#include <limits.h>
#include "sp5x.h"
+#define QUANT_BIAS_SHIFT 8
+
+#define QMAT_SHIFT_MMX 16
+#define QMAT_SHIFT 21
+
static int encode_picture(MpegEncContext *s, int picture_number);
static int dct_quantize_refine(MpegEncContext *s, int16_t *block, int16_t *weight, int16_t *orig, int n, int qscale);
static int sse_mb(MpegEncContext *s);
s->mpeg_quant = avctx->mpeg_quant;
s->rtp_mode = !!avctx->rtp_payload_size;
s->intra_dc_precision = avctx->intra_dc_precision;
+
+ // workaround some differences between how applications specify dc precission
+ if (s->intra_dc_precision < 0) {
+ s->intra_dc_precision += 8;
+ } else if (s->intra_dc_precision >= 8)
+ s->intra_dc_precision -= 8;
+
+ if (s->intra_dc_precision < 0) {
+ av_log(avctx, AV_LOG_ERROR,
+ "intra dc precision must be positive, note some applications use"
+ " 0 and some 8 as base meaning 8bit, the value must not be smaller than that\n");
+ return AVERROR(EINVAL);
+ }
+
+ if (s->intra_dc_precision > (avctx->codec_id == AV_CODEC_ID_MPEG2VIDEO ? 3 : 0)) {
+ av_log(avctx, AV_LOG_ERROR, "intra dc precision too large\n");
+ return AVERROR(EINVAL);
+ }
s->user_specified_pts = AV_NOPTS_VALUE;
if (s->gop_size <= 1) {
if ((!avctx->rc_max_rate) != (!avctx->rc_buffer_size)) {
av_log(avctx, AV_LOG_ERROR, "Either both buffer size and max rate or neither must be specified\n");
- if (avctx->rc_max_rate && !avctx->rc_buffer_size)
- return -1;
+ return -1;
}
if (avctx->rc_min_rate && avctx->rc_max_rate != avctx->rc_min_rate) {
return -1;
ff_fdctdsp_init(&s->fdsp, avctx);
+ ff_me_cmp_init(&s->mecc, avctx);
ff_mpegvideoencdsp_init(&s->mpvencdsp, avctx);
ff_pixblockdsp_init(&s->pdsp, avctx);
ff_qpeldsp_init(&s->qdsp);
s->quant_precision = 5;
- ff_set_cmp(&s->dsp, s->dsp.ildct_cmp, s->avctx->ildct_cmp);
- ff_set_cmp(&s->dsp, s->dsp.frame_skip_cmp, s->avctx->frame_skip_cmp);
+ ff_set_cmp(&s->mecc, s->mecc.ildct_cmp, s->avctx->ildct_cmp);
+ ff_set_cmp(&s->mecc, s->mecc.frame_skip_cmp, s->avctx->frame_skip_cmp);
if (CONFIG_H261_ENCODER && s->out_format == FMT_H261)
ff_h261_encode_init(s);
for (y = 0; y < h; y += 16) {
for (x = 0; x < w; x += 16) {
int offset = x + y * stride;
- int sad = s->dsp.sad[0](NULL, src + offset, ref + offset, stride,
- 16);
+ int sad = s->mecc.sad[0](NULL, src + offset, ref + offset,
+ stride, 16);
int mean = (s->mpvencdsp.pix_sum(src + offset, stride) + 128) >> 8;
int sae = get_sae(src + offset, mean, stride);
int vpad = 16;
if ( s->codec_id == AV_CODEC_ID_MPEG2VIDEO
- && !s->progressive_sequence)
+ && !s->progressive_sequence
+ && FFALIGN(s->height, 32) - s->height > 16)
vpad = 32;
if (!s->avctx->rc_buffer_size)
int off = p->shared ? 0 : 16;
uint8_t *dptr = p->f->data[plane] + 8 * (x + y * stride) + off;
uint8_t *rptr = ref->f->data[plane] + 8 * (x + y * stride);
- int v = s->dsp.frame_skip_cmp[1](s, dptr, rptr, stride, 8);
+ int v = s->mecc.frame_skip_cmp[1](s, dptr, rptr, stride, 8);
