* AAC encoder
* Copyright (C) 2008 Konstantin Shishkov
*
- * This file is part of FFmpeg.
+ * This file is part of Libav.
*
- * FFmpeg is free software; you can redistribute it and/or
+ * Libav is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
- * FFmpeg is distributed in the hope that it will be useful,
+ * Libav is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
- * License along with FFmpeg; if not, write to the Free Software
+ * License along with Libav; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
* add temporal noise shaping
***********************************/
+#include "libavutil/opt.h"
#include "avcodec.h"
#include "put_bits.h"
#include "dsputil.h"
#include "mpeg4audio.h"
+#include "kbdwin.h"
+#include "sinewin.h"
#include "aac.h"
#include "aactab.h"
AACEncContext *s = avctx->priv_data;
int i;
const uint8_t *sizes[2];
+ uint8_t grouping[AAC_MAX_CHANNELS];
int lengths[2];
avctx->frame_size = 1024;
ff_init_ff_sine_windows(10);
ff_init_ff_sine_windows(7);
+ s->chan_map = aac_chan_configs[avctx->channels-1];
s->samples = av_malloc(2 * 1024 * avctx->channels * sizeof(s->samples[0]));
- s->cpe = av_mallocz(sizeof(ChannelElement) * aac_chan_configs[avctx->channels-1][0]);
+ s->cpe = av_mallocz(sizeof(ChannelElement) * s->chan_map[0]);
avctx->extradata = av_mallocz(5 + FF_INPUT_BUFFER_PADDING_SIZE);
avctx->extradata_size = 5;
put_audio_specific_config(avctx);
sizes[1] = swb_size_128[i];
lengths[0] = ff_aac_num_swb_1024[i];
lengths[1] = ff_aac_num_swb_128[i];
- ff_psy_init(&s->psy, avctx, 2, sizes, lengths);
+ for (i = 0; i < s->chan_map[0]; i++)
+ grouping[i] = s->chan_map[i + 1] == TYPE_CPE;
+ ff_psy_init(&s->psy, avctx, 2, sizes, lengths, s->chan_map[0], grouping);
s->psypp = ff_psy_preprocess_init(avctx);
s->coder = &ff_aac_coders[2];
const float * lwindow = sce->ics.use_kb_window[0] ? ff_aac_kbd_long_1024 : ff_sine_1024;
const float * swindow = sce->ics.use_kb_window[0] ? ff_aac_kbd_short_128 : ff_sine_128;
const float * pwindow = sce->ics.use_kb_window[1] ? ff_aac_kbd_short_128 : ff_sine_128;
+ float *output = sce->ret;
if (sce->ics.window_sequence[0] != EIGHT_SHORT_SEQUENCE) {
- memcpy(s->output, sce->saved, sizeof(float)*1024);
+ memcpy(output, sce->saved, sizeof(float)*1024);
if (sce->ics.window_sequence[0] == LONG_STOP_SEQUENCE) {
- memset(s->output, 0, sizeof(s->output[0]) * 448);
+ memset(output, 0, sizeof(output[0]) * 448);
for (i = 448; i < 576; i++)
- s->output[i] = sce->saved[i] * pwindow[i - 448];
+ output[i] = sce->saved[i] * pwindow[i - 448];
for (i = 576; i < 704; i++)
- s->output[i] = sce->saved[i];
+ output[i] = sce->saved[i];
}
if (sce->ics.window_sequence[0] != LONG_START_SEQUENCE) {
for (i = 0; i < 1024; i++) {
- s->output[i+1024] = audio[i * chans] * lwindow[1024 - i - 1];
+ output[i+1024] = audio[i * chans] * lwindow[1024 - i - 1];
sce->saved[i] = audio[i * chans] * lwindow[i];
}
} else {
for (i = 0; i < 448; i++)
