Skip to content

F
ffmpeg.wasm-core
Commit ce87e630 authored by Rostislav Pehlivanov's avatar Rostislav Pehlivanov
Browse files
Options

opus_celt: deduplicate band quantization/dequantization function 

No point in having the same code twice to do exactly the same thing.
Signed-off-by: 's avatarRostislav Pehlivanov <atomnuker@gmail.com>
parent 86fda8be
Hide whitespace changes
Inline Side-by-side
Showing with 162 additions and 244 deletions
libavcodec/opus.c
View file @ ce87e630
...@@ -29,7 +29,8 @@ ...@@ -29,7 +29,8 @@
#include "libavutil/error.h" #include "libavutil/error.h"
#include "libavutil/ffmath.h" #include "libavutil/ffmath.h"
#include "opus.h" #include "opus_celt.h"
#include "opustab.h"
#include "vorbis.h" #include "vorbis.h"
static const uint16_t opus_frame_duration[32] = { static const uint16_t opus_frame_duration[32] = {
...@@ -438,3 +439,111 @@ av_cold int ff_opus_parse_extradata(AVCodecContext *avctx, ...@@ -438,3 +439,111 @@ av_cold int ff_opus_parse_extradata(AVCodecContext *avctx,
return 0; return 0;
} }
void ff_celt_quant_bands(CeltFrame *f, OpusRangeCoder *rc)
{
float lowband_scratch[8 * 22];
float norm1[2 * 8 * 100];
float *norm2 = norm1 + 8 * 100;
int totalbits = (f->framebits << 3) - f->anticollapse_needed;
int update_lowband = 1;
int lowband_offset = 0;
int i, j;
for (i = f->start_band; i < f->end_band; i++) {
uint32_t cm[2] = { (1 << f->blocks) - 1, (1 << f->blocks) - 1 };
int band_offset = ff_celt_freq_bands[i] << f->size;
int band_size = ff_celt_freq_range[i] << f->size;
float *X = f->block[0].coeffs + band_offset;
float *Y = (f->channels == 2) ? f->block[1].coeffs + band_offset : NULL;
float *norm_loc1, *norm_loc2;
int consumed = opus_rc_tell_frac(rc);
int effective_lowband = -1;
int b = 0;
/* Compute how many bits we want to allocate to this band */
if (i != f->start_band)
f->remaining -= consumed;
f->remaining2 = totalbits - consumed - 1;
if (i <= f->coded_bands - 1) {
int curr_balance = f->remaining / FFMIN(3, f->coded_bands-i);
b = av_clip_uintp2(FFMIN(f->remaining2 + 1, f->pulses[i] + curr_balance), 14);
}
if ((ff_celt_freq_bands[i] - ff_celt_freq_range[i] >= ff_celt_freq_bands[f->start_band] ||
i == f->start_band + 1) && (update_lowband || lowband_offset == 0))
lowband_offset = i;
if (i == f->start_band + 1) {
/* Special Hybrid Folding (RFC 8251 section 9). Copy the first band into
the second to ensure the second band never has to use the LCG. */
int offset = 8 * ff_celt_freq_bands[i];
int count = 8 * (ff_celt_freq_range[i] - ff_celt_freq_range[i-1]);
memcpy(&norm1[offset], &norm1[offset - count], count * sizeof(float));
if (f->channels == 2)
memcpy(&norm2[offset], &norm2[offset - count], count * sizeof(float));
}
/* Get a conservative estimate of the collapse_mask's for the bands we're
going to be folding from. */
if (lowband_offset != 0 && (f->spread != CELT_SPREAD_AGGRESSIVE ||
f->blocks > 1 || f->tf_change[i] < 0)) {
int foldstart, foldend;
/* This ensures we never repeat spectral content within one band */
effective_lowband = FFMAX(ff_celt_freq_bands[f->start_band],
ff_celt_freq_bands[lowband_offset] - ff_celt_freq_range[i]);
foldstart = lowband_offset;
while (ff_celt_freq_bands[--foldstart] > effective_lowband);
foldend = lowband_offset - 1;
while (++foldend < i && ff_celt_freq_bands[foldend] < effective_lowband + ff_celt_freq_range[i]);
cm[0] = cm[1] = 0;
for (j = foldstart; j < foldend; j++) {
cm[0] |= f->block[0].collapse_masks[j];
cm[1] |= f->block[f->channels - 1].collapse_masks[j];
}
}
if (f->dual_stereo && i == f->intensity_stereo) {
/* Switch off dual stereo to do intensity */
f->dual_stereo = 0;
for (j = ff_celt_freq_bands[f->start_band] << f->size; j < band_offset; j++)
norm1[j] = (norm1[j] + norm2[j]) / 2;
}
norm_loc1 = effective_lowband != -1 ? norm1 + (effective_lowband << f->size) : NULL;
