Go to the documentation of this file.
21 #define BITSTREAM_READER_LE
87 { 0, -1, -1, -1, -1 },
97 { 0, -1, -1, -1, -1 },
100 { 1, 2, -1, -1, -1 },
131 for (
i = 0;
i < 256;
i++)
134 for (
i = 0;
i < 16;
i++)
141 int i, ps, si,
code, step_i;
147 value = (((ps & 0x7fffff) ^ -si) + si) * (1.0
f / 0x7fffff);
150 if (step_i > step_max) {
157 for (
i = 0;
i < 64;
i++) {
183 step_i =
av_clip(step_i, 0, step_max);
186 s->lfe_data[
i] =
value *
s->lfe_scale;
195 int i, ps, si,
code, step_i;
201 value = (((ps & 0x7fff) ^ -si) + si) * (1.0
f / 0x7fff);
204 if (step_i > step_max) {
211 for (
i = 0;
i < 64;
i++) {
233 step_i =
av_clip(step_i, 0, step_max);
236 s->lfe_data[
i] =
value *
s->lfe_scale;
257 if (chunk->
len >= 52)
259 if (chunk->
len >= 35)
280 int sf, sf_idx, ch, main_ch, freq;
284 for (sf = 0; sf < 1 << group; sf +=
diff ? 8 : 1) {
285 sf_idx = ((
s->framenum << group) + sf) & 31;
286 s->tonal_bounds[group][sf_idx][0] =
s->ntones;
289 for (freq = 1;; freq++) {
306 if (freq >> (5 - group) >
s->nsubbands * 4 - 6) {
315 +
s->limited_range - 2;
316 amp[main_ch] = main_amp <
AMP_MAX ? main_amp : 0;
320 for (ch = 0; ch <
s->nchannels_total; ch++) {
337 t->
x_freq = freq >> (5 - group);
338 t->
f_delt = (freq & ((1 << (5 - group)) - 1)) << group;
344 for (ch = 0; ch <
s->nchannels; ch++) {
346 t->
phs[ch] = 128 - phs[ch] * 32 +
shift;
351 s->tonal_bounds[group][sf_idx][1] =
s->ntones;
375 for (sb = 0; sb < 6; sb++)
381 for (group = 0; group < 5; group++) {
422 int i, sf, prev, next, dist;
431 for (sf = 0; sf < 7; sf += dist) {
451 next = prev + ((next + 1) >> 1);
453 next = prev - ( next >> 1);
459 scf[sf + 1] = prev + ((next - prev) >> 1);
461 scf[sf + 1] = prev - ((prev - next) >> 1);
466 scf[sf + 1] = prev + ( (next - prev) >> 2);
467 scf[sf + 2] = prev + ( (next - prev) >> 1);
468 scf[sf + 3] = prev + (((next - prev) * 3) >> 2);
470 scf[sf + 1] = prev - ( (prev - next) >> 2);
471 scf[sf + 2] = prev - ( (prev - next) >> 1);
472 scf[sf + 3] = prev - (((prev - next) * 3) >> 2);
477 for (
i = 1;
i < dist;
i++)
478 scf[sf +
i] = prev + (next - prev) *
i / dist;
506 int ch, sb, sf, nsubbands,
ret;
517 for (sb = 2; sb < nsubbands; sb++) {
532 for (sb = 0; sb <
s->nsubbands - 4; sb++) {
535 if (sb + 4 <
s->min_mono_subband)
538 s->grid_3_avg[ch2][sb] =
s->grid_3_avg[ch1][sb];
557 nsubbands = (
s->nsubbands -
s->min_mono_subband + 3) / 4;
558 for (sb = 0; sb < nsubbands; sb++)
559 for (ch = ch1; ch <= ch2; ch++)
560 for (sf = 1; sf <= 4; sf++)
564 s->part_stereo_pres |= 1 << ch1;
574 int sb, nsubbands,
ret;
578 for (sb = 2; sb < nsubbands; sb++) {
587 for (sb = 0; sb <
s->nsubbands - 4; sb++) {
588 if (sb + 4 >=
s->min_mono_subband) {
602 for (ch = ch1; ch <= ch2; ch++) {
603 if ((ch != ch1 && sb + 4 >=
s->min_mono_subband) !=
flag)
606 if (
s->grid_3_pres[ch] & (1
U << sb))
609 for (
i = 0;
i < 8;
i++) {
616 s->grid_3_pres[ch] |= 1
U << sb;
622 s->lbr_rand = 1103515245
U *
s->lbr_rand + 12345
U;
623 return s->lbr_rand *
s->sb_scf[sb];
631 float *
samples =
s->time_samples[ch][sb];
632 int i, j,
code, nblocks, coding_method;
639 switch (quant_level) {
644 for (j = 0; j < 8; j++)
662 for (j = 0; j < 5; j++)
673 for (j = 0; j < 3; j++)
686 for (
i = 0;
i < nblocks;
i++)
700 s->ch_pres[ch] |= 1
U << sb;
704 int start_sb,
int end_sb,
int flag)
706 int sb, sb_g3, sb_reorder, quant_level;
708 for (sb = start_sb; sb < end_sb; sb++) {
712 }
else if (
flag && sb < s->max_mono_subband) {
713 sb_reorder =
s->sb_indices[sb];
717 sb_reorder =
get_bits(&
s->gb,
s->limited_range + 3);
720 s->sb_indices[sb] = sb_reorder;
722 if (sb_reorder >=
s->nsubbands)
727 for (sb_g3 = 0; sb_g3 <
s->g3_avg_only_start_sb - 4; sb_g3++)
729 }
else if (sb < 12 && sb_reorder >= 4) {
737 if (!