switch (FFABS(s->avctx->frame_skip_exp)) {
case 0: score = FFMAX(score, v); break;
int progressive_score, interlaced_score;
s->interlaced_dct = 0;
- progressive_score = s->dsp.ildct_cmp[4](s, ptr_y,
- NULL, wrap_y, 8) +
- s->dsp.ildct_cmp[4](s, ptr_y + wrap_y * 8,
- NULL, wrap_y, 8) - 400;
+ progressive_score = s->mecc.ildct_cmp[4](s, ptr_y, NULL, wrap_y, 8) +
+ s->mecc.ildct_cmp[4](s, ptr_y + wrap_y * 8,
+ NULL, wrap_y, 8) - 400;
if (progressive_score > 0) {
- interlaced_score = s->dsp.ildct_cmp[4](s, ptr_y,
- NULL, wrap_y * 2, 8) +
- s->dsp.ildct_cmp[4](s, ptr_y + wrap_y,
- NULL, wrap_y * 2, 8);
+ interlaced_score = s->mecc.ildct_cmp[4](s, ptr_y,
+ NULL, wrap_y * 2, 8) +
+ s->mecc.ildct_cmp[4](s, ptr_y + wrap_y,
+ NULL, wrap_y * 2, 8);
if (progressive_score > interlaced_score) {
s->interlaced_dct = 1;
int progressive_score, interlaced_score;
s->interlaced_dct = 0;
- progressive_score = s->dsp.ildct_cmp[0](s, dest_y,
- ptr_y, wrap_y,
- 8) +
- s->dsp.ildct_cmp[0](s, dest_y + wrap_y * 8,
- ptr_y + wrap_y * 8, wrap_y,
- 8) - 400;
+ progressive_score = s->mecc.ildct_cmp[0](s, dest_y, ptr_y, wrap_y, 8) +
+ s->mecc.ildct_cmp[0](s, dest_y + wrap_y * 8,
+ ptr_y + wrap_y * 8,
+ wrap_y, 8) - 400;
if (s->avctx->ildct_cmp == FF_CMP_VSSE)
progressive_score -= 400;
if (progressive_score > 0) {
- interlaced_score = s->dsp.ildct_cmp[0](s, dest_y,
- ptr_y,
- wrap_y * 2, 8) +
- s->dsp.ildct_cmp[0](s, dest_y + wrap_y,
- ptr_y + wrap_y,
- wrap_y * 2, 8);
+ interlaced_score = s->mecc.ildct_cmp[0](s, dest_y, ptr_y,
+ wrap_y * 2, 8) +
+ s->mecc.ildct_cmp[0](s, dest_y + wrap_y,
+ ptr_y + wrap_y,
+ wrap_y * 2, 8);
if (progressive_score > interlaced_score) {
s->interlaced_dct = 1;
if (s->current_picture.mc_mb_var[s->mb_stride * mb_y + mb_x] <
2 * s->qscale * s->qscale) {
// FIXME optimize
- if (s->dsp.sad[1](NULL, ptr_y , dest_y,
- wrap_y, 8) < 20 * s->qscale)
+ if (s->mecc.sad[1](NULL, ptr_y, dest_y, wrap_y, 8) < 20 * s->qscale)
skip_dct[0] = 1;
- if (s->dsp.sad[1](NULL, ptr_y + 8,
- dest_y + 8, wrap_y, 8) < 20 * s->qscale)
+ if (s->mecc.sad[1](NULL, ptr_y + 8, dest_y + 8, wrap_y, 8) < 20 * s->qscale)
skip_dct[1] = 1;
- if (s->dsp.sad[1](NULL, ptr_y + dct_offset,
- dest_y + dct_offset, wrap_y, 8) < 20 * s->qscale)
+ if (s->mecc.sad[1](NULL, ptr_y + dct_offset, dest_y + dct_offset,
+ wrap_y, 8) < 20 * s->qscale)
skip_dct[2] = 1;
- if (s->dsp.sad[1](NULL, ptr_y + dct_offset + 8,
- dest_y + dct_offset + 8,
- wrap_y, 8) < 20 * s->qscale)
+ if (s->mecc.sad[1](NULL, ptr_y + dct_offset + 8, dest_y + dct_offset + 8,
+ wrap_y, 8) < 20 * s->qscale)
skip_dct[3] = 1;
- if (s->dsp.sad[1](NULL, ptr_cb, dest_cb,
- wrap_c, 8) < 20 * s->qscale)
+ if (s->mecc.sad[1](NULL, ptr_cb, dest_cb, wrap_c, 8) < 20 * s->qscale)
skip_dct[4] = 1;
- if (s->dsp.sad[1](NULL, ptr_cr, dest_cr,
- wrap_c, 8) < 20 * s->qscale)