norm_loc2 = effective_lowband != -1 ? norm2 + (effective_lowband << f->size) : NULL;
if (f->dual_stereo) {
cm[0] = f->pvq->quant_band(f->pvq, f, rc, i, X, NULL, band_size, b >> 1,
f->blocks, norm_loc1, f->size,
norm1 + band_offset, 0, 1.0f,
lowband_scratch, cm[0]);
cm[1] = f->pvq->quant_band(f->pvq, f, rc, i, Y, NULL, band_size, b >> 1,
f->blocks, norm_loc2, f->size,
norm2 + band_offset, 0, 1.0f,
lowband_scratch, cm[1]);
} else {
cm[0] = f->pvq->quant_band(f->pvq, f, rc, i, X, Y, band_size, b >> 0,
f->blocks, norm_loc1, f->size,
norm1 + band_offset, 0, 1.0f,
lowband_scratch, cm[0] | cm[1]);
cm[1] = cm[0];
}
f->block[0].collapse_masks[i] = (uint8_t)cm[0];
f->block[f->channels - 1].collapse_masks[i] = (uint8_t)cm[1];
f->remaining += f->pulses[i] + consumed;
/* Update the folding position only as long as we have 1 bit/sample depth */
update_lowband = (b > band_size << 3);
}
}
libavcodec/opus.h
View file @ ce87e630
...@@ -191,4 +191,7 @@ int ff_silk_decode_superframe(SilkContext *s, OpusRangeCoder *rc, ...@@ -191,4 +191,7 @@ int ff_silk_decode_superframe(SilkContext *s, OpusRangeCoder *rc,
enum OpusBandwidth bandwidth, int coded_channels, enum OpusBandwidth bandwidth, int coded_channels,
int duration_ms); int duration_ms);
/* Encode or decode CELT bands */
void ff_celt_quant_bands(CeltFrame *f, OpusRangeCoder *rc);
#endif /* AVCODEC_OPUS_H */ #endif /* AVCODEC_OPUS_H */
libavcodec/opus_celt.c
View file @ ce87e630
...@@ -676,110 +676,6 @@ static void process_anticollapse(CeltFrame *f, CeltBlock *block, float *X) ...@@ -676,110 +676,6 @@ static void process_anticollapse(CeltFrame *f, CeltBlock *block, float *X)
} }
} }
static void celt_decode_bands(CeltFrame *f, OpusRangeCoder *rc)
{
float lowband_scratch[8 * 22];
float norm[2 * 8 * 100];
int totalbits = (f->framebits << 3) - f->anticollapse_needed;
int update_lowband = 1;
int lowband_offset = 0;
int i, j;
memset(f->block[0].coeffs, 0, sizeof(f->block[0].coeffs));
memset(f->block[1].coeffs, 0, sizeof(f->block[0].coeffs));
for (i = f->start_band; i < f->end_band; i++) {
uint32_t cm[2] = { (1 << f->blocks) - 1, (1 << f->blocks) - 1 };
int band_offset = ff_celt_freq_bands[i] << f->size;
int band_size = ff_celt_freq_range[i] << f->size;
float *X = f->block[0].coeffs + band_offset;
float *Y = (f->channels == 2) ? f->block[1].coeffs + band_offset : NULL;
int consumed = opus_rc_tell_frac(rc);
float *norm2 = norm + 8 * 100;
int effective_lowband = -1;
int b = 0;
/* Compute how many bits we want to allocate to this band */
if (i != f->start_band)
f->remaining -= consumed;
f->remaining2 = totalbits - consumed - 1;
if (i <= f->coded_bands - 1) {
int curr_balance = f->remaining / FFMIN(3, f->coded_bands-i);
b = av_clip_uintp2(FFMIN(f->remaining2 + 1, f->pulses[i] + curr_balance), 14);
}
if ((ff_celt_freq_bands[i] - ff_celt_freq_range[i] >= ff_celt_freq_bands[f->start_band] ||
i == f->start_band + 1) && (update_lowband || lowband_offset == 0))
lowband_offset = i;
if (i == f->start_band + 1) {
/* Special Hybrid Folding (RFC 8251 section 9). Copy the first band into
the second to ensure the second band never has to use the LCG. */
int offset = 8 * ff_celt_freq_bands[i];
int count = 8 * (ff_celt_freq_range[i] - ff_celt_freq_range[i-1]);
memcpy(&norm[offset], &norm[offset - count], count * sizeof(float));
if (f->channels == 2)
memcpy(&norm2[offset], &norm2[offset - count], count * sizeof(float));
}
/* Get a conservative estimate of the collapse_mask's for the bands we're
going to be folding from. */
if (lowband_offset != 0 && (f->spread != CELT_SPREAD_AGGRESSIVE ||
f->blocks > 1 || f->tf_change[i] < 0)) {
int foldstart, foldend;
/* This ensures we never repeat spectral content within one band */
effective_lowband = FFMAX(ff_celt_freq_bands[f->start_band],