flag || sb_reorder >=
s->max_mono_subband)
738 s->sec_ch_sbms[ch1 / 2][sb_reorder] =
get_bits(&
s->gb, 8);
739 if (
flag && sb_reorder >=
s->min_mono_subband)
740 s->sec_ch_lrms[ch1 / 2][sb_reorder] =
get_bits(&
s->gb, 8);
743 quant_level =
s->quant_levels[ch1 / 2][sb];
748 if (sb < s->max_mono_subband && sb_reorder >=
s->min_mono_subband) {
750 parse_ch(
s, ch1, sb_reorder, quant_level, 0);
752 parse_ch(
s, ch2, sb_reorder, quant_level, 1);
754 parse_ch(
s, ch1, sb_reorder, quant_level, 0);
756 parse_ch(
s, ch2, sb_reorder, quant_level, 0);
770 for (
i = 0;
i < 8;
i++) {
772 for (j = 0; j < (
i + 1) / 2; j++) {
773 float tmp1 =
coeff[ j ];
774 float tmp2 =
coeff[
i - j - 1];
775 coeff[ j ] = tmp1 + rc * tmp2;
776 coeff[
i - j - 1] = tmp2 + rc * tmp1;
784 int f =
s->framenum & 1;
785 int i, sb, ch, codes[16];
788 for (sb = start_sb; sb < end_sb; sb++) {
789 int ncodes = 8 * (1 + (sb < 2));
790 for (ch = ch1; ch <= ch2; ch++) {
793 for (
i = 0;
i < ncodes;
i++)
795 for (
i = 0;
i < ncodes / 8;
i++)
825 for (sb = 0; sb <
s->nsubbands; sb++) {
826 int f = sb *
s->limited_rate /
s->nsubbands;
827 int a = 18000 / (12 *
f / 1000 + 100 + 40 * st) + 20 * ol;
829 quant_levels[sb] = 1;
831 quant_levels[sb] = 2;
833 quant_levels[sb] = 3;
835 quant_levels[sb] = 4;
837 quant_levels[sb] = 5;
841 for (sb = 0; sb < 8; sb++)
843 for (; sb <
s->nsubbands; sb++)
844 s->quant_levels[ch1 / 2][sb] = quant_levels[sb];
857 for (sb = 0; sb < 2; sb++)
858 for (ch = ch1; ch <= ch2; ch++)
866 int start_sb,
int end_sb,
int flag)
868 int i, j, sb, ch, nsubbands;
871 if (end_sb > nsubbands)
874 for (sb = start_sb; sb < end_sb; sb++) {
875 for (ch = ch1; ch <= ch2; ch++) {
876 uint8_t *g2_scf =
s->grid_2_scf[ch][sb];
880 memcpy(g2_scf,
s->grid_2_scf[ch1][sb], 64);
885 for (
i = 0;
i < 8;
i++, g2_scf += 8) {
887 memset(g2_scf, 0, 64 -
i * 8);
892 for (j = 0; j < 8; j++) {
898 memset(g2_scf, 0, 8);
935 if ((
ret =
parse_ts(
s, ch1, ch2, 6,
s->max_mono_subband, 0)) < 0)
943 if ((
ret =
parse_ts(
s, ch1, ch2,
s->min_mono_subband,
s->nsubbands, 1)) < 0)
950 double scale = (-1.0 / (1 << 17)) * sqrt(1 << (2 -
s->limited_range));
951 int i, br_per_ch =
s->bit_rate_scaled /
s->nchannels_total;
960 for (
i = 0;
i < 32 <<
s->freq_range;
i++)
963 if (br_per_ch < 14000)
965 else if (br_per_ch < 32000)
966 scale = (br_per_ch - 14000) * (1.0 / 120000) + 0.85;
970 scale *= 1.0 / INT_MAX;
972 for (
i = 0;
i <
s->nsubbands;
i++) {
976 s->sb_scf[
i] = (
i - 1) * 0.25 * 0.785 *
scale;
981 s->lfe_scale = (16 <<
s->freq_range) * 0.0000078265894;
990 int nsamples = nchsamples *
s->nchannels *
s->nsubbands;
1000 for (ch = 0; ch <
s->nchannels; ch++) {
1001 for (sb = 0; sb <
s->nsubbands; sb++) {
1002 s->time_samples[ch][sb] = ptr;
1012 int old_rate =
s->sample_rate;
1013 int old_band_limit =
s->band_limit;
1014 int old_nchannels =
s->nchannels;
1016 unsigned int sr_code;
1019 sr_code = bytestream2_get_byte(gb);
1025 if (
s->sample_rate > 48000) {
1031 s->ch_mask = bytestream2_get_le16(gb);
1032 if (!(
s->ch_mask & 0x7)) {
1036 if ((
s->ch_mask & 0xfff0) && !(
s->warned & 1)) {
1042 version = bytestream2_get_le16(gb);
1043 if ((
version & 0xff00) != 0x0800) {
1049 s->flags = bytestream2_get_byte(gb);
1055 if (!(
s->warned & 2)) {
1063 bit_rate_hi = bytestream2_get_byte(gb);
1066 s->bit_rate_orig = bytestream2_get_le16(gb) | ((bit_rate_hi & 0x0F) << 16);
1069 s->bit_rate_scaled = bytestream2_get_le16(gb) | ((bit_rate_hi & 0xF0) << 12);
1095 if (
s->bit_rate_orig >= 44000 * (
s->nchannels_total + 2))
1097 else if (
s->bit_rate_orig >= 25000 * (
s->nchannels_total + 2))
1103 s->limited_rate =
s->sample_rate >>
s->band_limit;
1104 s->limited_range =
s->freq_range -
s->band_limit;
1105 if (
s->limited_range < 0) {
1110 s->nsubbands = 8 <<
s->limited_range;
1113 if (
s->g3_avg_only_start_sb >
s->nsubbands)
1114 s->g3_avg_only_start_sb =
s->nsubbands;
1116 s->min_mono_subband =
s->nsubbands * 2000 / (
s->limited_rate / 2);
1117 if (
s->min_mono_subband >
s->nsubbands)
1118 s->min_mono_subband =
s->nsubbands;
1120 s->max_mono_subband =
s->nsubbands * 14000 / (
s->limited_rate / 2);
1121 if (
s->max_mono_subband >
s->nsubbands)
1122 s->max_mono_subband =
s->nsubbands;
1125 if ((old_rate !=
s->sample_rate || old_band_limit !=
s->band_limit) &&
init_sample_rate(
s) < 0)
1144 s->nchannels_total += 2;
1151 if (old_rate !=
s->sample_rate
1152 || old_band_limit !=
s->band_limit
1153 || old_nchannels !=
s->nchannels) {
1176 int i, ch, sb, sf,
ret, group, chunk_id, chunk_len;
1187 switch (bytestream2_get_byte(&gb)) {
1189 if (!