- s->output[i+1024] = audio[i * chans];
+ output[i+1024] = audio[i * chans];
for (; i < 576; i++)
- s->output[i+1024] = audio[i * chans] * swindow[576 - i - 1];
- memset(s->output+1024+576, 0, sizeof(s->output[0]) * 448);
+ output[i+1024] = audio[i * chans] * swindow[576 - i - 1];
+ memset(output+1024+576, 0, sizeof(output[0]) * 448);
for (i = 0; i < 1024; i++)
sce->saved[i] = audio[i * chans];
}
- ff_mdct_calc(&s->mdct1024, sce->coeffs, s->output);
+ s->mdct1024.mdct_calc(&s->mdct1024, sce->coeffs, output);
} else {
for (k = 0; k < 1024; k += 128) {
for (i = 448 + k; i < 448 + k + 256; i++)
- s->output[i - 448 - k] = (i < 1024)
+ output[i - 448 - k] = (i < 1024)
? sce->saved[i]
: audio[(i-1024)*chans];
- s->dsp.vector_fmul (s->output, s->output, k ? swindow : pwindow, 128);
- s->dsp.vector_fmul_reverse(s->output+128, s->output+128, swindow, 128);
- ff_mdct_calc(&s->mdct128, sce->coeffs + k, s->output);
+ s->dsp.vector_fmul (output, output, k ? swindow : pwindow, 128);
+ s->dsp.vector_fmul_reverse(output+128, output+128, swindow, 128);
+ s->mdct128.mdct_calc(&s->mdct128, sce->coeffs + k, output);
}
for (i = 0; i < 1024; i++)
sce->saved[i] = audio[i * chans];
if (msc == 0 || ics0->max_sfb == 0)
cpe->ms_mode = 0;
else
- cpe->ms_mode = msc < ics0->max_sfb ? 1 : 2;
+ cpe->ms_mode = msc < ics0->max_sfb * ics0->num_windows ? 1 : 2;
}
}
AACEncContext *s = avctx->priv_data;
int16_t *samples = s->samples, *samples2, *la;
ChannelElement *cpe;
- int i, j, chans, tag, start_ch;
- const uint8_t *chan_map = aac_chan_configs[avctx->channels-1];
+ int i, ch, w, g, chans, tag, start_ch;
int chan_el_counter[4];
FFPsyWindowInfo windows[AAC_MAX_CHANNELS];
} else {
start_ch = 0;
samples2 = s->samples + 1024 * avctx->channels;
- for (i = 0; i < chan_map[0]; i++) {
- tag = chan_map[i+1];
+ for (i = 0; i < s->chan_map[0]; i++) {
+ tag = s->chan_map[i+1];
chans = tag == TYPE_CPE ? 2 : 1;
ff_psy_preprocess(s->psypp, (uint16_t*)data + start_ch,
samples2 + start_ch, start_ch, chans);
}
start_ch = 0;
- for (i = 0; i < chan_map[0]; i++) {
+ for (i = 0; i < s->chan_map[0]; i++) {
FFPsyWindowInfo* wi = windows + start_ch;
- tag = chan_map[i+1];
+ tag = s->chan_map[i+1];
chans = tag == TYPE_CPE ? 2 : 1;
cpe = &s->cpe[i];
- for (j = 0; j < chans; j++) {
- IndividualChannelStream *ics = &cpe->ch[j].ics;
- int k;
- int cur_channel = start_ch + j;
+ for (ch = 0; ch < chans; ch++) {
+ IndividualChannelStream *ics = &cpe->ch[ch].ics;
+ int cur_channel = start_ch + ch;
samples2 = samples + cur_channel;
la = samples2 + (448+64) * avctx->channels;
if (!data)
la = NULL;
if (tag == TYPE_LFE) {
- wi[j].window_type[0] = ONLY_LONG_SEQUENCE;
- wi[j].window_shape = 0;
- wi[j].num_windows = 1;
- wi[j].grouping[0] = 1;
+ wi[ch].window_type[0] = ONLY_LONG_SEQUENCE;
+ wi[ch].window_shape = 0;
+ wi[ch].num_windows = 1;
+ wi[ch].grouping[0] = 1;
+
+ /* Only the lowest 12 coefficients are used in a LFE channel.