ff_celt_freq_bands[lowband_offset] - ff_celt_freq_range[i]);
foldstart = lowband_offset;
while (ff_celt_freq_bands[--foldstart] > effective_lowband);
foldend = lowband_offset - 1;
while (++foldend < i && ff_celt_freq_bands[foldend] < effective_lowband + ff_celt_freq_range[i]);
cm[0] = cm[1] = 0;
for (j = foldstart; j < foldend; j++) {
cm[0] |= f->block[0].collapse_masks[j];
cm[1] |= f->block[f->channels - 1].collapse_masks[j];
}
}
if (f->dual_stereo && i == f->intensity_stereo) {
/* Switch off dual stereo to do intensity */
f->dual_stereo = 0;
for (j = ff_celt_freq_bands[f->start_band] << f->size; j < band_offset; j++)
norm[j] = (norm[j] + norm2[j]) / 2;
}
if (f->dual_stereo) {
cm[0] = f->pvq->decode_band(f->pvq, f, rc, i, X, NULL, band_size, b / 2, f->blocks,
effective_lowband != -1 ? norm + (effective_lowband << f->size) : NULL, f->size,
norm + band_offset, 0, 1.0f, lowband_scratch, cm[0]);
cm[1] = f->pvq->decode_band(f->pvq, f, rc, i, Y, NULL, band_size, b/2, f->blocks,
effective_lowband != -1 ? norm2 + (effective_lowband << f->size) : NULL, f->size,
norm2 + band_offset, 0, 1.0f, lowband_scratch, cm[1]);
} else {
cm[0] = f->pvq->decode_band(f->pvq, f, rc, i, X, Y, band_size, b, f->blocks,
effective_lowband != -1 ? norm + (effective_lowband << f->size) : NULL, f->size,
norm + band_offset, 0, 1.0f, lowband_scratch, cm[0]|cm[1]);
cm[1] = cm[0];
}
f->block[0].collapse_masks[i] = (uint8_t)cm[0];
f->block[f->channels - 1].collapse_masks[i] = (uint8_t)cm[1];
f->remaining += f->pulses[i] + consumed;
/* Update the folding position only as long as we have 1 bit/sample depth */
update_lowband = (b > band_size << 3);
}
}
int ff_celt_decode_frame(CeltFrame *f, OpusRangeCoder *rc, int ff_celt_decode_frame(CeltFrame *f, OpusRangeCoder *rc,
float **output, int channels, int frame_size, float **output, int channels, int frame_size,
int start_band, int end_band) int start_band, int end_band)
...@@ -819,8 +715,10 @@ int ff_celt_decode_frame(CeltFrame *f, OpusRangeCoder *rc, ...@@ -819,8 +715,10 @@ int ff_celt_decode_frame(CeltFrame *f, OpusRangeCoder *rc,
if (!f->output_channels) if (!f->output_channels)
f->output_channels = channels; f->output_channels = channels;
memset(f->block[0].collapse_masks, 0, sizeof(f->block[0].collapse_masks)); for (i = 0; i < f->channels; i++) {
memset(f->block[1].collapse_masks, 0, sizeof(f->block[1].collapse_masks)); memset(f->block[i].coeffs, 0, sizeof(f->block[i].coeffs));
memset(f->block[i].collapse_masks, 0, sizeof(f->block[i].collapse_masks));
}
consumed = opus_rc_tell(rc); consumed = opus_rc_tell(rc);
...@@ -857,7 +755,7 @@ int ff_celt_decode_frame(CeltFrame *f, OpusRangeCoder *rc, ...@@ -857,7 +755,7 @@ int ff_celt_decode_frame(CeltFrame *f, OpusRangeCoder *rc,
celt_decode_tf_changes (f, rc); celt_decode_tf_changes (f, rc);
celt_decode_allocation (f, rc); celt_decode_allocation (f, rc);
celt_decode_fine_energy (f, rc); celt_decode_fine_energy (f, rc);
celt_decode_bands (f, rc); ff_celt_quant_bands (f, rc);
if (f->anticollapse_needed) if (f->anticollapse_needed)
f->anticollapse = ff_opus_rc_get_raw(rc, 1); f->anticollapse = ff_opus_rc_get_raw(rc, 1);
...@@ -1021,7 +919,7 @@ int ff_celt_init(AVCodecContext *avctx, CeltFrame **f, int output_channels, ...@@ -1021,7 +919,7 @@ int ff_celt_init(AVCodecContext *avctx, CeltFrame **f, int output_channels,
if ((ret = ff_mdct15_init(&frm->imdct[i], 1, i + 3, -1.0f/32768)) < 0) if ((ret = ff_mdct15_init(&frm->imdct[i], 1, i + 3, -1.0f/32768)) < 0)
goto fail; goto fail;
if ((ret = ff_celt_pvq_init(&frm->pvq)) < 0) if ((ret = ff_celt_pvq_init(&frm->pvq, 0)) < 0)
goto fail; goto fail;
frm->dsp = avpriv_float_dsp_alloc(avctx->flags & AV_CODEC_FLAG_BITEXACT); frm->dsp = avpriv_float_dsp_alloc(avctx->flags & AV_CODEC_FLAG_BITEXACT);
......