s->sample_rate) {
1206 chunk_id = bytestream2_get_byte(&gb);
1207 chunk_len = (chunk_id & 0x80) ? bytestream2_get_be16(&gb) : bytestream2_get_byte(&gb);
1218 switch (chunk_id & 0x7f) {
1221 int checksum = bytestream2_get_be16(&gb);
1222 uint16_t res = chunk_id;
1223 res += (chunk_len >> 8) & 0xff;
1224 res += chunk_len & 0xff;
1225 for (
i = 0;
i < chunk_len - 2;
i++)
1244 memset(
s->quant_levels, 0,
sizeof(
s->quant_levels));
1245 memset(
s->sb_indices, 0xff,
sizeof(
s->sb_indices));
1246 memset(
s->sec_ch_sbms, 0,
sizeof(
s->sec_ch_sbms));
1247 memset(
s->sec_ch_lrms, 0,
sizeof(
s->sec_ch_lrms));
1248 memset(
s->ch_pres, 0,
sizeof(
s->ch_pres));
1249 memset(
s->grid_1_scf, 0,
sizeof(
s->grid_1_scf));
1250 memset(
s->grid_2_scf, 0,
sizeof(
s->grid_2_scf));
1251 memset(
s->grid_3_avg, 0,
sizeof(
s->grid_3_avg));
1252 memset(
s->grid_3_scf, 0,
sizeof(
s->grid_3_scf));
1253 memset(
s->grid_3_pres, 0,
sizeof(
s->grid_3_pres));
1254 memset(
s->tonal_scf, 0,
sizeof(
s->tonal_scf));
1255 memset(
s->lfe_data, 0,
sizeof(
s->lfe_data));
1256 s->part_stereo_pres = 0;
1257 s->framenum = (
s->framenum + 1) & 31;
1259 for (ch = 0; ch <
s->nchannels; ch++) {
1260 for (sb = 0; sb <
s->nsubbands / 4; sb++) {
1261 s->part_stereo[ch][sb][0] =
s->part_stereo[ch][sb][4];
1262 s->part_stereo[ch][sb][4] = 16;
1266 memset(
s->lpc_coeff[
s->framenum & 1], 0,
sizeof(
s->lpc_coeff[0]));
1268 for (group = 0; group < 5; group++) {
1269 for (sf = 0; sf < 1 << group; sf++) {
1270 int sf_idx = ((
s->framenum << group) + sf) & 31;
1271 s->tonal_bounds[group][sf_idx][0] =
1272 s->tonal_bounds[group][sf_idx][1] =
s->ntones;
1278 chunk_id = bytestream2_get_byte(&gb);
1279 chunk_len = (chunk_id & 0x80) ? bytestream2_get_be16(&gb) : bytestream2_get_byte(&gb);
1291 chunk.lfe.len = chunk_len;
1292 chunk.lfe.data = gb.
buffer;
1298 chunk.tonal.id = chunk_id;
1299 chunk.tonal.len = chunk_len;
1300 chunk.tonal.data = gb.
buffer;
1309 chunk.tonal_grp[
i].id =
i;
1310 chunk.tonal_grp[
i].len = chunk_len;
1311 chunk.tonal_grp[
i].data = gb.
buffer;
1320 chunk.tonal_grp[
i].id =
i;
1321 chunk.tonal_grp[
i].len = chunk_len;
1322 chunk.tonal_grp[
i].data = gb.
buffer;
1329 chunk.grid1[
i].len = chunk_len;
1330 chunk.grid1[
i].data = gb.
buffer;
1337 chunk.hr_grid[
i].len = chunk_len;
1338 chunk.hr_grid[
i].data = gb.
buffer;
1345 chunk.ts1[
i].len = chunk_len;
1346 chunk.ts1[
i].data = gb.
buffer;
1353 chunk.ts2[
i].len = chunk_len;
1354 chunk.ts2[
i].data = gb.