+ * The expression below results in only the bottom 8 coefficients
+ * being used for 11.025kHz to 16kHz sample rates.
+ */
+ ics->num_swb = s->samplerate_index >= 8 ? 1 : 3;
} else {
- wi[j] = ff_psy_suggest_window(&s->psy, samples2, la, cur_channel,
+ wi[ch] = s->psy.model->window(&s->psy, samples2, la, cur_channel,
ics->window_sequence[0]);
}
ics->window_sequence[1] = ics->window_sequence[0];
- ics->window_sequence[0] = wi[j].window_type[0];
+ ics->window_sequence[0] = wi[ch].window_type[0];
ics->use_kb_window[1] = ics->use_kb_window[0];
- ics->use_kb_window[0] = wi[j].window_shape;
- ics->num_windows = wi[j].num_windows;
+ ics->use_kb_window[0] = wi[ch].window_shape;
+ ics->num_windows = wi[ch].num_windows;
ics->swb_sizes = s->psy.bands [ics->num_windows == 8];
- ics->num_swb = tag == TYPE_LFE ? 12 : s->psy.num_bands[ics->num_windows == 8];
- for (k = 0; k < ics->num_windows; k++)
- ics->group_len[k] = wi[j].grouping[k];
+ ics->num_swb = tag == TYPE_LFE ? ics->num_swb : s->psy.num_bands[ics->num_windows == 8];
+ for (w = 0; w < ics->num_windows; w++)
+ ics->group_len[w] = wi[ch].grouping[w];
- apply_window_and_mdct(avctx, s, &cpe->ch[j], samples2);
+ apply_window_and_mdct(avctx, s, &cpe->ch[ch], samples2);
}
start_ch += chans;
}
put_bitstream_info(avctx, s, LIBAVCODEC_IDENT);
start_ch = 0;
memset(chan_el_counter, 0, sizeof(chan_el_counter));
- for (i = 0; i < chan_map[0]; i++) {
+ for (i = 0; i < s->chan_map[0]; i++) {
FFPsyWindowInfo* wi = windows + start_ch;
- tag = chan_map[i+1];
+ const float *coeffs[2];
+ tag = s->chan_map[i+1];
chans = tag == TYPE_CPE ? 2 : 1;
cpe = &s->cpe[i];
put_bits(&s->pb, 3, tag);
put_bits(&s->pb, 4, chan_el_counter[tag]++);
- for (j = 0; j < chans; j++) {
- s->cur_channel = start_ch + j;
- ff_psy_set_band_info(&s->psy, s->cur_channel, cpe->ch[j].coeffs, &wi[j]);
- s->coder->search_for_quantizers(avctx, s, &cpe->ch[j], s->lambda);
+ for (ch = 0; ch < chans; ch++)
+ coeffs[ch] = cpe->ch[ch].coeffs;
+ s->psy.model->analyze(&s->psy, start_ch, coeffs, wi);
+ for (ch = 0; ch < chans; ch++) {
+ s->cur_channel = start_ch * 2 + ch;
+ s->coder->search_for_quantizers(avctx, s, &cpe->ch[ch], s->lambda);
}
cpe->common_window = 0;
if (chans > 1
&& wi[0].window_shape == wi[1].window_shape) {
cpe->common_window = 1;
- for (j = 0; j < wi[0].num_windows; j++) {
- if (wi[0].grouping[j] != wi[1].grouping[j]) {
+ for (w = 0; w < wi[0].num_windows; w++) {
+ if (wi[0].grouping[w] != wi[1].grouping[w]) {
cpe->common_window = 0;
break;
}
}
}
- s->cur_channel = start_ch;
- if (cpe->common_window && s->coder->search_for_ms)