libavcodec/opus_pvq.c
View file @ ce87e630
...@@ -486,8 +486,7 @@ static av_always_inline uint32_t quant_band_template(CeltPVQ *pvq, CeltFrame *f, ...@@ -486,8 +486,7 @@ static av_always_inline uint32_t quant_band_template(CeltPVQ *pvq, CeltFrame *f,
int duration, float *lowband_out, int duration, float *lowband_out,
int level, float gain, int level, float gain,
float *lowband_scratch, float *lowband_scratch,
int fill, int quant, int fill, int quant)
QUANT_FN(*rec))
{ {
int i; int i;
const uint8_t *cache; const uint8_t *cache;
...@@ -700,8 +699,8 @@ static av_always_inline uint32_t quant_band_template(CeltPVQ *pvq, CeltFrame *f, ...@@ -700,8 +699,8 @@ static av_always_inline uint32_t quant_band_template(CeltPVQ *pvq, CeltFrame *f,
sign = 1 - 2 * sign; sign = 1 - 2 * sign;
/* We use orig_fill here because we want to fold the side, but if /* We use orig_fill here because we want to fold the side, but if
itheta==16384, we'll have cleared the low bits of fill. */ itheta==16384, we'll have cleared the low bits of fill. */
cm = rec(pvq, f, rc, band, x2, NULL, N, mbits, blocks, lowband, duration, cm = pvq->quant_band(pvq, f, rc, band, x2, NULL, N, mbits, blocks, lowband, duration,
lowband_out, level, gain, lowband_scratch, orig_fill); lowband_out, level, gain, lowband_scratch, orig_fill);
/* We don't split N=2 bands, so cm is either 1 or 0 (for a fold-collapse), /* We don't split N=2 bands, so cm is either 1 or 0 (for a fold-collapse),
and there's no need to worry about mixing with the other channel. */ and there's no need to worry about mixing with the other channel. */
y2[0] = -sign * x2[1]; y2[0] = -sign * x2[1];
...@@ -753,24 +752,25 @@ static av_always_inline uint32_t quant_band_template(CeltPVQ *pvq, CeltFrame *f, ...@@ -753,24 +752,25 @@ static av_always_inline uint32_t quant_band_template(CeltPVQ *pvq, CeltFrame *f,
if (mbits >= sbits) { if (mbits >= sbits) {
/* In stereo mode, we do not apply a scaling to the mid /* In stereo mode, we do not apply a scaling to the mid
* because we need the normalized mid for folding later */ * because we need the normalized mid for folding later */
cm = rec(pvq, f, rc, band, X, NULL, N, mbits, blocks, lowband, cm = pvq->quant_band(pvq, f, rc, band, X, NULL, N, mbits, blocks,
duration, next_lowband_out1, next_level, lowband, duration, next_lowband_out1, next_level,
stereo ? 1.0f : (gain * mid), lowband_scratch, fill); stereo ? 1.0f : (gain * mid), lowband_scratch, fill);
rebalance = mbits - (rebalance - f->remaining2); rebalance = mbits - (rebalance - f->remaining2);
if (rebalance > 3 << 3 && itheta != 0) if (rebalance > 3 << 3 && itheta != 0)
sbits += rebalance - (3 << 3); sbits += rebalance - (3 << 3);
/* For a stereo split, the high bits of fill are always zero, /* For a stereo split, the high bits of fill are always zero,
* so no folding will be done to the side. */ * so no folding will be done to the side. */
cmt = rec(pvq, f, rc, band, Y, NULL, N, sbits, blocks, next_lowband2, cmt = pvq->quant_band(pvq, f, rc, band, Y, NULL, N, sbits, blocks,
duration, NULL, next_level, gain * side, NULL, next_lowband2, duration, NULL, next_level,
fill >> blocks); gain * side, NULL, fill >> blocks);
cm |= cmt << ((B0 >> 1) & (stereo - 1)); cm |= cmt << ((B0 >> 1) & (stereo - 1));
} else { } else {
/* For a stereo split, the high bits of fill are always zero, /* For a stereo split, the high bits of fill are always zero,