buffer;
1366 for (
i = 0;
i < 5;
i++)
1369 for (
i = 0;
i < (
s->nchannels + 1) / 2;
i++) {
1371 int ch2 =
FFMIN(ch1 + 1,
s->nchannels - 1);
1380 if (!chunk.grid1[
i].len || !chunk.hr_grid[
i].len || !chunk.ts1[
i].len)
1403 for (ch = ch1; ch <= ch2; ch++) {
1404 for (sb = 0; sb <
s->nsubbands; sb++) {
1407 uint8_t *g1_scf_a =
s->grid_1_scf[ch][g1_sb ];
1408 uint8_t *g1_scf_b =
s->grid_1_scf[ch][g1_sb + 1];
1413 uint8_t *hr_scf =
s->high_res_scf[ch][sb];
1416 for (
i = 0;
i < 8;
i++) {
1417 int scf = w1 * g1_scf_a[
i] + w2 * g1_scf_b[
i];
1418 hr_scf[
i] = scf >> 7;
1421 int8_t *g3_scf =
s->grid_3_scf[ch][sb - 4];
1422 int g3_avg =
s->grid_3_avg[ch][sb - 4];
1424 for (
i = 0;
i < 8;
i++) {
1425 int scf = w1 * g1_scf_a[
i] + w2 * g1_scf_b[
i];
1426 hr_scf[
i] = (scf >> 7) - g3_avg - g3_scf[
i];
1438 int i, j, k, ch, sb;
1440 for (ch = ch1; ch <= ch2; ch++) {
1441 for (sb = 0; sb <
s->nsubbands; sb++) {
1442 float *
samples =
s->time_samples[ch][sb];
1444 if (
s->ch_pres[ch] & (1
U << sb))
1450 }
else if (sb < 10) {
1455 float accum[8] = { 0 };
1458 for (k = 2; k < 6; k++) {
1459 float *other = &
s->time_samples[ch][k][
i * 8];
1460 for (j = 0; j < 8; j++)
1461 accum[j] +=
fabs(other[j]);
1464 for (j = 0; j < 8; j++)
1476 for (
i = 0;
i < nsamples;
i++) {
1478 for (j = 0; j < 8; j++)
1486 int f =
s->framenum & 1;
1489 for (ch = ch1; ch <= ch2; ch++) {
1490 float *
samples =
s->time_samples[ch][sb];
1492 if (!(
s->ch_pres[ch] & (1
U << sb)))
1510 for (sb = 0; sb <
s->nsubbands; sb++) {
1512 for (ch = ch1; ch <= ch2; ch++) {
1513 float *
samples =
s->time_samples[ch][sb];
1514 uint8_t *hr_scf =
s->high_res_scf[ch][sb];
1517 unsigned int scf = hr_scf[
i];
1520 for (j = 0; j < 16; j++)
1526 unsigned int scf = hr_scf[
i / 8] - g2_scf[
i];
1537 float *samples_l =
s->time_samples[ch1][sb];
1538 float *samples_r =
s->time_samples[ch2][sb];
1539 int ch2_pres =
s->ch_pres[ch2] & (1
U << sb);
1542 int sbms = (
s->sec_ch_sbms[ch1 / 2][sb] >>
i) & 1;
1543 int lrms = (
s->sec_ch_lrms[ch1 / 2][sb] >>
i) & 1;
1545 if (sb >=
s->min_mono_subband) {
1546 if (lrms && ch2_pres) {
1548 for (j = 0; j < 16; j++) {
1549 float tmp = samples_l[j];
1550 samples_l[j] = samples_r[j];
1551 samples_r[j] = -
tmp;
1554 for (j = 0; j < 16; j++) {
1555 float tmp = samples_l[j];
1556 samples_l[j] = samples_r[j];
1560 }
else if (!ch2_pres) {
1561 if (sbms && (
s->part_stereo_pres & (1 << ch1))) {
1562 for (j = 0; j < 16; j++)
1563 samples_r[j] = -samples_l[j];
1565 for (j = 0; j < 16; j++)
1566 samples_r[j] = samples_l[j];
1569 }
else if (sbms && ch2_pres) {
1570 for (j = 0; j < 16; j++) {
1571 float tmp = samples_l[j];
1572 samples_l[j] = (
tmp + samples_r[j]) * 0.5
f;
1573 samples_r[j] = (
tmp - samples_r[j]) * 0.5
f;
1595 for (ch = ch1; ch <= ch2; ch++) {
1596 for (sb =
s->min_mono_subband; sb < s->nsubbands; sb++) {
1597 uint8_t *pt_st =
s->part_stereo[ch][(sb -
s->min_mono_subband) / 4];
1598 float *
samples =
s->time_samples[ch][sb];
1600 if (
s->ch_pres[ch2] & (1
U << sb))
1603 for (sf = 1; sf <= 4; sf++,
samples += 32) {
1607 for (
i = 0;
i < 32;
i++)
1618 int group,
int group_sf,
int synth_idx)
1620 int i, start, count;
1625 start =
s->tonal_bounds[group][group_sf][0];
1626 count = (
s->tonal_bounds[group][group_sf][1] - start) & (
DCA_LBR_TONES - 1);
1628 for (
i = 0;
i < count;
i++) {
1661 values[x_freq - 5] += cf[ 0] * -
s;
1662 p4:
values[x_freq - 4] += cf[ 1] *
c;
1663 p3:
values[x_freq - 3] += cf[ 2] *
s;
1664 p2:
values[x_freq - 2] += cf[ 3] * -
c;
1665 p1:
values[x_freq - 1] += cf[ 4] * -
s;
1666 p0:
values[x_freq ] += cf[ 5] *
c;
1667 values[x_freq + 1] += cf[ 6] *
s;
1668 values[x_freq + 2] += cf[ 7] * -
c;
1669 values[x_freq + 3] += cf[ 8] * -
s;
1670 values[x_freq + 4] += cf[ 9] *
c;
1671 values[x_freq + 5] += cf[10] *
s;
1686 for (group = 0; group < 5; group++) {
1687 int group_sf = (
s->framenum << group) + ((sf - 22) >> (5 - group));
1688 int synth_idx = ((((sf - 22) & 31) << group) & 31) + (1 << group) - 1;