- s->coder->search_for_ms(s, cpe, s->lambda);
+ s->cur_channel = start_ch * 2;
+ if (s->options.stereo_mode && cpe->common_window) {
+ if (s->options.stereo_mode > 0) {
+ IndividualChannelStream *ics = &cpe->ch[0].ics;
+ for (w = 0; w < ics->num_windows; w += ics->group_len[w])
+ for (g = 0; g < ics->num_swb; g++)
+ cpe->ms_mask[w*16+g] = 1;
+ } else if (s->coder->search_for_ms) {
+ s->coder->search_for_ms(s, cpe, s->lambda);
+ }
+ }
adjust_frame_information(s, cpe, chans);
if (chans == 2) {
put_bits(&s->pb, 1, cpe->common_window);
encode_ms_info(&s->pb, cpe);
}
}
- for (j = 0; j < chans; j++) {
- s->cur_channel = start_ch + j;
- encode_individual_channel(avctx, s, &cpe->ch[j], cpe->common_window);
+ for (ch = 0; ch < chans; ch++) {
+ s->cur_channel = start_ch + ch;
+ encode_individual_channel(avctx, s, &cpe->ch[ch], cpe->common_window);
}
start_ch += chans;
}
frame_bits = put_bits_count(&s->pb);
- if (frame_bits <= 6144 * avctx->channels - 3)
+ if (frame_bits <= 6144 * avctx->channels - 3) {
+ s->psy.bitres.bits = frame_bits / avctx->channels;
break;
+ }
s->lambda *= avctx->bit_rate * 1024.0f / avctx->sample_rate / frame_bits;
return 0;
}
+#define AACENC_FLAGS AV_OPT_FLAG_ENCODING_PARAM | AV_OPT_FLAG_AUDIO_PARAM
+static const AVOption aacenc_options[] = {
+ {"stereo_mode", "Stereo coding method", offsetof(AACEncContext, options.stereo_mode), FF_OPT_TYPE_INT, {.dbl = 0}, -1, 1, AACENC_FLAGS, "stereo_mode"},
+ {"auto", "Selected by the Encoder", 0, FF_OPT_TYPE_CONST, {.dbl = -1 }, INT_MIN, INT_MAX, AACENC_FLAGS, "stereo_mode"},
+ {"ms_off", "Disable Mid/Side coding", 0, FF_OPT_TYPE_CONST, {.dbl = 0 }, INT_MIN, INT_MAX, AACENC_FLAGS, "stereo_mode"},
+ {"ms_force", "Force Mid/Side for the whole frame if possible", 0, FF_OPT_TYPE_CONST, {.dbl = 1 }, INT_MIN, INT_MAX, AACENC_FLAGS, "stereo_mode"},
+ {NULL}
+};
+
+static const AVClass aacenc_class = {
+ "AAC encoder",
+ av_default_item_name,
+ aacenc_options,
+ LIBAVUTIL_VERSION_INT,
+};
+
AVCodec ff_aac_encoder = {
- "aac",
- AVMEDIA_TYPE_AUDIO,
- CODEC_ID_AAC,
- sizeof(AACEncContext),
- aac_encode_init,
- aac_encode_frame,
- aac_encode_end,
+ .name = "aac",
+ .type = AVMEDIA_TYPE_AUDIO,
+ .id = CODEC_ID_AAC,
+ .priv_data_size = sizeof(AACEncContext),
+ .init = aac_encode_init,
+ .encode = aac_encode_frame,
+ .close = aac_encode_end,
.capabilities = CODEC_CAP_SMALL_LAST_FRAME | CODEC_CAP_DELAY | CODEC_CAP_EXPERIMENTAL,
.sample_fmts = (const enum AVSampleFormat[]){AV_SAMPLE_FMT_S16,AV_SAMPLE_FMT_NONE},
.long_name = NULL_IF_CONFIG_SMALL("Advanced Audio Coding"),
+ .priv_class = &aacenc_class,
};