* so no folding will be done to the side. */ * so no folding will be done to the side. */
cm = rec(pvq, f, rc, band, Y, NULL, N, sbits, blocks, next_lowband2, cm = pvq->quant_band(pvq, f, rc, band, Y, NULL, N, sbits, blocks,
duration, NULL, next_level, gain * side, NULL, fill >> blocks); next_lowband2, duration, NULL, next_level,
gain * side, NULL, fill >> blocks);
cm <<= ((B0 >> 1) & (stereo - 1)); cm <<= ((B0 >> 1) & (stereo - 1));
rebalance = sbits - (rebalance - f->remaining2); rebalance = sbits - (rebalance - f->remaining2);
if (rebalance > 3 << 3 && itheta != 16384) if (rebalance > 3 << 3 && itheta != 16384)
...@@ -778,9 +778,9 @@ static av_always_inline uint32_t quant_band_template(CeltPVQ *pvq, CeltFrame *f, ...@@ -778,9 +778,9 @@ static av_always_inline uint32_t quant_band_template(CeltPVQ *pvq, CeltFrame *f,
/* In stereo mode, we do not apply a scaling to the mid because /* In stereo mode, we do not apply a scaling to the mid because
* we need the normalized mid for folding later */ * we need the normalized mid for folding later */
cm |= rec(pvq, f, rc, band, X, NULL, N, mbits, blocks, lowband, duration, cm |= pvq->quant_band(pvq, f, rc, band, X, NULL, N, mbits, blocks,
next_lowband_out1, next_level, stereo ? 1.0f : (gain * mid), lowband, duration, next_lowband_out1, next_level,
lowband_scratch, fill); stereo ? 1.0f : (gain * mid), lowband_scratch, fill);
} }
} }
} else { } else {
...@@ -874,19 +874,16 @@ static av_always_inline uint32_t quant_band_template(CeltPVQ *pvq, CeltFrame *f, ...@@ -874,19 +874,16 @@ static av_always_inline uint32_t quant_band_template(CeltPVQ *pvq, CeltFrame *f,
return cm; return cm;
} }
static QUANT_FN(pvq_decode_band) static QUANT_FN(pvq_decode_band)
{ {
return quant_band_template(pvq, f, rc, band, X, Y, N, b, blocks, lowband, duration, return quant_band_template(pvq, f, rc, band, X, Y, N, b, blocks, lowband, duration,
lowband_out, level, gain, lowband_scratch, fill, 0, lowband_out, level, gain, lowband_scratch, fill, 0);
pvq->decode_band);
} }
static QUANT_FN(pvq_encode_band) static QUANT_FN(pvq_encode_band)
{ {
return quant_band_template(pvq, f, rc, band, X, Y, N, b, blocks, lowband, duration, return quant_band_template(pvq, f, rc, band, X, Y, N, b, blocks, lowband, duration,
lowband_out, level, gain, lowband_scratch, fill, 1, lowband_out, level, gain, lowband_scratch, fill, 1);
pvq->encode_band);
} }
static float pvq_band_cost(CeltPVQ *pvq, CeltFrame *f, OpusRangeCoder *rc, int band, static float pvq_band_cost(CeltPVQ *pvq, CeltFrame *f, OpusRangeCoder *rc, int band,
...@@ -914,14 +911,14 @@ static float pvq_band_cost(CeltPVQ *pvq, CeltFrame *f, OpusRangeCoder *rc, int b ...@@ -914,14 +911,14 @@ static float pvq_band_cost(CeltPVQ *pvq, CeltFrame *f, OpusRangeCoder *rc, int b
} }
if (f->dual_stereo) { if (f->dual_stereo) {
pvq->encode_band(pvq, f, rc, band, X, NULL, band_size, b / 2, f->blocks, NULL, pvq->quant_band(pvq, f, rc, band, X, NULL, band_size, b / 2, f->blocks, NULL,
f->size, norm1, 0, 1.0f, lowband_scratch, cm[0]); f->size, norm1, 0, 1.0f, lowband_scratch, cm[0]);
pvq->encode_band(pvq, f, rc, band, Y, NULL, band_size, b / 2, f->blocks, NULL, pvq->quant_band(pvq, f, rc, band, Y, NULL, band_size, b / 2, f->blocks, NULL,
f->size, norm2, 0, 1.0f, lowband_scratch, cm[1]); f->size, norm2, 0, 1.0f, lowband_scratch, cm[1]);
} else { } else {
pvq->encode_band(pvq, f, rc, band, X, Y, band_size, b, f->blocks, NULL, f->size, pvq->quant_band(pvq, f, rc, band, X, Y, band_size, b, f->blocks, NULL, f->size,