1699 int sf, sb, nsubbands =
s->nsubbands, noutsubbands = 8 <<
s->freq_range;
1702 if (nsubbands < noutsubbands)
1703 memset(
values[nsubbands], 0, (noutsubbands - nsubbands) *
sizeof(
values[0]));
1707 s->dcadsp->lbr_bank(
values,
s->time_samples[ch],
1716 s->history[ch], noutsubbands * 4);
1717 s->fdsp->vector_fmul_reverse(
s->history[ch],
result[noutsubbands],
1718 s->window, noutsubbands * 4);
1719 output += noutsubbands * 4;
1723 for (sb = 0; sb < nsubbands; sb++) {
1732 int i,
ret, nchannels, ch_conf = (
s->ch_mask & 0x7) - 1;
1733 const int8_t *reorder;
1751 frame->nb_samples = 1024 <<
s->freq_range;
1756 for (
i = 0;
i < (
s->nchannels + 1) / 2;
i++) {
1758 int ch2 =
FFMIN(ch1 + 1,
s->nchannels - 1);
1766 if (ch1 != ch2 && (
s->part_stereo_pres & (1 << ch1)))
1769 if (ch1 < nchannels)
1772 if (ch1 != ch2 && ch2 < nchannels)
1780 s->lfe_history, 16 <<
s->freq_range);
1793 if (!
s->sample_rate)
1797 memset(
s->part_stereo, 16,
sizeof(
s->part_stereo));
1798 memset(
s->lpc_coeff, 0,
sizeof(
s->lpc_coeff));
1799 memset(
s->history, 0,
sizeof(
s->history));
1800 memset(
s->tonal_bounds, 0,
sizeof(
s->tonal_bounds));
1801 memset(
s->lfe_history, 0,
sizeof(
s->lfe_history));
1805 for (ch = 0; ch <
s->nchannels; ch++) {
1806 for (sb = 0; sb <
s->nsubbands; sb++) {
const uint8_t ff_dca_grid_2_to_scf[3]
@ AV_SAMPLE_FMT_FLTP
float, planar
static void skip_bits_long(GetBitContext *s, int n)
Skips the specified number of bits.
#define AV_LOG_WARNING
Something somehow does not look correct.
static int get_bits_left(GetBitContext *gb)
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel sample they are references to shared objects When the negotiation mechanism computes the intersection of the formats supported at each end of a all references to both lists are replaced with a reference to the intersection And when a single format is eventually chosen for a link amongst the remaining all references to the list are updated That means that if a filter requires that its input and output have the same format amongst a supported all it has to do is use a reference to the same list of formats query_formats can leave some formats unset and return AVERROR(EAGAIN) to cause the negotiation mechanism toagain later. That can be used by filters with complex requirements to use the format negotiated on one link to set the formats supported on another. Frame references ownership and permissions
const uint8_t ff_dca_grid_1_weights[12][32]
uint64_t channel_layout
Audio channel layout.
const uint8_t ff_dca_sb_reorder[8][8]
int sample_rate
samples per second
const float ff_dca_rsd_level_2b[2]
static void parse_ch(DCALbrDecoder *s, int ch, int sb, int quant_level, int flag)
Parse time samples for one subband, filling truncated samples with randomness.
static int ensure_bits(GetBitContext *s, int n)
Check point to ensure that enough bits are left.
const float ff_dca_lfe_step_size_24[144]
#define AV_CH_LAYOUT_MONO
filter_frame For filters that do not use the this method is called when a frame is pushed to the filter s input It can be called at any time except in a reentrant way If the input frame is enough to produce output
static void filter_ts(DCALbrDecoder *s, int ch1, int ch2)
static const uint8_t lfe_index[7]
This structure describes decoded (raw) audio or video data.
trying all byte sequences megabyte in length and selecting the best looking sequence will yield cases to try But a word about which is also called distortion Distortion can be quantified by almost any quality measurement one chooses the sum of squared differences is used but more complex methods that consider psychovisual effects can be used as well It makes no difference in this discussion First step
const float ff_dca_quant_amp[57]
@ LBR_CHUNK_RES_TS_2_LAST
static void synth_tones(DCALbrDecoder *s, int ch, float *values, int group, int group_sf, int synth_idx)
Synthesise tones in the given group for the given tonal subframe.
