref2picture;
if (s->picture_structure == s->field_select[dir][i] + 1 ||
s->pict_type == AV_PICTURE_TYPE_B || s->first_field) {
ref2picture = ref_picture;
} else {
ref2picture = s->current_picture_ptr->f.data;
}
mpeg_motion_lowres(s, dest_y, dest_cb, dest_cr,
0, 0, s->field_select[dir][i],
ref2picture, pix_op,
s->mv[dir][i][0], s->mv[dir][i][1] +
2 * block_s * i, block_s, mb_y >> 1);
dest_y += 2 * block_s * s->linesize;
dest_cb += (2 * block_s >> s->chroma_y_shift) * s->uvlinesize;
dest_cr += (2 * block_s >> s->chroma_y_shift) * s->uvlinesize;
}
break;
case MV_TYPE_DMV:
if (s->picture_structure == PICT_FRAME) {
for (i = 0; i < 2; i++) {
int j;
for (j = 0; j < 2; j++) {
mpeg_motion_lowres(s, dest_y, dest_cb, dest_cr,
1, j, j ^ i,
ref_picture, pix_op,
s->mv[dir][2 * i + j][0],
s->mv[dir][2 * i + j][1],
block_s, mb_y);
}
pix_op = s->dsp.avg_h264_chroma_pixels_tab;
}
} else {
for (i = 0; i < 2; i++) {
mpeg_motion_lowres(s, dest_y, dest_cb, dest_cr,
0, 0, s->picture_structure != i + 1,
ref_picture, pix_op,
s->mv[dir][2 * i][0],s->mv[dir][2 * i][1],
2 * block_s, mb_y >> 1);
// after put we make avg of the same block
pix_op = s->dsp.avg_h264_chroma_pixels_tab;
// opposite parity is always in the same
// frame if this is second field
if (!s->first_field) {
ref_picture = s->current_picture_ptr->f.data;
}
}
}
break;
default:
assert(0);
}
}
/**
* find the lowest MB row referenced in the MVs
*/
@ -2268,7 +1893,7 @@ void ff_clean_intra_table_entries(MpegEncContext *s)
*/
static av_always_inline
void MPV_decode_mb_internal(MpegEncContext *s, DCTELEM block[12][64],
int lowres_flag, int is_mpeg12)
int is_mpeg12)
{
const int mb_xy = s->mb_y * s->mb_stride + s->mb_x;
if(CONFIG_MPEG_XVMC_DECODER && s->avctx->xvmc_acceleration){
@ -2313,8 +1938,8 @@ void MPV_decode_mb_internal(MpegEncContext *s, DCTELEM block[12][64],
qpel_mc_func (*op_qpix)[16];
const int linesize = s->current_picture.f.linesize[0]; //not s->linesize as this would be wrong for field pics
const int uvlinesize = s->current_picture.f.linesize[1];
const int readable= s->pict_type != AV_PICTURE_TYPE_B || s->encoding || s->avctx->draw_horiz_band || lowres_flag;
const int block_size= lowres_flag ? 8>>s->avctx->lowres : 8;
const int readable= s->pict_type != AV_PICTURE_TYPE_B || s->encoding || s->avctx->draw_horiz_band;
const int block_size = 8;
/* avoid copy if macroblock skipped in last frame too */
/* skip only during decoding as we might trash the buffers during encoding a bit */
@ -2363,31 +1988,19 @@ void MPV_decode_mb_internal(MpegEncContext *s, DCTELEM block[12][64],
}
}
if(lowres_flag){
h264_chroma_mc_func *op_pix = s->dsp.put_h264_chroma_pixels_tab;
if (s->mv_dir & MV_DIR_FORWARD) {
MPV_motion_lowres(s, dest_y, dest_cb, dest_cr, 0, s->last_picture.f.data, op_pix);
op_pix = s->dsp.avg_h264_chroma_pixels_tab;
}
if (s->mv_dir & MV_DIR_BACKWARD) {
MPV_motion_lowres(s, dest_y, dest_cb, dest_cr, 1, s->next_picture.f.data, op_pix);
}
op_qpix= s->me.qpel_put;
if ((!s->no_rounding) || s->pict_type==AV_PICTURE_TYPE_B){
op_pix = s->dsp.put_pixels_tab;
}else{
op_qpix= s->me.qpel_put;
if ((!s->no_rounding) || s->pict_type==AV_PICTURE_TYPE_B){
op_pix = s->dsp.put_pixels_tab;
}else{
op_pix = s->dsp.put_no_rnd_pixels_tab;
}
if (s->mv_dir & MV_DIR_FORWARD) {
MPV_motion(s, dest_y, dest_cb, dest_cr, 0, s->last_picture.f.data, op_pix, op_qpix);
op_pix = s->dsp.avg_pixels_tab;
op_qpix= s->me.qpel_avg;
}
if (s->mv_dir & MV_DIR_BACKWARD) {
MPV_motion(s, dest_y, dest_cb, dest_cr, 1, s->next_picture.f.data, op_pix, op_qpix);
}
op_pix = s->dsp.put_no_rnd_pixels_tab;
}
if (s->mv_dir & MV_DIR_FORWARD) {
MPV_motion(s, dest_y, dest_cb, dest_cr, 0, s->last_picture.f.data, op_pix, op_qpix);
op_pix = s->dsp.avg_pixels_tab;
op_qpix= s->me.qpel_avg;
}
if (s->mv_dir & MV_DIR_BACKWARD) {
MPV_motion(s, dest_y, dest_cb, dest_cr, 1, s->next_picture.f.data, op_pix, op_qpix);
}
}
@ -2513,12 +2126,10 @@ skip_idct:
void ff_MPV_decode_mb(MpegEncContext *s, DCTELEM block[12][64]){
#if !CONFIG_SMALL
if(s->out_format == FMT_MPEG1) {
if(s->avctx->lowres) MPV_decode_mb_internal(s, block, 1, 1);
else MPV_decode_mb_internal(s, block, 0, 1);
MPV_decode_mb_internal(s, block, 1);
} else
#endif
if(s->avctx->lowres) MPV_decode_mb_internal(s, block, 1, 0);
else MPV_decode_mb_internal(s, block, 0, 0);
MPV_decode_mb_internal(s, block, 0);
}
/**
@ -2593,7 +2204,7 @@ void ff_draw_horiz_band(MpegEncContext *s, int y, int h){
void ff_init_block_index(MpegEncContext *s){ //FIXME maybe rename
const int linesize = s->current_picture.f.linesize[0]; //not s->linesize as this would be wrong for field pics
const int uvlinesize = s->current_picture.f.linesize[1];
const int mb_size= 4 - s->avctx->lowres;
const int mb_size= 4;
s->block_index[0]= s->b8_stride*(s->mb_y*2 ) - 2 + s->mb_x*2;
s->block_index[1]= s->b8_stride*(s->mb_y*2 ) - 1 + s->mb_x*2;