norm1, 0, 1.0f, lowband_scratch, cm[0] | cm[1]); norm1, 0, 1.0f, lowband_scratch, cm[0] | cm[1]);
} }
for (i = 0; i < band_size; i++) { for (i = 0; i < band_size; i++) {
...@@ -939,16 +936,15 @@ static float pvq_band_cost(CeltPVQ *pvq, CeltFrame *f, OpusRangeCoder *rc, int b ...@@ -939,16 +936,15 @@ static float pvq_band_cost(CeltPVQ *pvq, CeltFrame *f, OpusRangeCoder *rc, int b
return lambda*dist*cost; return lambda*dist*cost;
} }
int av_cold ff_celt_pvq_init(CeltPVQ **pvq) int av_cold ff_celt_pvq_init(CeltPVQ **pvq, int encode)
{ {
CeltPVQ *s = av_malloc(sizeof(CeltPVQ)); CeltPVQ *s = av_malloc(sizeof(CeltPVQ));
if (!s) if (!s)
return AVERROR(ENOMEM); return AVERROR(ENOMEM);
s->pvq_search = ppp_pvq_search_c; s->pvq_search = ppp_pvq_search_c;
s->decode_band = pvq_decode_band; s->quant_band = encode ? pvq_encode_band : pvq_decode_band;
s->encode_band = pvq_encode_band; s->band_cost = pvq_band_cost;
s->band_cost = pvq_band_cost;
if (ARCH_X86) if (ARCH_X86)
ff_opus_dsp_init_x86(s); ff_opus_dsp_init_x86(s);
......
libavcodec/opus_pvq.h
View file @ ce87e630
...@@ -38,14 +38,14 @@ struct CeltPVQ { ...@@ -38,14 +38,14 @@ struct CeltPVQ {
float (*pvq_search)(float *X, int *y, int K, int N); float (*pvq_search)(float *X, int *y, int K, int N);
QUANT_FN(*decode_band); QUANT_FN(*quant_band);
QUANT_FN(*encode_band);
float (*band_cost)(struct CeltPVQ *pvq, CeltFrame *f, OpusRangeCoder *rc, float (*band_cost)(struct CeltPVQ *pvq, CeltFrame *f, OpusRangeCoder *rc,
int band, float *bits, float lambda); int band, float *bits, float lambda);
}; };
int ff_celt_pvq_init (struct CeltPVQ **pvq);
void ff_opus_dsp_init_x86(struct CeltPVQ *s); void ff_opus_dsp_init_x86(struct CeltPVQ *s);
int ff_celt_pvq_init(struct CeltPVQ **pvq, int encode);
void ff_celt_pvq_uninit(struct CeltPVQ **pvq); void ff_celt_pvq_uninit(struct CeltPVQ **pvq);
#endif /* AVCODEC_OPUS_PVQ_H */ #endif /* AVCODEC_OPUS_PVQ_H */
libavcodec/opusenc.c
View file @ ce87e630
...@@ -255,7 +255,7 @@ static void celt_frame_mdct(OpusEncContext *s, CeltFrame *f) ...@@ -255,7 +255,7 @@ static void celt_frame_mdct(OpusEncContext *s, CeltFrame *f)
} }
} }
static void celt_enc_tf(OpusRangeCoder *rc, CeltFrame *f) static void celt_enc_tf(CeltFrame *f, OpusRangeCoder *rc)
{ {
int i, tf_select = 0, diff = 0, tf_changed = 0, tf_select_needed; int i, tf_select = 0, diff = 0, tf_changed = 0, tf_select_needed;
int bits = f->transient ? 2 : 4; int bits = f->transient ? 2 : 4;
...@@ -282,7 +282,7 @@ static void celt_enc_tf(OpusRangeCoder *rc, CeltFrame *f) ...@@ -282,7 +282,7 @@ static void celt_enc_tf(OpusRangeCoder *rc, CeltFrame *f)
f->tf_change[i] = ff_celt_tf_select[f->size][f->transient][tf_select][f->tf_change[i]]; f->tf_change[i] = ff_celt_tf_select[f->size][f->transient][tf_select][f->tf_change[i]];
} }
void ff_celt_enc_bitalloc(OpusRangeCoder *rc, CeltFrame *f) void ff_celt_enc_bitalloc(CeltFrame *f, OpusRangeCoder *rc)
{ {
int i, j, low, high, total, done, bandbits, remaining, tbits_8ths; int i, j, low, high, total, done, bandbits, remaining, tbits_8ths;
int skip_startband = f->start_band; int skip_startband = f->start_band;
...@@ -690,7 +690,7 @@ static void exp_quant_coarse(OpusRangeCoder *rc, CeltFrame *f, ...@@ -690,7 +690,7 @@ static void exp_quant_coarse(OpusRangeCoder *rc, CeltFrame *f,
} }
} }
static void celt_quant_coarse(OpusRangeCoder *rc, CeltFrame *f, static void celt_quant_coarse(CeltFrame *f, OpusRangeCoder *rc,