#define DCA_SPEAKER_LAYOUT_STEREO
int request_channel_layout
Converted from avctx.request_channel_layout.
uint8_t phs[DCA_LBR_CHANNELS]
Per-channel phase.
static av_always_inline int get_vlc2(GetBitContext *s, VLC_TYPE(*table)[2], int bits, int max_depth)
Parse a vlc code.
const uint8_t ff_dca_scf_to_grid_1[32]
@ LBR_FLAG_BAND_LIMIT_1_8
uint8_t x_freq
Spectral line offset.
static void decode_grid(DCALbrDecoder *s, int ch1, int ch2)
Reconstruct high-frequency resolution grid from first and third grids.
int lbr_offset
Offset to LBR component from start of substream.
static av_always_inline void bytestream2_skip(GetByteContext *g, unsigned int size)
static unsigned int get_bits(GetBitContext *s, int n)
Read 1-25 bits.
static void synth_lpc(DCALbrDecoder *s, int ch1, int ch2, int sb)
uint8_t amp[DCA_LBR_CHANNELS]
Per-channel amplitude.
const float ff_dca_rsd_level_5[5]
static av_always_inline float scale(float x, float s)
#define AV_CH_LAYOUT_STEREO
#define AV_LOG_ERROR
Something went wrong and cannot losslessly be recovered.
#define FF_ARRAY_ELEMS(a)
static int alloc_sample_buffer(DCALbrDecoder *s)
static int init_get_bits8(GetBitContext *s, const uint8_t *buffer, int byte_size)
Initialize GetBitContext.
static int parse_st_code(GetBitContext *s, int min_v)
#define AV_CH_LOW_FREQUENCY
av_cold void ff_dca_lbr_init_tables(void)
@ LBR_FLAG_BAND_LIMIT_1_2
static int parse_lpc(DCALbrDecoder *s, int ch1, int ch2, int start_sb, int end_sb)
static int parse_ts(DCALbrDecoder *s, int ch1, int ch2, int start_sb, int end_sb, int flag)
const float ff_dca_rsd_level_16[16]
#define av_assert0(cond)
assert() equivalent, that is always enabled.
static int init_sample_rate(DCALbrDecoder *s)
int bits_per_raw_sample
Bits per sample/pixel of internal libavcodec pixel/sample format.
static int parse_grid_1_chunk(DCALbrDecoder *s, LBRChunk *chunk, int ch1, int ch2)
@ LBR_FLAG_BAND_LIMIT_2_3
@ LBR_CHUNK_FRAME_NO_CSUM
@ LBR_FLAG_BAND_LIMIT_1_3
const int8_t ff_dca_lfe_delta_index_16[8]
static void random_ts(DCALbrDecoder *s, int ch1, int ch2)
Fill unallocated subbands with randomness.
int ff_dca_lbr_filter_frame(DCALbrDecoder *s, AVFrame *frame)
const float ff_dca_synth_env[32]
and forward the result(frame or status change) to the corresponding input. If nothing is possible
static __device__ float fabs(float a)
#define LOCAL_ALIGNED_32(t, v,...)
#define AVERROR_PATCHWELCOME
Not yet implemented in FFmpeg, patches welcome.
int lbr_size
Size of LBR component in extension substream.
static int parse_ts1_chunk(DCALbrDecoder *s, LBRChunk *chunk, int ch1, int ch2)
@ LBR_FLAG_BAND_LIMIT_MASK
int64_t bit_rate
the average bitrate
const uint16_t ff_dca_avg_g3_freqs[3]
const float ff_dca_corr_cf[32][11]
static unsigned int get_bits1(GetBitContext *s)
void av_fast_mallocz(void *ptr, unsigned int *size, size_t min_size)
Allocate and clear a buffer, reusing the given one if large enough.
const float ff_dca_lfe_iir[5][4]
@ LBR_FLAG_BAND_LIMIT_NONE
static int parse_lfe_chunk(DCALbrDecoder *s, LBRChunk *chunk)
#define AV_CH_FRONT_CENTER
static void transform_channel(DCALbrDecoder *s, int ch, float *output)
#define AV_EF_EXPLODE
abort decoding on minor error detection
#define DCA_LBR_CHANNELS_TOTAL
Undefined Behavior In the C some operations are like signed integer dereferencing freed accessing outside allocated Undefined Behavior must not occur in a C it is not safe even if the output of undefined operations is unused The unsafety may seem nit picking but Optimizing compilers have in fact optimized code on the assumption that no undefined Behavior occurs Optimizing code based on wrong assumptions can and has in some cases lead to effects beyond the output of computations The signed integer overflow problem in speed critical code Code which is highly optimized and works with signed integers sometimes has the problem that often the output of the computation does not c
@ LBR_CHUNK_RES_GRID_HR_LAST
static av_always_inline int bytestream2_get_bytes_left(GetByteContext *g)
#define AV_EF_CAREFUL
consider things that violate the spec, are fast to calculate and have not been seen in the wild as er...
const float ff_dca_bank_coeff[10]
@ DCA_LBR_HEADER_SYNC_ONLY
uint8_t ph_rot
Phase rotation.
static int parse_lfe_24(DCALbrDecoder *s)
int ff_get_buffer(AVCodecContext *avctx, AVFrame *frame, int flags)
Get a buffer for a frame.
static float lbr_rand(DCALbrDecoder *s, int sb)
@ AV_MATRIX_ENCODING_NONE
enum AVSampleFormat sample_fmt
audio sample format
void avpriv_report_missing_feature(void *avc, const char *msg,...) av_printf_format(2
Log a generic warning message about a missing feature.
static int parse_tonal_group(DCALbrDecoder *s, LBRChunk *chunk)
const float ff_dca_rsd_level_3[3]
The reader does not expect b to be semantically here and if the code is changed by maybe adding a a division or other the signedness will almost certainly be mistaken To avoid this confusion a new type was SUINT is the C unsigned type but it holds a signed int to use the same example SUINT a
@ LBR_CHUNK_TONAL_SCF_GRP_3
@ DCA_LBR_HEADER_DECODER_INIT
uint8_t f_delt
Difference between original and center frequency.