+ if (s->mecc.sad[1](NULL, ptr_cr, dest_cr, wrap_c, 8) < 20 * s->qscale)
skip_dct[5] = 1;
if (!s->chroma_y_shift) { /* 422 */
- if (s->dsp.sad[1](NULL, ptr_cb + uv_dct_offset,
- dest_cb + uv_dct_offset,
- wrap_c, 8) < 20 * s->qscale)
+ if (s->mecc.sad[1](NULL, ptr_cb + uv_dct_offset,
+ dest_cb + uv_dct_offset,
+ wrap_c, 8) < 20 * s->qscale)
skip_dct[6] = 1;
- if (s->dsp.sad[1](NULL, ptr_cr + uv_dct_offset,
- dest_cr + uv_dct_offset,
- wrap_c, 8) < 20 * s->qscale)
+ if (s->mecc.sad[1](NULL, ptr_cr + uv_dct_offset,
+ dest_cr + uv_dct_offset,
+ wrap_c, 8) < 20 * s->qscale)
skip_dct[7] = 1;
}
}
int x,y;
if(w==16 && h==16)
- return s->dsp.sse[0](NULL, src1, src2, stride, 16);
+ return s->mecc.sse[0](NULL, src1, src2, stride, 16);
else if(w==8 && h==8)
- return s->dsp.sse[1](NULL, src1, src2, stride, 8);
+ return s->mecc.sse[1](NULL, src1, src2, stride, 8);
for(y=0; y<h; y++){
for(x=0; x<w; x++){
if(w==16 && h==16)
if(s->avctx->mb_cmp == FF_CMP_NSSE){
- return s->dsp.nsse[0](s, s->new_picture.f->data[0] + s->mb_x*16 + s->mb_y*s->linesize*16, s->dest[0], s->linesize, 16)
- +s->dsp.nsse[1](s, s->new_picture.f->data[1] + s->mb_x*8 + s->mb_y*s->uvlinesize*8,s->dest[1], s->uvlinesize, 8)
- +s->dsp.nsse[1](s, s->new_picture.f->data[2] + s->mb_x*8 + s->mb_y*s->uvlinesize*8,s->dest[2], s->uvlinesize, 8);
+ return s->mecc.nsse[0](s, s->new_picture.f->data[0] + s->mb_x * 16 + s->mb_y * s->linesize * 16, s->dest[0], s->linesize, 16) +
+ s->mecc.nsse[1](s, s->new_picture.f->data[1] + s->mb_x * 8 + s->mb_y * s->uvlinesize * 8, s->dest[1], s->uvlinesize, 8) +
+ s->mecc.nsse[1](s, s->new_picture.f->data[2] + s->mb_x * 8 + s->mb_y * s->uvlinesize * 8, s->dest[2], s->uvlinesize, 8);
}else{
- return s->dsp.sse[0](NULL, s->new_picture.f->data[0] + s->mb_x*16 + s->mb_y*s->linesize*16, s->dest[0], s->linesize, 16)
- +s->dsp.sse[1](NULL, s->new_picture.f->data[1] + s->mb_x*8 + s->mb_y*s->uvlinesize*8,s->dest[1], s->uvlinesize, 8)
- +s->dsp.sse[1](NULL, s->new_picture.f->data[2] + s->mb_x*8 + s->mb_y*s->uvlinesize*8,s->dest[2], s->uvlinesize, 8);
+ return s->mecc.sse[0](NULL, s->new_picture.f->data[0] + s->mb_x * 16 + s->mb_y * s->linesize * 16, s->dest[0], s->linesize, 16) +
+ s->mecc.sse[1](NULL, s->new_picture.f->data[1] + s->mb_x * 8 + s->mb_y * s->uvlinesize * 8, s->dest[1], s->uvlinesize, 8) +
+ s->mecc.sse[1](NULL, s->new_picture.f->data[2] + s->mb_x * 8 + s->mb_y * s->uvlinesize * 8, s->dest[2], s->uvlinesize, 8);
}
else
return sse(s, s->new_picture.f->data[0] + s->mb_x*16 + s->mb_y*s->linesize*16, s->dest[0], w, h, s->linesize)
*overflow= s->max_qcoeff < max; //overflow might have happened
/* we need this permutation so that we correct the IDCT, we only permute the !=0 elements */
- if (s->idsp.idct_permutation_type != FF_NO_IDCT_PERM)
+ if (s->idsp.perm_type != FF_IDCT_PERM_NONE)
ff_block_permute(block, s->idsp.idct_permutation,
scantable, last_non_zero);