float last_energy[][CELT_MAX_BANDS]) float last_energy[][CELT_MAX_BANDS])
{ {
uint32_t inter, intra; uint32_t inter, intra;
...@@ -710,7 +710,7 @@ static void celt_quant_coarse(OpusRangeCoder *rc, CeltFrame *f, ...@@ -710,7 +710,7 @@ static void celt_quant_coarse(OpusRangeCoder *rc, CeltFrame *f,
} }
} }
static void celt_quant_fine(OpusRangeCoder *rc, CeltFrame *f) static void celt_quant_fine(CeltFrame *f, OpusRangeCoder *rc)
{ {
int i, ch; int i, ch;
for (i = f->start_band; i < f->end_band; i++) { for (i = f->start_band; i < f->end_band; i++) {
...@@ -747,95 +747,6 @@ static void celt_quant_final(OpusEncContext *s, OpusRangeCoder *rc, CeltFrame *f ...@@ -747,95 +747,6 @@ static void celt_quant_final(OpusEncContext *s, OpusRangeCoder *rc, CeltFrame *f
} }
} }
static void celt_quant_bands(OpusRangeCoder *rc, CeltFrame *f)
{
float lowband_scratch[8 * 22];
float norm[2 * 8 * 100];
int totalbits = (f->framebits << 3) - f->anticollapse_needed;
int update_lowband = 1;
int lowband_offset = 0;
int i, j;
for (i = f->start_band; i < f->end_band; i++) {
uint32_t cm[2] = { (1 << f->blocks) - 1, (1 << f->blocks) - 1 };
int band_offset = ff_celt_freq_bands[i] << f->size;
int band_size = ff_celt_freq_range[i] << f->size;
float *X = f->block[0].coeffs + band_offset;
float *Y = (f->channels == 2) ? f->block[1].coeffs + band_offset : NULL;
int consumed = opus_rc_tell_frac(rc);
float *norm2 = norm + 8 * 100;
int effective_lowband = -1;
int b = 0;
/* Compute how many bits we want to allocate to this band */
if (i != f->start_band)
f->remaining -= consumed;
f->remaining2 = totalbits - consumed - 1;
if (i <= f->coded_bands - 1) {
int curr_balance = f->remaining / FFMIN(3, f->coded_bands-i);
b = av_clip_uintp2(FFMIN(f->remaining2 + 1, f->pulses[i] + curr_balance), 14);
}
if (ff_celt_freq_bands[i] - ff_celt_freq_range[i] >= ff_celt_freq_bands[f->start_band] &&
(update_lowband || lowband_offset == 0))
lowband_offset = i;
/* Get a conservative estimate of the collapse_mask's for the bands we're
going to be folding from. */
if (lowband_offset != 0 && (f->spread != CELT_SPREAD_AGGRESSIVE ||
f->blocks > 1 || f->tf_change[i] < 0)) {
int foldstart, foldend;
/* This ensures we never repeat spectral content within one band */
effective_lowband = FFMAX(ff_celt_freq_bands[f->start_band],
ff_celt_freq_bands[lowband_offset] - ff_celt_freq_range[i]);
foldstart = lowband_offset;
while (ff_celt_freq_bands[--foldstart] > effective_lowband);
foldend = lowband_offset - 1;
while (ff_celt_freq_bands[++foldend] < effective_lowband + ff_celt_freq_range[i]);
cm[0] = cm[1] = 0;
for (j = foldstart; j < foldend; j++) {
cm[0] |= f->block[0].collapse_masks[j];
cm[1] |= f->block[f->channels - 1].collapse_masks[j];
}
}
if (f->dual_stereo && i == f->intensity_stereo) {
/* Switch off dual stereo to do intensity */
f->dual_stereo = 0;
for (j = ff_celt_freq_bands[f->start_band] << f->size; j < band_offset; j++)
norm[j] = (norm[j] + norm2[j]) / 2;
}
if (f->dual_stereo) {
cm[0] = f->pvq->encode_band(f->pvq, f, rc, i, X, NULL, band_size, b / 2, f->blocks,
effective_lowband != -1 ? norm + (effective_lowband << f->size) : NULL, f->size,
norm + band_offset, 0, 1.0f, lowband_scratch, cm[0]);
cm[1] = f->pvq->encode_band(f->pvq, f, rc, i, Y, NULL, band_size, b / 2, f->blocks,
effective_lowband != -1 ? norm2 + (effective_lowband << f->size) : NULL, f->size,
norm2 + band_offset, 0, 1.0f, lowband_scratch, cm[1]);