VLC ff_dca_vlc_tnl_grp[5]
int channels
number of audio channels
const float ff_dca_rsd_level_8[8]
#define AV_CH_LAYOUT_5POINT0
const float ff_dca_lfe_step_size_16[101]
const uint8_t ff_dca_scf_to_grid_2[32]
const float ff_dca_rsd_level_2a[2]
static int parse_vlc(GetBitContext *s, VLC *vlc, int max_depth)
#define i(width, name, range_min, range_max)
and forward the test the status of outputs and forward it to the corresponding return FFERROR_NOT_READY If the filters stores internally one or a few frame for some it can consider them to be part of the FIFO and delay acknowledging a status change accordingly Example code
const uint8_t ff_dca_rsd_pack_3_in_7[128][3]
static void base_func_synth(DCALbrDecoder *s, int ch, float *values, int sf)
Synthesise all tones in all groups for the given residual subframe.
VLC ff_dca_vlc_fst_rsd_amp
static void decode_part_stereo(DCALbrDecoder *s, int ch1, int ch2)
Modulate by interpolated partial stereo coefficients.
@ LBR_FLAG_BAND_LIMIT_1_4
it s the only field you need to keep assuming you have a context There is some magic you don t need to care about around this just let it vf default value
av_cold int ff_dca_lbr_init(DCALbrDecoder *s)
const uint8_t ff_dca_freq_to_sb[32]
@ LBR_CHUNK_RES_GRID_LR_LAST
static int parse_scale_factors(DCALbrDecoder *s, uint8_t *scf)
static const int8_t channel_reorder_nolfe[7][5]
static void predict(float *samples, const float *coeff, int nsamples)
static int parse_high_res_grid(DCALbrDecoder *s, LBRChunk *chunk, int ch1, int ch2)
av_cold void ff_dca_lbr_flush(DCALbrDecoder *s)
av_cold void ff_dca_lbr_close(DCALbrDecoder *s)
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several the filter must be ready for frames arriving randomly on any input any filter with several inputs will most likely require some kind of queuing mechanism It is perfectly acceptable to have a limited queue and to drop frames when the inputs are too unbalanced request_frame For filters that do not use the this method is called when a frame is wanted on an output For a it should directly call filter_frame on the corresponding output For a if there are queued frames already one of these frames should be pushed If the filter should request a frame on one of its repeatedly until at least one frame has been pushed Return or at least make progress towards producing a frame
static int parse_ts2_chunk(DCALbrDecoder *s, LBRChunk *chunk, int ch1, int ch2)
#define AV_CH_LAYOUT_SURROUND
#define AV_EF_CRCCHECK
Verify checksums embedded in the bitstream (could be of either encoded or decoded data,...
static int parse_decoder_init(DCALbrDecoder *s, GetByteContext *gb)
static volatile int checksum
static const int8_t channel_reorder_lfe[7][5]
Tag MUST be and< 10hcoeff half pel interpolation filter coefficients, hcoeff[0] are the 2 middle coefficients[1] are the next outer ones and so on, resulting in a filter like:...eff[2], hcoeff[1], hcoeff[0], hcoeff[0], hcoeff[1], hcoeff[2] ... the sign of the coefficients is not explicitly stored but alternates after each coeff and coeff[0] is positive, so ...,+,-,+,-,+,+,-,+,-,+,... hcoeff[0] is not explicitly stored but found by subtracting the sum of all stored coefficients with signs from 32 hcoeff[0]=32 - hcoeff[1] - hcoeff[2] - ... a good choice for hcoeff and htaps is htaps=6 hcoeff={40,-10, 2} an alternative which requires more computations at both encoder and decoder side and may or may not be better is htaps=8 hcoeff={42,-14, 6,-2}ref_frames minimum of the number of available reference frames and max_ref_frames for example the first frame after a key frame always has ref_frames=1spatial_decomposition_type wavelet type 0 is a 9/7 symmetric compact integer wavelet 1 is a 5/3 symmetric compact integer wavelet others are reserved stored as delta from last, last is reset to 0 if always_reset||keyframeqlog quality(logarithmic quantizer scale) stored as delta from last, last is reset to 0 if always_reset||keyframemv_scale stored as delta from last, last is reset to 0 if always_reset||keyframe FIXME check that everything works fine if this changes between framesqbias dequantization bias stored as delta from last, last is reset to 0 if always_reset||keyframeblock_max_depth maximum depth of the block tree stored as delta from last, last is reset to 0 if always_reset||keyframequant_table quantization tableHighlevel bitstream structure:==============================--------------------------------------------|Header|--------------------------------------------|------------------------------------|||Block0||||split?||||yes no||||......... intra?||||:Block01 :yes no||||:Block02 :....... ..........||||:Block03 ::y DC ::ref index:||||:Block04 ::cb DC ::motion x :||||......... :cr DC ::motion y :||||....... ..........|||------------------------------------||------------------------------------|||Block1|||...|--------------------------------------------|------------ ------------ ------------|||Y subbands||Cb subbands||Cr subbands||||--- ---||--- ---||--- ---|||||LL0||HL0||||LL0||HL0||||LL0||HL0|||||--- ---||--- ---||--- ---||||--- ---||--- ---||--- ---|||||LH0||HH0||||LH0||HH0||||LH0||HH0|||||--- ---||--- ---||--- ---||||--- ---||--- ---||--- ---|||||HL1||LH1||||HL1||LH1||||HL1||LH1|||||--- ---||--- ---||--- ---||||--- ---||--- ---||--- ---|||||HH1||HL2||||HH1||HL2||||HH1||HL2|||||...||...||...|||------------ ------------ ------------|--------------------------------------------Decoding process:=================------------|||Subbands|------------||||------------|Intra DC||||LL0 subband prediction ------------|\ Dequantization ------------------- \||Reference frames|\ IDWT|------- -------|Motion \|||Frame 0||Frame 1||Compensation . OBMC v -------|------- -------|--------------. \------> Frame n output Frame Frame<----------------------------------/|...|------------------- Range Coder:============Binary Range Coder:------------------- The implemented range coder is an adapted version based upon "Range encoding: an algorithm for removing redundancy from a digitised message." by G. N. N. Martin. The symbols encoded by the Snow range coder are bits(0|1). The associated probabilities are not fix but change depending on the symbol mix seen so far. bit seen|new state ---------+----------------------------------------------- 0|256 - state_transition_table[256 - old_state];1|state_transition_table[old_state];state_transition_table={ 0, 0, 0, 0, 0, 0, 0, 0, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 190, 191, 192, 194, 194, 195, 196, 197, 198, 199, 200, 201, 202, 202, 204, 205, 206, 207, 208, 209, 209, 210, 211, 212, 213, 215, 215, 216, 217, 218, 219, 220, 220, 222, 223, 224, 225, 226, 227, 227, 229, 229, 230, 231, 232, 234, 234, 235, 236, 237, 238, 239, 240, 241, 242, 243, 244, 245, 246, 247, 248, 248, 0, 0, 0, 0, 0, 0, 0};FIXME Range Coding of integers:------------------------- FIXME Neighboring Blocks:===================left and top are set to the respective blocks unless they are outside of the image in which case they are set to the Null block top-left is set to the top left block unless it is outside of the image in which case it is set to the left block if this block has no larger parent block or it is at the left side of its parent block and the top right block is not outside of the image then the top right block is used for top-right else the top-left block is used Null block y, cb, cr are 128 level, ref, mx and my are 0 Motion Vector Prediction:=========================1. the motion vectors of all the neighboring blocks are scaled to compensate for the difference of reference frames scaled_mv=(mv *(256 *(current_reference+1)/(mv.reference+1))+128)> the median of the scaled left
static void parse_grid_3(DCALbrDecoder *s, int ch1, int ch2, int sb, int flag)
static const uint8_t channel_counts[7]
main external API structure.
static int parse_grid_1_sec_ch(DCALbrDecoder *s, int ch2)
const uint32_t ff_dca_sampling_freqs[16]
static int parse_tonal_chunk(DCALbrDecoder *s, LBRChunk *chunk)
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several the filter must be ready for frames arriving randomly on any input any filter with several inputs will most likely require some kind of queuing mechanism It is perfectly acceptable to have a limited queue and to drop frames when the inputs are too unbalanced request_frame For filters that do not use the this method is called when a frame is wanted on an output For a it should directly call filter_frame on the corresponding output For a if there are queued frames already one of these frames should be pushed If the filter should request a frame on one of its repeatedly until at least one frame has been pushed Return values
Filter the word “frame” indicates either a video frame or a group of audio samples
int ff_dca_lbr_parse(DCALbrDecoder *s, uint8_t *data, DCAExssAsset *asset)
const uint8_t ff_dca_grid_1_to_scf[11]
@ LBR_CHUNK_TONAL_SCF_GRP_2
static int shift(int a, int b)
const float ff_dca_long_window[128]
int ff_side_data_update_matrix_encoding(AVFrame *frame, enum AVMatrixEncoding matrix_encoding)
Add or update AV_FRAME_DATA_MATRIXENCODING side data.
static void convert_lpc(float *coeff, const int *codes)
Convert from reflection coefficients to direct form coefficients.
const uint8_t ff_dca_freq_ranges[16]
static av_always_inline int get_bitsz(GetBitContext *s, int n)
Read 0-25 bits.
const uint16_t ff_dca_fst_amp[44]
static av_always_inline int diff(const uint32_t a, const uint32_t b)
static int parse_tonal(DCALbrDecoder *s, int group)
static const uint16_t channel_layouts[7]
av_cold AVFloatDSPContext * avpriv_float_dsp_alloc(int bit_exact)
Allocate a float DSP context.
static av_always_inline void bytestream2_init(GetByteContext *g, const uint8_t *buf, int buf_size)
@ LBR_CHUNK_RES_TS_1_LAST
static const double coeff[2][5]
const int8_t ff_dca_ph0_shift[8]
#define AVERROR_INVALIDDATA
Invalid data found when processing input.
#define DCA_LBR_TIME_HISTORY
static float cos_tab[256]
@ LBR_CHUNK_TONAL_SCF_GRP_5
VLC_TYPE(* table)[2]
code, bits
#define DCA_LBR_TIME_SAMPLES
@ LBR_CHUNK_TONAL_SCF_GRP_1
static int ff_dca_count_chs_for_mask(unsigned int mask)
Return number of individual channels in DCASpeakerPair mask.
@ LBR_CHUNK_TONAL_SCF_GRP_4
const uint16_t ff_dca_rsd_pack_5_in_8[256]
#define FF_PROFILE_DTS_EXPRESS
static int parse_lfe_16(DCALbrDecoder *s)
static int parse_grid_2(DCALbrDecoder *s, int ch1, int ch2, int start_sb, int end_sb, int flag)
const float ff_dca_st_coeff[34]
const int8_t ff_dca_lfe_delta_index_24[32]