} else {
cm[0] = f->pvq->encode_band(f->pvq, f, rc, i, X, Y, band_size, b, f->blocks,
effective_lowband != -1 ? norm + (effective_lowband << f->size) : NULL, f->size,
norm + band_offset, 0, 1.0f, lowband_scratch, cm[0] | cm[1]);
cm[1] = cm[0];
}
f->block[0].collapse_masks[i] = (uint8_t)cm[0];
f->block[f->channels - 1].collapse_masks[i] = (uint8_t)cm[1];
f->remaining += f->pulses[i] + consumed;
/* Update the folding position only as long as we have 1 bit/sample depth */
update_lowband = (b > band_size << 3);
}
}
static void celt_encode_frame(OpusEncContext *s, OpusRangeCoder *rc, static void celt_encode_frame(OpusEncContext *s, OpusRangeCoder *rc,
CeltFrame *f, int index) CeltFrame *f, int index)
{ {
...@@ -883,11 +794,11 @@ static void celt_encode_frame(OpusEncContext *s, OpusRangeCoder *rc, ...@@ -883,11 +794,11 @@ static void celt_encode_frame(OpusEncContext *s, OpusRangeCoder *rc,
ff_opus_rc_enc_log(rc, f->transient, 3); ff_opus_rc_enc_log(rc, f->transient, 3);
/* Main encoding */ /* Main encoding */
celt_quant_coarse(rc, f, s->last_quantized_energy); celt_quant_coarse (f, rc, s->last_quantized_energy);
celt_enc_tf (rc, f); celt_enc_tf (f, rc);
ff_celt_enc_bitalloc(rc, f); ff_celt_enc_bitalloc(f, rc);
celt_quant_fine (rc, f); celt_quant_fine (f, rc);
celt_quant_bands (rc, f); ff_celt_quant_bands (f, rc);
/* Anticollapse bit */ /* Anticollapse bit */
if (f->anticollapse_needed) if (f->anticollapse_needed)
...@@ -1080,7 +991,7 @@ static av_cold int opus_encode_init(AVCodecContext *avctx) ...@@ -1080,7 +991,7 @@ static av_cold int opus_encode_init(AVCodecContext *avctx)
ff_af_queue_init(avctx, &s->afq); ff_af_queue_init(avctx, &s->afq);
if ((ret = ff_celt_pvq_init(&s->pvq)) < 0) if ((ret = ff_celt_pvq_init(&s->pvq, 1)) < 0)
return ret; return ret;
if (!(s->dsp = avpriv_float_dsp_alloc(avctx->flags & AV_CODEC_FLAG_BITEXACT))) if (!(s->dsp = avpriv_float_dsp_alloc(avctx->flags & AV_CODEC_FLAG_BITEXACT)))
...@@ -1117,6 +1028,7 @@ static av_cold int opus_encode_init(AVCodecContext *avctx) ...@@ -1117,6 +1028,7 @@ static av_cold int opus_encode_init(AVCodecContext *avctx)
s->frame[i].avctx = s->avctx; s->frame[i].avctx = s->avctx;
s->frame[i].seed = 0; s->frame[i].seed = 0;
s->frame[i].pvq = s->pvq; s->frame[i].pvq = s->pvq;
s->frame[i].apply_phase_inv = 1;
s->frame[i].block[0].emph_coeff = s->frame[i].block[1].emph_coeff = 0.0f; s->frame[i].block[0].emph_coeff = s->frame[i].block[1].emph_coeff = 0.0f;
} }
......
libavcodec/opusenc.h
View file @ ce87e630
...@@ -51,6 +51,6 @@ typedef struct OpusPacketInfo { ...@@ -51,6 +51,6 @@ typedef struct OpusPacketInfo {
int frames; int frames;
} OpusPacketInfo; } OpusPacketInfo;
void ff_celt_enc_bitalloc(OpusRangeCoder *rc, CeltFrame *f); void ff_celt_enc_bitalloc(CeltFrame *f, OpusRangeCoder *rc);
#endif /* AVCODEC_OPUSENC_H */ #endif /* AVCODEC_OPUSENC_H */
libavcodec/opusenc_psy.c
View file @ ce87e630
...@@ -316,7 +316,7 @@ static int bands_dist(OpusPsyContext *s, CeltFrame *f, float *total_dist) ...@@ -316,7 +316,7 @@ static int bands_dist(OpusPsyContext *s, CeltFrame *f, float *total_dist)
OpusRangeCoder dump; OpusRangeCoder dump;
ff_opus_rc_enc_init(&dump); ff_opus_rc_enc_init(&dump);
ff_celt_enc_bitalloc(&dump, f); ff_celt_enc_bitalloc(f, &dump);
for (i = 0; i < CELT_MAX_BANDS; i++) { for (i = 0; i < CELT_MAX_BANDS; i++) {
float bits = 0.0f; float bits = 0.0f;
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment