static void swap_sample_fmts_on_filter(AVFilterContext *filter)
{
AVFilterLink *link = NULL;
int format, bps;
int i, j;
for (i = 0; i < filter->nb_inputs; i++) {
link = filter->inputs[i];
if (link->type == AVMEDIA_TYPE_AUDIO &&
link->out_formats->format_count == 1)
break;
}
if (i == filter->nb_inputs)
return;
format = link->out_formats->formats[0];
bps = av_get_bytes_per_sample(format);
for (i = 0; i < filter->nb_outputs; i++) {
AVFilterLink *outlink = filter->outputs[i];
int best_idx = -1, best_score = INT_MIN;
if (outlink->type != AVMEDIA_TYPE_AUDIO ||
outlink->in_formats->format_count < 2)
continue;
for (j = 0; j < outlink->in_formats->format_count; j++) {
int out_format = outlink->in_formats->formats[j];
int out_bps = av_get_bytes_per_sample(out_format);
int score;
if (av_get_packed_sample_fmt(out_format) == format ||
av_get_planar_sample_fmt(out_format) == format) {
best_idx = j;
break;
}
/* for s32 and float prefer double to prevent loss of information */
if (bps == 4 && out_bps == 8) {
best_idx = j;
break;
}
/* prefer closest higher or equal bps */
score = -abs(out_bps - bps);
if (out_bps >= bps)
score += INT_MAX/2;
if (score > best_score) {
best_score = score;
best_idx = j;
}
}
av_assert0(best_idx >= 0);
FFSWAP(int, outlink->in_formats->formats[0],
outlink->in_formats->formats[best_idx]);
}
}
bool AVFormatWriter::OpenAudio(void)
{
AVCodecContext *c;
AVCodec *codec;
c = m_audioStream->codec;
c->strict_std_compliance = FF_COMPLIANCE_EXPERIMENTAL;
codec = avcodec_find_encoder(c->codec_id);
if (!codec)
{
LOG(VB_RECORD, LOG_ERR,
LOC + "OpenAudio(): avcodec_find_encoder() failed");
return false;
}
// try to find suitable format we can use. avcodec_open2 will fail if we don't
// find one, so no need to worry otherwise. Can only handle S16 or FLOAT
// we give priority to S16 as libmp3lame requires aligned floats which we can't guarantee
if (!FindAudioFormat(c, codec, AV_SAMPLE_FMT_S16))
{
FindAudioFormat(c, codec, AV_SAMPLE_FMT_FLT);
}
if (avcodec_open2(c, codec, NULL) < 0)
{
LOG(VB_RECORD, LOG_ERR,
LOC + "OpenAudio(): avcodec_open() failed");
return false;
}
m_audioFrameSize = c->frame_size; // number of *samples* per channel in an AVFrame
m_audPicture = av_frame_alloc();
if (!m_audPicture)
{
LOG(VB_RECORD, LOG_ERR,
LOC + "OpenAudio(): alloc_frame() failed");
return false;
}
int samples_per_frame = m_audioFrameSize * m_audioChannels;
int bps = av_get_bytes_per_sample(c->sample_fmt);
if (av_get_packed_sample_fmt(c->sample_fmt) == AV_SAMPLE_FMT_FLT)
{
// allocate buffer to convert from S16 to float
if (!(m_audioInBuf = (unsigned char*)av_malloc(bps * samples_per_frame)))
return false;
}
if (av_sample_fmt_is_planar(c->sample_fmt))
{
// allocate buffer to convert interleaved to planar audio
if (!(m_audioInPBuf = (unsigned char*)av_malloc(bps * samples_per_frame)))
return false;
}
return true;
}
AudioConvert *ff_audio_convert_alloc(AVAudioResampleContext *avr,
enum AVSampleFormat out_fmt,
enum AVSampleFormat in_fmt,
int channels, int sample_rate,
int apply_map)
{
AudioConvert *ac;
int in_planar, out_planar;
ac = av_mallocz(sizeof(*ac));
if (!ac)
return NULL;
ac->avr = avr;
ac->out_fmt = out_fmt;
ac->in_fmt = in_fmt;
ac->channels = channels;
ac->apply_map = apply_map;
if (avr->dither_method != AV_RESAMPLE_DITHER_NONE &&
av_get_packed_sample_fmt(out_fmt) == AV_SAMPLE_FMT_S16 &&
av_get_bytes_per_sample(in_fmt) > 2) {
ac->dc = ff_dither_alloc(avr, out_fmt, in_fmt, channels, sample_rate,
apply_map);
if (!ac->dc) {
av_free(ac);
return NULL;
}
return ac;
}
in_planar = ff_sample_fmt_is_planar(in_fmt, channels);
out_planar = ff_sample_fmt_is_planar(out_fmt, channels);
if (in_planar == out_planar) {
ac->func_type = CONV_FUNC_TYPE_FLAT;
ac->planes = in_planar ? ac->channels : 1;
} else if (in_planar)
ac->func_type = CONV_FUNC_TYPE_INTERLEAVE;
else
ac->func_type = CONV_FUNC_TYPE_DEINTERLEAVE;
set_generic_function(ac);
if (ARCH_AARCH64)
ff_audio_convert_init_aarch64(ac);
if (ARCH_ARM)
ff_audio_convert_init_arm(ac);
if (ARCH_X86)
ff_audio_convert_init_x86(ac);
return ac;
}
bool AVFormatWriter::FindAudioFormat(AVCodecContext *ctx, AVCodec *c, AVSampleFormat format)
{
if (c->sample_fmts)
{
for (int i = 0; c->sample_fmts[i] != AV_SAMPLE_FMT_NONE; i++)
{
if (av_get_packed_sample_fmt(c->sample_fmts[i]) == format)
{
ctx->sample_fmt = c->sample_fmts[i];
return true;
}
}
}
return false;
}
void FillAP(FFMS_AudioProperties &AP, AVCodecContext *CTX, FFMS_Track &Frames) {
AP.SampleFormat = static_cast<FFMS_SampleFormat>(av_get_packed_sample_fmt(CTX->sample_fmt));
AP.BitsPerSample = av_get_bytes_per_sample(CTX->sample_fmt) * 8;
AP.Channels = CTX->channels;
AP.ChannelLayout = CTX->channel_layout;
AP.SampleRate = CTX->sample_rate;
if (!Frames.empty()) {
AP.NumSamples = (Frames.back()).SampleStart + (Frames.back()).SampleCount;
AP.FirstTime = ((Frames.front().PTS * Frames.TB.Num) / (double)Frames.TB.Den) / 1000;
AP.LastTime = ((Frames.back().PTS * Frames.TB.Num) / (double)Frames.TB.Den) / 1000;
}
if (AP.ChannelLayout == 0)
AP.ChannelLayout = av_get_default_channel_layout(AP.Channels);
}
AudioConvert *swri_audio_convert_alloc(enum AVSampleFormat out_fmt,
enum AVSampleFormat in_fmt,
int channels, const int *ch_map,
int flags)
{
AudioConvert *ctx;
conv_func_type *f = fmt_pair_to_conv_functions[av_get_packed_sample_fmt(out_fmt) + AV_SAMPLE_FMT_NB*av_get_packed_sample_fmt(in_fmt)];
if (!f)
return NULL;
ctx = av_mallocz(sizeof(*ctx));
if (!ctx)
return NULL;
if(channels == 1){
in_fmt = av_get_planar_sample_fmt( in_fmt);
out_fmt = av_get_planar_sample_fmt(out_fmt);
}
ctx->channels = channels;
ctx->conv_f = f;
ctx->ch_map = ch_map;
if (in_fmt == AV_SAMPLE_FMT_U8 || in_fmt == AV_SAMPLE_FMT_U8P)
memset(ctx->silence, 0x80, sizeof(ctx->silence));
if(out_fmt == in_fmt && !ch_map) {
switch(av_get_bytes_per_sample(in_fmt)){
case 1:ctx->simd_f = cpy1; break;
case 2:ctx->simd_f = cpy2; break;
case 4:ctx->simd_f = cpy4; break;
case 8:ctx->simd_f = cpy8; break;
}
}
#if (HAVE_YASM == 1) && (HAVE_MMX == 1)
if(HAVE_YASM && HAVE_MMX) swri_audio_convert_init_x86(ctx, out_fmt, in_fmt, channels);
#endif
#if (ARCH_ARM == 1)
if(ARCH_ARM) swri_audio_convert_init_arm(ctx, out_fmt, in_fmt, channels);
#endif
#if (ARCH_AARCH64 == 1)
if(ARCH_AARCH64) swri_audio_convert_init_aarch64(ctx, out_fmt, in_fmt, channels);
#endif
return ctx;
}
int BKE_ffmpeg_start(void *context_v, struct Scene *scene, RenderData *rd, int rectx, int recty,
ReportList *reports, bool preview, const char *suffix)
{
int success;
FFMpegContext *context = context_v;
context->ffmpeg_autosplit_count = 0;
context->ffmpeg_preview = preview;
success = start_ffmpeg_impl(context, rd, rectx, recty, suffix, reports);
#ifdef WITH_AUDASPACE
if (context->audio_stream) {
AVCodecContext *c = context->audio_stream->codec;
AUD_DeviceSpecs specs;
specs.channels = c->channels;
switch (av_get_packed_sample_fmt(c->sample_fmt)) {
case AV_SAMPLE_FMT_U8:
specs.format = AUD_FORMAT_U8;
break;
case AV_SAMPLE_FMT_S16:
specs.format = AUD_FORMAT_S16;
break;
case AV_SAMPLE_FMT_S32:
specs.format = AUD_FORMAT_S32;
break;
case AV_SAMPLE_FMT_FLT:
specs.format = AUD_FORMAT_FLOAT32;
break;
case AV_SAMPLE_FMT_DBL:
specs.format = AUD_FORMAT_FLOAT64;
break;
default:
return -31415;
}
specs.rate = rd->ffcodecdata.audio_mixrate;
context->audio_mixdown_device = BKE_sound_mixdown(scene, specs, preview ? rd->psfra : rd->sfra, rd->ffcodecdata.audio_volume);
#ifdef FFMPEG_CODEC_TIME_BASE
c->time_base.den = specs.rate;
c->time_base.num = 1;
#endif
}
#endif
return success;
}
/**
* Allocates and initializes the SwrContext so that we don't have to deal with planar sample formats.
*
* @param env JNIEnv
* @param aio FFAudioIO
* @return a negative value should an error occur
*/
static int init_swr(JNIEnv *env, FFAudioIO *aio) {
int res = 0;
aio->swr_context = swr_alloc();
if (!aio->swr_context) {
res = AVERROR(ENOMEM);
throwIOExceptionIfError(env, res, "Could not allocate swr context.");
goto bail;
}
av_opt_set_sample_fmt(aio->swr_context, "in_sample_fmt", aio->stream->codecpar->format, 0);
// make sure we get interleaved/packed output
av_opt_set_sample_fmt(aio->swr_context, "out_sample_fmt", av_get_packed_sample_fmt(aio->stream->codecpar->format), 0);
// keep everything else the way it was...
av_opt_set_int(aio->swr_context, "in_channel_count", aio->stream->codecpar->channels, 0);
av_opt_set_int(aio->swr_context, "out_channel_count", aio->stream->codecpar->channels, 0);
av_opt_set_int(aio->swr_context, "in_channel_layout", aio->stream->codecpar->channel_layout, 0);
av_opt_set_int(aio->swr_context, "out_channel_layout", aio->stream->codecpar->channel_layout, 0);
av_opt_set_int(aio->swr_context, "in_sample_rate", aio->stream->codecpar->sample_rate, 0);
av_opt_set_int(aio->swr_context, "out_sample_rate", aio->stream->codecpar->sample_rate, 0);
res = swr_init(aio->swr_context);
if (res < 0) {
res = AVERROR(ENOMEM);
throwIOExceptionIfError(env, res, "Could not initialize swr context");
goto bail;
}
//fprintf(stderr, "init_swr: dither context: %d\n", aio->swr_context->dither);
//fprintf(stderr, "init_swr: output sample bits: %d\n", aio->swr_context->dither.output_sample_bits);
bail:
return res;
}
int AVFormatWriter::WriteAudioFrame(unsigned char *buf, int fnum, long long &timecode)
{
#if HAVE_BIGENDIAN
bswap_16_buf((short int*) buf, m_audioFrameSize, m_audioChannels);
#endif
int got_packet = 0;
int ret = 0;
int samples_per_avframe = m_audioFrameSize * m_audioChannels;
int sampleSizeIn = AudioOutputSettings::SampleSize(FORMAT_S16);
AudioFormat format =
AudioOutputSettings::AVSampleFormatToFormat(m_audioStream->codec->sample_fmt);
int sampleSizeOut = AudioOutputSettings::SampleSize(format);
AVPacket pkt;
av_init_packet(&pkt);
pkt.data = NULL;
pkt.size = 0;
if (av_get_packed_sample_fmt(m_audioStream->codec->sample_fmt) == AV_SAMPLE_FMT_FLT)
{
AudioOutputUtil::toFloat(FORMAT_S16, (void *)m_audioInBuf, (void *)buf,
samples_per_avframe * sampleSizeIn);
buf = m_audioInBuf;
}
if (av_sample_fmt_is_planar(m_audioStream->codec->sample_fmt))
{
AudioOutputUtil::DeinterleaveSamples(format,
m_audioChannels,
m_audioInPBuf,
buf,
samples_per_avframe * sampleSizeOut);
// init AVFrame for planar data (input is interleaved)
for (int j = 0, jj = 0; j < m_audioChannels; j++, jj += m_audioFrameSize)
{
m_audPicture->data[j] = (uint8_t*)(m_audioInPBuf + jj * sampleSizeOut);
}
}
else
{
m_audPicture->data[0] = buf;
}
m_audPicture->linesize[0] = m_audioFrameSize;
m_audPicture->nb_samples = m_audioFrameSize;
m_audPicture->format = m_audioStream->codec->sample_fmt;
m_audPicture->extended_data = m_audPicture->data;
m_bufferedAudioFrameTimes.push_back(timecode);
{
QMutexLocker locker(avcodeclock);
ret = avcodec_encode_audio2(m_audioStream->codec, &pkt,
m_audPicture, &got_packet);
}
if (ret < 0)
{
LOG(VB_RECORD, LOG_ERR, "avcodec_encode_audio2() failed");
return ret;
}
if (!got_packet)
{
//LOG(VB_RECORD, LOG_ERR, QString("WriteAudioFrame(): Frame Buffered: cs: %1, mfw: %2, f->tc: %3, fn: %4").arg(m_audPkt->size).arg(m_framesWritten).arg(timecode).arg(fnum));
return ret;
}
long long tc = timecode;
if (m_bufferedAudioFrameTimes.size())
tc = m_bufferedAudioFrameTimes.takeFirst();
if (m_startingTimecodeOffset == -1)
m_startingTimecodeOffset = tc - 1;
tc -= m_startingTimecodeOffset;
if (m_avVideoCodec)
pkt.pts = tc * m_videoStream->time_base.den / m_videoStream->time_base.num / 1000;
else
pkt.pts = tc * m_audioStream->time_base.den / m_audioStream->time_base.num / 1000;
pkt.dts = AV_NOPTS_VALUE;
pkt.flags |= AV_PKT_FLAG_KEY;
pkt.stream_index = m_audioStream->index;
//LOG(VB_RECORD, LOG_ERR, QString("WriteAudioFrame(): cs: %1, mfw: %2, pkt->pts: %3, tc: %4, fn: %5, f->tc: %6").arg(m_audPkt->size).arg(m_framesWritten).arg(m_audPkt->pts).arg(tc).arg(fnum).arg(timecode));
ret = av_interleaved_write_frame(m_ctx, &pkt);
if (ret != 0)
LOG(VB_RECORD, LOG_ERR, LOC + "WriteAudioFrame(): "
"av_interleaved_write_frame couldn't write Audio");
timecode = tc + m_startingTimecodeOffset;
av_packet_unref(&pkt);
return 1;
}
av_cold static int auto_matrix(SwrContext *s)
{
int i, j, out_i;
double matrix[NUM_NAMED_CHANNELS][NUM_NAMED_CHANNELS]={{0}};
int64_t unaccounted, in_ch_layout, out_ch_layout;
double maxcoef=0;
char buf[128];
const int matrix_encoding = s->matrix_encoding;
float maxval;
in_ch_layout = clean_layout(s, s->in_ch_layout);
out_ch_layout = clean_layout(s, s->out_ch_layout);
if( out_ch_layout == AV_CH_LAYOUT_STEREO_DOWNMIX
&& (in_ch_layout & AV_CH_LAYOUT_STEREO_DOWNMIX) == 0
)
out_ch_layout = AV_CH_LAYOUT_STEREO;
if( in_ch_layout == AV_CH_LAYOUT_STEREO_DOWNMIX
&& (out_ch_layout & AV_CH_LAYOUT_STEREO_DOWNMIX) == 0
)
in_ch_layout = AV_CH_LAYOUT_STEREO;
if(!sane_layout(in_ch_layout)){
av_get_channel_layout_string(buf, sizeof(buf), -1, s->in_ch_layout);
av_log(s, AV_LOG_ERROR, "Input channel layout '%s' is not supported\n", buf);
return AVERROR(EINVAL);
}
if(!sane_layout(out_ch_layout)){
av_get_channel_layout_string(buf, sizeof(buf), -1, s->out_ch_layout);
av_log(s, AV_LOG_ERROR, "Output channel layout '%s' is not supported\n", buf);
return AVERROR(EINVAL);
}
memset(s->matrix, 0, sizeof(s->matrix));
for(i=0; i<FF_ARRAY_ELEMS(matrix); i++){
if(in_ch_layout & out_ch_layout & (1ULL<<i))
matrix[i][i]= 1.0;
}
unaccounted= in_ch_layout & ~out_ch_layout;
//FIXME implement dolby surround
//FIXME implement full ac3
if(unaccounted & AV_CH_FRONT_CENTER){
if((out_ch_layout & AV_CH_LAYOUT_STEREO) == AV_CH_LAYOUT_STEREO){
if(in_ch_layout & AV_CH_LAYOUT_STEREO) {
matrix[ FRONT_LEFT][FRONT_CENTER]+= s->clev;
matrix[FRONT_RIGHT][FRONT_CENTER]+= s->clev;
} else {
matrix[ FRONT_LEFT][FRONT_CENTER]+= M_SQRT1_2;
matrix[FRONT_RIGHT][FRONT_CENTER]+= M_SQRT1_2;
}
}else
av_assert0(0);
}
if(unaccounted & AV_CH_LAYOUT_STEREO){
if(out_ch_layout & AV_CH_FRONT_CENTER){
matrix[FRONT_CENTER][ FRONT_LEFT]+= M_SQRT1_2;
matrix[FRONT_CENTER][FRONT_RIGHT]+= M_SQRT1_2;
if(in_ch_layout & AV_CH_FRONT_CENTER)
matrix[FRONT_CENTER][ FRONT_CENTER] = s->clev*sqrt(2);
}else
av_assert0(0);
}
if(unaccounted & AV_CH_BACK_CENTER){
if(out_ch_layout & AV_CH_BACK_LEFT){
matrix[ BACK_LEFT][BACK_CENTER]+= M_SQRT1_2;
matrix[BACK_RIGHT][BACK_CENTER]+= M_SQRT1_2;
}else if(out_ch_layout & AV_CH_SIDE_LEFT){
matrix[ SIDE_LEFT][BACK_CENTER]+= M_SQRT1_2;
matrix[SIDE_RIGHT][BACK_CENTER]+= M_SQRT1_2;
}else if(out_ch_layout & AV_CH_FRONT_LEFT){
if (matrix_encoding == AV_MATRIX_ENCODING_DOLBY ||
matrix_encoding == AV_MATRIX_ENCODING_DPLII) {
if (unaccounted & (AV_CH_BACK_LEFT | AV_CH_SIDE_LEFT)) {
matrix[FRONT_LEFT ][BACK_CENTER] -= s->slev * M_SQRT1_2;
matrix[FRONT_RIGHT][BACK_CENTER] += s->slev * M_SQRT1_2;
} else {
matrix[FRONT_LEFT ][BACK_CENTER] -= s->slev;
matrix[FRONT_RIGHT][BACK_CENTER] += s->slev;
}
} else {
matrix[ FRONT_LEFT][BACK_CENTER]+= s->slev*M_SQRT1_2;
matrix[FRONT_RIGHT][BACK_CENTER]+= s->slev*M_SQRT1_2;
}
}else if(out_ch_layout & AV_CH_FRONT_CENTER){
matrix[ FRONT_CENTER][BACK_CENTER]+= s->slev*M_SQRT1_2;
}else
av_assert0(0);
}
if(unaccounted & AV_CH_BACK_LEFT){
if(out_ch_layout & AV_CH_BACK_CENTER){
matrix[BACK_CENTER][ BACK_LEFT]+= M_SQRT1_2;
matrix[BACK_CENTER][BACK_RIGHT]+= M_SQRT1_2;
}else if(out_ch_layout & AV_CH_SIDE_LEFT){
if(in_ch_layout & AV_CH_SIDE_LEFT){
matrix[ SIDE_LEFT][ BACK_LEFT]+= M_SQRT1_2;
matrix[SIDE_RIGHT][BACK_RIGHT]+= M_SQRT1_2;
}else{
matrix[ SIDE_LEFT][ BACK_LEFT]+= 1.0;
matrix[SIDE_RIGHT][BACK_RIGHT]+= 1.0;
}
}else if(out_ch_layout & AV_CH_FRONT_LEFT){
if (matrix_encoding == AV_MATRIX_ENCODING_DOLBY) {
matrix[FRONT_LEFT ][BACK_LEFT ] -= s->slev * M_SQRT1_2;
matrix[FRONT_LEFT ][BACK_RIGHT] -= s->slev * M_SQRT1_2;
matrix[FRONT_RIGHT][BACK_LEFT ] += s->slev * M_SQRT1_2;
matrix[FRONT_RIGHT][BACK_RIGHT] += s->slev * M_SQRT1_2;
} else if (matrix_encoding == AV_MATRIX_ENCODING_DPLII) {
matrix[FRONT_LEFT ][BACK_LEFT ] -= s->slev * SQRT3_2;
matrix[FRONT_LEFT ][BACK_RIGHT] -= s->slev * M_SQRT1_2;
matrix[FRONT_RIGHT][BACK_LEFT ] += s->slev * M_SQRT1_2;
matrix[FRONT_RIGHT][BACK_RIGHT] += s->slev * SQRT3_2;
} else {
matrix[ FRONT_LEFT][ BACK_LEFT] += s->slev;
matrix[FRONT_RIGHT][BACK_RIGHT] += s->slev;
}
}else if(out_ch_layout & AV_CH_FRONT_CENTER){
matrix[ FRONT_CENTER][BACK_LEFT ]+= s->slev*M_SQRT1_2;
matrix[ FRONT_CENTER][BACK_RIGHT]+= s->slev*M_SQRT1_2;
}else
av_assert0(0);
}
if(unaccounted & AV_CH_SIDE_LEFT){
if(out_ch_layout & AV_CH_BACK_LEFT){
/* if back channels do not exist in the input, just copy side
channels to back channels, otherwise mix side into back */
if (in_ch_layout & AV_CH_BACK_LEFT) {
matrix[BACK_LEFT ][SIDE_LEFT ] += M_SQRT1_2;
matrix[BACK_RIGHT][SIDE_RIGHT] += M_SQRT1_2;
} else {
matrix[BACK_LEFT ][SIDE_LEFT ] += 1.0;
matrix[BACK_RIGHT][SIDE_RIGHT] += 1.0;
}
}else if(out_ch_layout & AV_CH_BACK_CENTER){
matrix[BACK_CENTER][ SIDE_LEFT]+= M_SQRT1_2;
matrix[BACK_CENTER][SIDE_RIGHT]+= M_SQRT1_2;
}else if(out_ch_layout & AV_CH_FRONT_LEFT){
if (matrix_encoding == AV_MATRIX_ENCODING_DOLBY) {
matrix[FRONT_LEFT ][SIDE_LEFT ] -= s->slev * M_SQRT1_2;
matrix[FRONT_LEFT ][SIDE_RIGHT] -= s->slev * M_SQRT1_2;
matrix[FRONT_RIGHT][SIDE_LEFT ] += s->slev * M_SQRT1_2;
matrix[FRONT_RIGHT][SIDE_RIGHT] += s->slev * M_SQRT1_2;
} else if (matrix_encoding == AV_MATRIX_ENCODING_DPLII) {
matrix[FRONT_LEFT ][SIDE_LEFT ] -= s->slev * SQRT3_2;
matrix[FRONT_LEFT ][SIDE_RIGHT] -= s->slev * M_SQRT1_2;
matrix[FRONT_RIGHT][SIDE_LEFT ] += s->slev * M_SQRT1_2;
matrix[FRONT_RIGHT][SIDE_RIGHT] += s->slev * SQRT3_2;
} else {
matrix[ FRONT_LEFT][ SIDE_LEFT] += s->slev;
matrix[FRONT_RIGHT][SIDE_RIGHT] += s->slev;
}
}else if(out_ch_layout & AV_CH_FRONT_CENTER){
matrix[ FRONT_CENTER][SIDE_LEFT ]+= s->slev*M_SQRT1_2;
matrix[ FRONT_CENTER][SIDE_RIGHT]+= s->slev*M_SQRT1_2;
}else
av_assert0(0);
}
if(unaccounted & AV_CH_FRONT_LEFT_OF_CENTER){
if(out_ch_layout & AV_CH_FRONT_LEFT){
matrix[ FRONT_LEFT][ FRONT_LEFT_OF_CENTER]+= 1.0;
matrix[FRONT_RIGHT][FRONT_RIGHT_OF_CENTER]+= 1.0;
}else if(out_ch_layout & AV_CH_FRONT_CENTER){
matrix[ FRONT_CENTER][ FRONT_LEFT_OF_CENTER]+= M_SQRT1_2;
matrix[ FRONT_CENTER][FRONT_RIGHT_OF_CENTER]+= M_SQRT1_2;
}else
av_assert0(0);
}
/* mix LFE into front left/right or center */
if (unaccounted & AV_CH_LOW_FREQUENCY) {
if (out_ch_layout & AV_CH_FRONT_CENTER) {
matrix[FRONT_CENTER][LOW_FREQUENCY] += s->lfe_mix_level;
} else if (out_ch_layout & AV_CH_FRONT_LEFT) {
matrix[FRONT_LEFT ][LOW_FREQUENCY] += s->lfe_mix_level * M_SQRT1_2;
matrix[FRONT_RIGHT][LOW_FREQUENCY] += s->lfe_mix_level * M_SQRT1_2;
} else
av_assert0(0);
}
for(out_i=i=0; i<64; i++){
double sum=0;
int in_i=0;
if((out_ch_layout & (1ULL<<i)) == 0)
continue;
for(j=0; j<64; j++){
if((in_ch_layout & (1ULL<<j)) == 0)
continue;
if (i < FF_ARRAY_ELEMS(matrix) && j < FF_ARRAY_ELEMS(matrix[0]))
s->matrix[out_i][in_i]= matrix[i][j];
else
s->matrix[out_i][in_i]= i == j && (in_ch_layout & out_ch_layout & (1ULL<<i));
sum += fabs(s->matrix[out_i][in_i]);
in_i++;
}
maxcoef= FFMAX(maxcoef, sum);
out_i++;
}
if(s->rematrix_volume < 0)
maxcoef = -s->rematrix_volume;
if (s->rematrix_maxval > 0) {
maxval = s->rematrix_maxval;
} else if ( av_get_packed_sample_fmt(s->out_sample_fmt) < AV_SAMPLE_FMT_FLT
|| av_get_packed_sample_fmt(s->int_sample_fmt) < AV_SAMPLE_FMT_FLT) {
maxval = 1.0;
} else
maxval = INT_MAX;
if(maxcoef > maxval || s->rematrix_volume < 0){
maxcoef /= maxval;
for(i=0; i<SWR_CH_MAX; i++)
for(j=0; j<SWR_CH_MAX; j++){
s->matrix[i][j] /= maxcoef;
}
}
if(s->rematrix_volume > 0){
for(i=0; i<SWR_CH_MAX; i++)
for(j=0; j<SWR_CH_MAX; j++){
s->matrix[i][j] *= s->rematrix_volume;
}
}
av_log(s, AV_LOG_DEBUG, "Matrix coefficients:\n");
for(i=0; i<av_get_channel_layout_nb_channels(out_ch_layout); i++){
const char *c =
av_get_channel_name(av_channel_layout_extract_channel(out_ch_layout, i));
av_log(s, AV_LOG_DEBUG, "%s: ", c ? c : "?");
for(j=0; j<av_get_channel_layout_nb_channels(in_ch_layout); j++){
c = av_get_channel_name(av_channel_layout_extract_channel(in_ch_layout, j));
av_log(s, AV_LOG_DEBUG, "%s:%f ", c ? c : "?", s->matrix[i][j]);
}
av_log(s, AV_LOG_DEBUG, "\n");
}
return 0;
}
int main (int argc, char **argv)
{
int ret = 0, got_frame;
if (argc != 4 && argc != 5) {
fprintf(stderr, "usage: %s [-refcount] input_file video_output_file audio_output_file\n"
"API example program to show how to read frames from an input file.\n"
"This program reads frames from a file, decodes them, and writes decoded\n"
"video frames to a rawvideo file named video_output_file, and decoded\n"
"audio frames to a rawaudio file named audio_output_file.\n\n"
"If the -refcount option is specified, the program use the\n"
"reference counting frame system which allows keeping a copy of\n"
"the data for longer than one decode call.\n"
"\n", argv[0]);
exit(1);
}
if (argc == 5 && !strcmp(argv[1], "-refcount")) {
refcount = 1;
argv++;
}
src_filename = argv[1];
video_dst_filename = argv[2];
audio_dst_filename = argv[3];
/* register all formats and codecs */
av_register_all();
/* open input file, and allocate format context */
if (avformat_open_input(&fmt_ctx, src_filename, NULL, NULL) < 0) {
fprintf(stderr, "Could not open source file %s\n", src_filename);
exit(1);
}
/* retrieve stream information */
if (avformat_find_stream_info(fmt_ctx, NULL) < 0) {
fprintf(stderr, "Could not find stream information\n");
exit(1);
}
if (open_codec_context(&video_stream_idx, fmt_ctx, AVMEDIA_TYPE_VIDEO) >= 0) {
video_stream = fmt_ctx->streams[video_stream_idx];
video_dec_ctx = video_stream->codec;
video_dst_file = fopen(video_dst_filename, "wb");
if (!video_dst_file) {
fprintf(stderr, "Could not open destination file %s\n", video_dst_filename);
ret = 1;
goto end;
}
/* allocate image where the decoded image will be put */
width = video_dec_ctx->width;
height = video_dec_ctx->height;
pix_fmt = video_dec_ctx->pix_fmt;
ret = av_image_alloc(video_dst_data, video_dst_linesize,
width, height, pix_fmt, 1);
if (ret < 0) {
fprintf(stderr, "Could not allocate raw video buffer\n");
goto end;
}
video_dst_bufsize = ret;
}
if (open_codec_context(&audio_stream_idx, fmt_ctx, AVMEDIA_TYPE_AUDIO) >= 0) {
audio_stream = fmt_ctx->streams[audio_stream_idx];
audio_dec_ctx = audio_stream->codec;
audio_dst_file = fopen(audio_dst_filename, "wb");
if (!audio_dst_file) {
fprintf(stderr, "Could not open destination file %s\n", audio_dst_filename);
ret = 1;
goto end;
}
}
/* dump input information to stderr */
av_dump_format(fmt_ctx, 0, src_filename, 0);
if (!audio_stream && !video_stream) {
fprintf(stderr, "Could not find audio or video stream in the input, aborting\n");
ret = 1;
goto end;
}
frame = av_frame_alloc();
if (!frame) {
fprintf(stderr, "Could not allocate frame\n");
ret = AVERROR(ENOMEM);
goto end;
}
/* initialize packet, set data to NULL, let the demuxer fill it */
av_init_packet(&pkt);
pkt.data = NULL;
pkt.size = 0;
if (video_stream)
printf("Demuxing video from file '%s' into '%s'\n", src_filename, video_dst_filename);
if (audio_stream)
printf("Demuxing audio from file '%s' into '%s'\n", src_filename, audio_dst_filename);
/* read frames from the file */
while (av_read_frame(fmt_ctx, &pkt) >= 0) {
AVPacket orig_pkt = pkt;
do {
ret = decode_packet(&got_frame, 0);
if (ret < 0)
break;
pkt.data += ret;
pkt.size -= ret;
} while (pkt.size > 0);
av_packet_unref(&orig_pkt);
}
/* flush cached frames */
pkt.data = NULL;
pkt.size = 0;
do {
decode_packet(&got_frame, 1);
} while (got_frame);
printf("Demuxing succeeded.\n");
if (video_stream) {
printf("Play the output video file with the command:\n"
"ffplay -f rawvideo -pix_fmt %s -video_size %dx%d %s\n",
av_get_pix_fmt_name(pix_fmt), width, height,
video_dst_filename);
}
if (audio_stream) {
enum AVSampleFormat sfmt = audio_dec_ctx->sample_fmt;
int n_channels = audio_dec_ctx->channels;
const char *fmt;
if (av_sample_fmt_is_planar(sfmt)) {
const char *packed = av_get_sample_fmt_name(sfmt);
printf("Warning: the sample format the decoder produced is planar "
"(%s). This example will output the first channel only.\n",
packed ? packed : "?");
sfmt = av_get_packed_sample_fmt(sfmt);
n_channels = 1;
}
if ((ret = get_format_from_sample_fmt(&fmt, sfmt)) < 0)
goto end;
printf("Play the output audio file with the command:\n"
"ffplay -f %s -ac %d -ar %d %s\n",
fmt, n_channels, audio_dec_ctx->sample_rate,
audio_dst_filename);
}
end:
avcodec_close(video_dec_ctx);
avcodec_close(audio_dec_ctx);
avformat_close_input(&fmt_ctx);
if (video_dst_file)
fclose(video_dst_file);
if (audio_dst_file)
fclose(audio_dst_file);
av_frame_free(&frame);
av_free(video_dst_data[0]);
return ret < 0;
}
DitherContext *ff_dither_alloc(AVAudioResampleContext *avr,
enum AVSampleFormat out_fmt,
enum AVSampleFormat in_fmt,
int channels, int sample_rate)
{
AVLFG seed_gen;
DitherContext *c;
int ch;
if (av_get_packed_sample_fmt(out_fmt) != AV_SAMPLE_FMT_S16 ||
av_get_bytes_per_sample(in_fmt) <= 2) {
av_log(avr, AV_LOG_ERROR, "dithering %s to %s is not supported\n",
av_get_sample_fmt_name(in_fmt), av_get_sample_fmt_name(out_fmt));
return NULL;
}
c = av_mallocz(sizeof(*c));
if (!c)
return NULL;
if (avr->dither_method == AV_RESAMPLE_DITHER_TRIANGULAR_NS &&
sample_rate != 48000 && sample_rate != 44100) {
av_log(avr, AV_LOG_WARNING, "sample rate must be 48000 or 44100 Hz "
"for triangular_ns dither. using triangular_hp instead.\n");
avr->dither_method = AV_RESAMPLE_DITHER_TRIANGULAR_HP;
}
c->method = avr->dither_method;
dither_init(&c->ddsp, c->method);
if (c->method == AV_RESAMPLE_DITHER_TRIANGULAR_NS) {
if (sample_rate == 48000) {
c->ns_coef_b = ns_48_coef_b;
c->ns_coef_a = ns_48_coef_a;
} else {
c->ns_coef_b = ns_44_coef_b;
c->ns_coef_a = ns_44_coef_a;
}
}
/* Either s16 or s16p output format is allowed, but s16p is used
internally, so we need to use a temp buffer and interleave if the output
format is s16 */
if (out_fmt != AV_SAMPLE_FMT_S16P) {
c->s16_data = ff_audio_data_alloc(channels, 1024, AV_SAMPLE_FMT_S16P,
"dither s16 buffer");
if (!c->s16_data)
goto fail;
c->ac_out = ff_audio_convert_alloc(avr, out_fmt, AV_SAMPLE_FMT_S16P,
channels, sample_rate);
if (!c->ac_out)
goto fail;
}
if (in_fmt != AV_SAMPLE_FMT_FLTP) {
c->flt_data = ff_audio_data_alloc(channels, 1024, AV_SAMPLE_FMT_FLTP,
"dither flt buffer");
if (!c->flt_data)
goto fail;
c->ac_in = ff_audio_convert_alloc(avr, AV_SAMPLE_FMT_FLTP, in_fmt,
channels, sample_rate);
if (!c->ac_in)
goto fail;
}
c->state = av_mallocz(channels * sizeof(*c->state));
if (!c->state)
goto fail;
c->channels = channels;
/* calculate thresholds for turning off dithering during periods of
silence to avoid replacing digital silence with quiet dither noise */
c->mute_dither_threshold = lrintf(sample_rate * MUTE_THRESHOLD_SEC);
c->mute_reset_threshold = c->mute_dither_threshold * 4;
/* initialize dither states */
av_lfg_init(&seed_gen, 0xC0FFEE);
for (ch = 0; ch < channels; ch++) {
DitherState *state = &c->state[ch];
state->mute = c->mute_reset_threshold + 1;
state->seed = av_lfg_get(&seed_gen);
generate_dither_noise(c, state, FFMAX(32768, sample_rate / 2));
}
return c;
fail:
ff_dither_free(&c);
return NULL;
}
/* Prefer playing audio with the samplerate given in container data
* if available, but take number the number of channels and sample format
* from the codec, since if the codec isn't using the correct values for
* those everything breaks anyway.
*/
static int setup_format(sh_audio_t *sh_audio)
{
struct priv *priv = sh_audio->context;
AVCodecContext *codec = priv->avctx;
int sample_format = sample_fmt_lavc2native(codec->sample_fmt);
if (sample_format == AF_FORMAT_UNKNOWN) {
#ifndef CONFIG_LIBAVRESAMPLE
if (av_sample_fmt_is_planar(codec->sample_fmt))
mp_msg(MSGT_DECAUDIO, MSGL_ERR,
"The player has been compiled without libavresample "
"support,\nwhich is needed with this libavcodec decoder "
"version.\nCompile with libavresample enabled to make "
"audio decoding work!\n");
else
mp_msg(MSGT_DECAUDIO, MSGL_ERR, "Unsupported sample format\n");
goto error;
#else
if (priv->avr && (priv->resample_fmt != codec->sample_fmt ||
priv->resample_channels != codec->channels))
avresample_free(&priv->avr);
if (!priv->avr) {
int ret;
uint8_t error[128];
enum AVSampleFormat out_fmt =
av_get_packed_sample_fmt(codec->sample_fmt);
uint64_t ch_layout = codec->channel_layout;
mp_msg(MSGT_DECAUDIO, MSGL_V,
"(Re)initializing libavresample format conversion...\n");
if (!ch_layout)
ch_layout = av_get_default_channel_layout(codec->channels);
/* if lavc format is planar, try just getting packed equivalent */
sample_format = sample_fmt_lavc2native(out_fmt);
if (sample_format == AF_FORMAT_UNKNOWN) {
/* fallback to s16 */
out_fmt = AV_SAMPLE_FMT_S16;
sample_format = AF_FORMAT_S16_NE;
}
priv->avr = avresample_alloc_context();
if (!priv->avr) {
mp_msg(MSGT_DECAUDIO, MSGL_FATAL, "Out of memory.\n");
abort();
}
av_opt_set_int(priv->avr, "in_channel_layout", ch_layout, 0);
av_opt_set_int(priv->avr, "out_channel_layout", ch_layout, 0);
av_opt_set_int(priv->avr, "in_sample_rate", codec->sample_rate, 0);
av_opt_set_int(priv->avr, "out_sample_rate", codec->sample_rate, 0);
av_opt_set_int(priv->avr, "in_sample_fmt", codec->sample_fmt, 0);
av_opt_set_int(priv->avr, "out_sample_fmt", out_fmt, 0);
if ((ret = avresample_open(priv->avr)) < 0) {
av_strerror(ret, error, sizeof(error));
mp_msg(MSGT_DECAUDIO, MSGL_ERR,
"Error opening libavresample: %s.\n", error);
goto error;
}
priv->resample_fmt = codec->sample_fmt;
priv->resample_channels = codec->channels;
priv->out_fmt = out_fmt;
priv->unitsize = av_get_bytes_per_sample(out_fmt) *
codec->channels;
} else
sample_format = sh_audio->sample_format;
} else if (priv->avr) {
avresample_free(&priv->avr);
#endif
}
bool broken_srate = false;
int samplerate = codec->sample_rate;
int container_samplerate = sh_audio->container_out_samplerate;
if (!container_samplerate && sh_audio->wf)
container_samplerate = sh_audio->wf->nSamplesPerSec;
if (codec->codec_id == AV_CODEC_ID_AAC
&& samplerate == 2 * container_samplerate)
broken_srate = true;
else if (container_samplerate)
samplerate = container_samplerate;
if (codec->channels != sh_audio->channels ||
samplerate != sh_audio->samplerate ||
sample_format != sh_audio->sample_format) {
sh_audio->channels = codec->channels;
sh_audio->samplerate = samplerate;
sh_audio->sample_format = sample_format;
sh_audio->samplesize = af_fmt2bits(sh_audio->sample_format) / 8;
if (broken_srate)
mp_msg(MSGT_DECAUDIO, MSGL_WARN,
"Ignoring broken container sample rate for AAC with SBR\n");
return 1;
}
return 0;
error:
#ifdef CONFIG_LIBAVRESAMPLE
avresample_free(&priv->avr);
#endif
return -1;
}
static int auto_matrix(SwrContext *s)
{
int i, j, out_i;
double matrix[64][64]={{0}};
int64_t unaccounted= s->in_ch_layout & ~s->out_ch_layout;
double maxcoef=0;
memset(s->matrix, 0, sizeof(s->matrix));
for(i=0; i<64; i++){
if(s->in_ch_layout & s->out_ch_layout & (1LL<<i))
matrix[i][i]= 1.0;
}
if(!sane_layout(s->in_ch_layout)){
av_log(s, AV_LOG_ERROR, "Input channel layout isnt supported\n");
return AVERROR(EINVAL);
}
if(!sane_layout(s->out_ch_layout)){
av_log(s, AV_LOG_ERROR, "Output channel layout isnt supported\n");
return AVERROR(EINVAL);
}
//FIXME implement dolby surround
//FIXME implement full ac3
if(unaccounted & AV_CH_FRONT_CENTER){
if((s->out_ch_layout & AV_CH_LAYOUT_STEREO) == AV_CH_LAYOUT_STEREO){
matrix[ FRONT_LEFT][FRONT_CENTER]+= M_SQRT1_2;
matrix[FRONT_RIGHT][FRONT_CENTER]+= M_SQRT1_2;
}else
av_assert0(0);
}
if(unaccounted & AV_CH_LAYOUT_STEREO){
if(s->out_ch_layout & AV_CH_FRONT_CENTER){
matrix[FRONT_CENTER][ FRONT_LEFT]+= M_SQRT1_2;
matrix[FRONT_CENTER][FRONT_RIGHT]+= M_SQRT1_2;
if(s->in_ch_layout & AV_CH_FRONT_CENTER)
matrix[FRONT_CENTER][ FRONT_CENTER] = s->clev*sqrt(2);
}else
av_assert0(0);
}
if(unaccounted & AV_CH_BACK_CENTER){
if(s->out_ch_layout & AV_CH_BACK_LEFT){
matrix[ BACK_LEFT][BACK_CENTER]+= M_SQRT1_2;
matrix[BACK_RIGHT][BACK_CENTER]+= M_SQRT1_2;
}else if(s->out_ch_layout & AV_CH_SIDE_LEFT){
matrix[ SIDE_LEFT][BACK_CENTER]+= M_SQRT1_2;
matrix[SIDE_RIGHT][BACK_CENTER]+= M_SQRT1_2;
}else if(s->out_ch_layout & AV_CH_FRONT_LEFT){
matrix[ FRONT_LEFT][BACK_CENTER]+= s->slev*M_SQRT1_2;
matrix[FRONT_RIGHT][BACK_CENTER]+= s->slev*M_SQRT1_2;
}else if(s->out_ch_layout & AV_CH_FRONT_CENTER){
matrix[ FRONT_CENTER][BACK_CENTER]+= s->slev*M_SQRT1_2;
}else
av_assert0(0);
}
if(unaccounted & AV_CH_BACK_LEFT){
if(s->out_ch_layout & AV_CH_BACK_CENTER){
matrix[BACK_CENTER][ BACK_LEFT]+= M_SQRT1_2;
matrix[BACK_CENTER][BACK_RIGHT]+= M_SQRT1_2;
}else if(s->out_ch_layout & AV_CH_SIDE_LEFT){
if(s->in_ch_layout & AV_CH_SIDE_LEFT){
matrix[ SIDE_LEFT][ BACK_LEFT]+= M_SQRT1_2;
matrix[SIDE_RIGHT][BACK_RIGHT]+= M_SQRT1_2;
}else{
matrix[ SIDE_LEFT][ BACK_LEFT]+= 1.0;
matrix[SIDE_RIGHT][BACK_RIGHT]+= 1.0;
}
}else if(s->out_ch_layout & AV_CH_FRONT_LEFT){
matrix[ FRONT_LEFT][ BACK_LEFT]+= s->slev;
matrix[FRONT_RIGHT][BACK_RIGHT]+= s->slev;
}else if(s->out_ch_layout & AV_CH_FRONT_CENTER){
matrix[ FRONT_CENTER][BACK_LEFT ]+= s->slev*M_SQRT1_2;
matrix[ FRONT_CENTER][BACK_RIGHT]+= s->slev*M_SQRT1_2;
}else
av_assert0(0);
}
if(unaccounted & AV_CH_SIDE_LEFT){
if(s->out_ch_layout & AV_CH_BACK_LEFT){
/* if back channels do not exist in the input, just copy side
channels to back channels, otherwise mix side into back */
if (s->in_ch_layout & AV_CH_BACK_LEFT) {
matrix[BACK_LEFT ][SIDE_LEFT ] += M_SQRT1_2;
matrix[BACK_RIGHT][SIDE_RIGHT] += M_SQRT1_2;
} else {
matrix[BACK_LEFT ][SIDE_LEFT ] += 1.0;
matrix[BACK_RIGHT][SIDE_RIGHT] += 1.0;
}
}else if(s->out_ch_layout & AV_CH_BACK_CENTER){
matrix[BACK_CENTER][ SIDE_LEFT]+= M_SQRT1_2;
matrix[BACK_CENTER][SIDE_RIGHT]+= M_SQRT1_2;
}else if(s->out_ch_layout & AV_CH_FRONT_LEFT){
matrix[ FRONT_LEFT][ SIDE_LEFT]+= s->slev;
matrix[FRONT_RIGHT][SIDE_RIGHT]+= s->slev;
}else if(s->out_ch_layout & AV_CH_FRONT_CENTER){
matrix[ FRONT_CENTER][SIDE_LEFT ]+= s->slev*M_SQRT1_2;
matrix[ FRONT_CENTER][SIDE_RIGHT]+= s->slev*M_SQRT1_2;
}else
av_assert0(0);
}
if(unaccounted & AV_CH_FRONT_LEFT_OF_CENTER){
if(s->out_ch_layout & AV_CH_FRONT_LEFT){
matrix[ FRONT_LEFT][ FRONT_LEFT_OF_CENTER]+= 1.0;
matrix[FRONT_RIGHT][FRONT_RIGHT_OF_CENTER]+= 1.0;
}else if(s->out_ch_layout & AV_CH_FRONT_CENTER){
matrix[ FRONT_CENTER][ FRONT_LEFT_OF_CENTER]+= M_SQRT1_2;
matrix[ FRONT_CENTER][FRONT_RIGHT_OF_CENTER]+= M_SQRT1_2;
}else
av_assert0(0);
}
/* mix LFE into front left/right or center */
if (unaccounted & AV_CH_LOW_FREQUENCY) {
if (s->out_ch_layout & AV_CH_FRONT_CENTER) {
matrix[FRONT_CENTER][LOW_FREQUENCY] += s->lfe_mix_level;
} else if (s->out_ch_layout & AV_CH_FRONT_LEFT) {
matrix[FRONT_LEFT ][LOW_FREQUENCY] += s->lfe_mix_level * M_SQRT1_2;
matrix[FRONT_RIGHT][LOW_FREQUENCY] += s->lfe_mix_level * M_SQRT1_2;
} else
av_assert0(0);
}
for(out_i=i=0; i<64; i++){
double sum=0;
int in_i=0;
for(j=0; j<64; j++){
s->matrix[out_i][in_i]= matrix[i][j];
if(matrix[i][j]){
sum += fabs(matrix[i][j]);
}
if(s->in_ch_layout & (1ULL<<j))
in_i++;
}
maxcoef= FFMAX(maxcoef, sum);
if(s->out_ch_layout & (1ULL<<i))
out_i++;
}
if(s->rematrix_volume < 0)
maxcoef = -s->rematrix_volume;
if(( av_get_packed_sample_fmt(s->out_sample_fmt) < AV_SAMPLE_FMT_FLT
|| av_get_packed_sample_fmt(s->int_sample_fmt) < AV_SAMPLE_FMT_FLT) && maxcoef > 1.0){
for(i=0; i<SWR_CH_MAX; i++)
for(j=0; j<SWR_CH_MAX; j++){
s->matrix[i][j] /= maxcoef;
}
}
if(s->rematrix_volume > 0){
for(i=0; i<SWR_CH_MAX; i++)
for(j=0; j<SWR_CH_MAX; j++){
s->matrix[i][j] *= s->rematrix_volume;
}
}
for(i=0; i<av_get_channel_layout_nb_channels(s->out_ch_layout); i++){
for(j=0; j<av_get_channel_layout_nb_channels(s->in_ch_layout); j++){
av_log(NULL, AV_LOG_DEBUG, "%f ", s->matrix[i][j]);
}
av_log(NULL, AV_LOG_DEBUG, "\n");
}
return 0;
}
static int filter_frame(AVFilterLink *inlink, AVFrame *buf)
{
VolumeContext *vol = inlink->dst->priv;
AVFilterLink *outlink = inlink->dst->outputs[0];
int nb_samples = buf->nb_samples;
AVFrame *out_buf;
AVFrameSideData *sd = av_frame_get_side_data(buf, AV_FRAME_DATA_REPLAYGAIN);
int ret;
if (sd && vol->replaygain != REPLAYGAIN_IGNORE) {
if (vol->replaygain != REPLAYGAIN_DROP) {
AVReplayGain *replaygain = (AVReplayGain*)sd->data;
int32_t gain = 100000;
uint32_t peak = 100000;
float g, p;
if (vol->replaygain == REPLAYGAIN_TRACK &&
replaygain->track_gain != INT32_MIN) {
gain = replaygain->track_gain;
if (replaygain->track_peak != 0)
peak = replaygain->track_peak;
} else if (replaygain->album_gain != INT32_MIN) {
gain = replaygain->album_gain;
if (replaygain->album_peak != 0)
peak = replaygain->album_peak;
} else {
av_log(inlink->dst, AV_LOG_WARNING, "Both ReplayGain gain "
"values are unknown.\n");
}
g = gain / 100000.0f;
p = peak / 100000.0f;
av_log(inlink->dst, AV_LOG_VERBOSE,
"Using gain %f dB from replaygain side data.\n", g);
vol->volume = pow(10, (g + vol->replaygain_preamp) / 20);
if (vol->replaygain_noclip)
vol->volume = FFMIN(vol->volume, 1.0 / p);
vol->volume_i = (int)(vol->volume * 256 + 0.5);
volume_init(vol);
}
av_frame_remove_side_data(buf, AV_FRAME_DATA_REPLAYGAIN);
}
if (vol->volume == 1.0 || vol->volume_i == 256)
return ff_filter_frame(outlink, buf);
/* do volume scaling in-place if input buffer is writable */
if (av_frame_is_writable(buf)) {
out_buf = buf;
} else {
out_buf = ff_get_audio_buffer(inlink, nb_samples);
if (!out_buf)
return AVERROR(ENOMEM);
ret = av_frame_copy_props(out_buf, buf);
if (ret < 0) {
av_frame_free(&out_buf);
av_frame_free(&buf);
return ret;
}
}
if (vol->precision != PRECISION_FIXED || vol->volume_i > 0) {
int p, plane_samples;
if (av_sample_fmt_is_planar(buf->format))
plane_samples = FFALIGN(nb_samples, vol->samples_align);
else
plane_samples = FFALIGN(nb_samples * vol->channels, vol->samples_align);
if (vol->precision == PRECISION_FIXED) {
for (p = 0; p < vol->planes; p++) {
vol->scale_samples(out_buf->extended_data[p],
buf->extended_data[p], plane_samples,
vol->volume_i);
}
} else if (av_get_packed_sample_fmt(vol->sample_fmt) == AV_SAMPLE_FMT_FLT) {
for (p = 0; p < vol->planes; p++) {
vol->fdsp.vector_fmul_scalar((float *)out_buf->extended_data[p],
(const float *)buf->extended_data[p],
vol->volume, plane_samples);
}
} else {
for (p = 0; p < vol->planes; p++) {
vol->fdsp.vector_dmul_scalar((double *)out_buf->extended_data[p],
(const double *)buf->extended_data[p],
vol->volume, plane_samples);
}
}
}
emms_c();
if (buf != out_buf)
av_frame_free(&buf);
return ff_filter_frame(outlink, out_buf);
}
void ff_audio_convert_set_func(AudioConvert *ac, enum AVSampleFormat out_fmt,
enum AVSampleFormat in_fmt, int channels,
int ptr_align, int samples_align,
const char *descr, void *conv)
{
int found = 0;
switch (ac->func_type) {
case CONV_FUNC_TYPE_FLAT:
if (av_get_packed_sample_fmt(ac->in_fmt) == in_fmt &&
av_get_packed_sample_fmt(ac->out_fmt) == out_fmt) {
ac->conv_flat = conv;
ac->func_descr = descr;
ac->ptr_align = ptr_align;
ac->samples_align = samples_align;
if (ptr_align == 1 && samples_align == 1) {
ac->conv_flat_generic = conv;
ac->func_descr_generic = descr;
} else {
ac->has_optimized_func = 1;
}
found = 1;
}
break;
case CONV_FUNC_TYPE_INTERLEAVE:
if (ac->in_fmt == in_fmt && ac->out_fmt == out_fmt &&
(!channels || ac->channels == channels)) {
ac->conv_interleave = conv;
ac->func_descr = descr;
ac->ptr_align = ptr_align;
ac->samples_align = samples_align;
if (ptr_align == 1 && samples_align == 1) {
ac->conv_interleave_generic = conv;
ac->func_descr_generic = descr;
} else {
ac->has_optimized_func = 1;
}
found = 1;
}
break;
case CONV_FUNC_TYPE_DEINTERLEAVE:
if (ac->in_fmt == in_fmt && ac->out_fmt == out_fmt &&
(!channels || ac->channels == channels)) {
ac->conv_deinterleave = conv;
ac->func_descr = descr;
ac->ptr_align = ptr_align;
ac->samples_align = samples_align;
if (ptr_align == 1 && samples_align == 1) {
ac->conv_deinterleave_generic = conv;
ac->func_descr_generic = descr;
} else {
ac->has_optimized_func = 1;
}
found = 1;
}
break;
}
if (found) {
av_log(ac->avr, AV_LOG_DEBUG, "audio_convert: found function: %-4s "
"to %-4s (%s)\n", av_get_sample_fmt_name(ac->in_fmt),
av_get_sample_fmt_name(ac->out_fmt), descr);
}
}
void AUD_FFMPEGReader::init()
{
m_position = 0;
m_pkgbuf_left = 0;
if(avformat_find_stream_info(m_formatCtx, NULL) < 0)
AUD_THROW(AUD_ERROR_FFMPEG, streaminfo_error);
// find audio stream and codec
m_stream = -1;
for(unsigned int i = 0; i < m_formatCtx->nb_streams; i++)
{
if((m_formatCtx->streams[i]->codec->codec_type == AVMEDIA_TYPE_AUDIO)
&& (m_stream < 0))
{
m_stream=i;
break;
}
}
if(m_stream == -1)
AUD_THROW(AUD_ERROR_FFMPEG, noaudio_error);
m_codecCtx = m_formatCtx->streams[m_stream]->codec;
// get a decoder and open it
AVCodec *aCodec = avcodec_find_decoder(m_codecCtx->codec_id);
if(!aCodec)
AUD_THROW(AUD_ERROR_FFMPEG, nodecoder_error);
if(avcodec_open2(m_codecCtx, aCodec, NULL) < 0)
AUD_THROW(AUD_ERROR_FFMPEG, codecopen_error);
// XXX this prints file information to stdout:
//dump_format(m_formatCtx, 0, NULL, 0);
m_specs.channels = (AUD_Channels) m_codecCtx->channels;
m_tointerleave = av_sample_fmt_is_planar(m_codecCtx->sample_fmt);
switch(av_get_packed_sample_fmt(m_codecCtx->sample_fmt))
{
case AV_SAMPLE_FMT_U8:
m_convert = AUD_convert_u8_float;
m_specs.format = AUD_FORMAT_U8;
break;
case AV_SAMPLE_FMT_S16:
m_convert = AUD_convert_s16_float;
m_specs.format = AUD_FORMAT_S16;
break;
case AV_SAMPLE_FMT_S32:
m_convert = AUD_convert_s32_float;
m_specs.format = AUD_FORMAT_S32;
break;
case AV_SAMPLE_FMT_FLT:
m_convert = AUD_convert_copy<float>;
m_specs.format = AUD_FORMAT_FLOAT32;
break;
case AV_SAMPLE_FMT_DBL:
m_convert = AUD_convert_double_float;
m_specs.format = AUD_FORMAT_FLOAT64;
break;
default:
AUD_THROW(AUD_ERROR_FFMPEG, format_error);
}
m_specs.rate = (AUD_SampleRate) m_codecCtx->sample_rate;
}
void AudioDecoderThread::decodePacket(AVPacket* pPacket)
{
char* pDecodedData = 0;
AVPacket* pTempPacket = new AVPacket;
av_init_packet(pTempPacket);
pTempPacket->data = pPacket->data;
pTempPacket->size = pPacket->size;
AVFrame* pDecodedFrame;
pDecodedFrame = avcodec_alloc_frame();
while (pTempPacket->size > 0) {
int gotFrame = 0;
int bytesDecoded;
int bytesConsumed = avcodec_decode_audio4(m_pStream->codec, pDecodedFrame,
&gotFrame, pTempPacket);
if (gotFrame) {
bytesDecoded = av_samples_get_buffer_size(0, m_pStream->codec->channels,
pDecodedFrame->nb_samples, m_pStream->codec->sample_fmt, 1);
pDecodedData = (char*)(pDecodedFrame->data[0]);
} else {
bytesDecoded = 0;
}
// This is triggered for some strange/broken videos.
// AVG_ASSERT(bytesConsumed != 0);
if (bytesConsumed < 0) {
// Error decoding -> throw away current packet.
bytesDecoded = 0;
pTempPacket->size = 0;
} else {
pTempPacket->data += bytesConsumed;
pTempPacket->size -= bytesConsumed;
}
if (bytesDecoded > 0) {
int framesDecoded = bytesDecoded/(m_pStream->codec->channels*
getBytesPerSample(m_InputSampleFormat));
AudioBufferPtr pBuffer;
bool bNeedsResample = (m_InputSampleRate != m_AP.m_SampleRate ||
m_InputSampleFormat != AV_SAMPLE_FMT_S16 ||
m_pStream->codec->channels != m_AP.m_Channels);
bool bIsPlanar = false;
#if LIBAVUTIL_VERSION_INT >= AV_VERSION_INT(51, 27, 0)
bIsPlanar = av_sample_fmt_is_planar((AVSampleFormat)m_InputSampleFormat);
if (bIsPlanar) {
char* pPackedData = (char*)av_malloc(AVCODEC_MAX_AUDIO_FRAME_SIZE +
FF_INPUT_BUFFER_PADDING_SIZE);
planarToInterleaved(pPackedData, pDecodedFrame, m_pStream->codec->channels,
framesDecoded);
pBuffer = resampleAudio(pPackedData, framesDecoded,
av_get_packed_sample_fmt((AVSampleFormat)m_InputSampleFormat));
av_free(pPackedData);
bNeedsResample = false;
}
#endif
if (bNeedsResample) {
pBuffer = resampleAudio(pDecodedData, framesDecoded,
m_InputSampleFormat);
} else if (!bIsPlanar) {
pBuffer = AudioBufferPtr(new AudioBuffer(framesDecoded, m_AP));
memcpy(pBuffer->getData(), pDecodedData, bytesDecoded);
}
m_LastFrameTime += float(pBuffer->getNumFrames())/m_AP.m_SampleRate;
pushAudioMsg(pBuffer, m_LastFrameTime);
}
}
#if LIBAVCODEC_VERSION_MAJOR > 53
avcodec_free_frame(&pDecodedFrame);
delete pDecodedFrame;
#endif
delete pTempPacket;
}
SCM make_ffmpeg_output(SCM scm_file_name,
SCM scm_format_name,
SCM scm_video_parameters,
SCM scm_have_video,
SCM scm_audio_parameters,
SCM scm_have_audio,
SCM scm_debug)
{
SCM retval;
struct ffmpeg_t *self;
scm_dynwind_begin(0);
const char *file_name = scm_to_locale_string(scm_file_name);
scm_dynwind_free(file_name);
self = (struct ffmpeg_t *)scm_gc_calloc(sizeof(struct ffmpeg_t), "ffmpeg");
self->video_stream_idx = -1;
self->audio_stream_idx = -1;
SCM_NEWSMOB(retval, ffmpeg_tag, self);
int err;
const char *format_name = NULL;
if (!scm_is_false(scm_format_name)) {
format_name = scm_to_locale_string(scm_symbol_to_string(scm_format_name));
scm_dynwind_free(format_name);
};
#ifdef HAVE_AVFORMAT_ALLOC_OUTPUT_CONTEXT2
err = avformat_alloc_output_context2(&self->fmt_ctx, NULL, format_name, file_name);
if (!self->fmt_ctx) {
ffmpeg_destroy(retval);
scm_misc_error("make-ffmpeg-output", "Error initializing output format for file '~a': ~a",
scm_list_2(scm_file_name, get_error_text(err)));
};
#else
AVOutputFormat *format;
if (format_name)
format = av_guess_format(format_name, NULL, NULL);
else
format = av_guess_format(NULL, file_name, NULL);
if (!format) {
ffmpeg_destroy(retval);
scm_misc_error("make-ffmpeg-output", "Unable to determine file format for file '~a'",
scm_list_1(scm_file_name));
};
self->fmt_ctx = avformat_alloc_context();
if (!self->fmt_ctx) {
ffmpeg_destroy(retval);
scm_misc_error("make-ffmpeg-output", "Error initializing output format for file '~a'",
scm_list_1(scm_file_name));
};
self->fmt_ctx->oformat = format;
strncpy(self->fmt_ctx->filename, file_name, sizeof(self->fmt_ctx->filename));
#endif
char have_video = scm_is_true(scm_have_video);
if (have_video) {
// Open codec and video stream
enum AVCodecID video_codec_id = self->fmt_ctx->oformat->video_codec;
AVCodec *video_encoder = find_encoder(retval, video_codec_id, "video");
AVStream *video_stream = open_output_stream(retval, video_encoder, &self->video_stream_idx, "video", scm_file_name);
// Get video parameters
SCM scm_shape = scm_car(scm_video_parameters);
SCM scm_frame_rate = scm_cadr(scm_video_parameters);
SCM scm_video_bit_rate = scm_caddr(scm_video_parameters);
SCM scm_aspect_ratio = scm_cadddr(scm_video_parameters);
// Configure the output video codec
self->video_codec_ctx =
configure_output_video_codec(video_stream, video_codec_id, scm_video_bit_rate, scm_shape, scm_frame_rate, scm_aspect_ratio);
// Some formats want stream headers to be separate.
if (self->fmt_ctx->oformat->flags & AVFMT_GLOBALHEADER)
self->video_codec_ctx->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
// Open output video codec
open_codec(retval, self->video_codec_ctx, video_encoder, "video", scm_file_name);
// Allocate frame
self->video_target_frame = allocate_output_video_frame(retval, self->video_codec_ctx);
};
char have_audio = scm_is_true(scm_have_audio);
if (have_audio) {
// Open audio codec and stream
enum AVCodecID audio_codec_id = self->fmt_ctx->oformat->audio_codec;
AVCodec *audio_encoder = find_encoder(retval, audio_codec_id, "audio");
AVStream *audio_stream = open_output_stream(retval, audio_encoder, &self->audio_stream_idx, "audio", scm_file_name);
// Get audio parameters
SCM scm_select_rate = scm_car(scm_audio_parameters);
SCM scm_channels = scm_cadr(scm_audio_parameters);
SCM scm_audio_bit_rate = scm_caddr(scm_audio_parameters);
SCM scm_select_format = scm_cadddr(scm_audio_parameters);
// Configure the output audio codec
self->audio_codec_ctx =
configure_output_audio_codec(retval, audio_stream, audio_codec_id,
scm_select_rate, scm_channels, scm_audio_bit_rate, scm_select_format);
// Some formats want stream headers to be separate.
if (self->fmt_ctx->oformat->flags & AVFMT_GLOBALHEADER)
self->audio_codec_ctx->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
// Open output audio codec
open_codec(retval, self->audio_codec_ctx, audio_encoder, "audio", scm_file_name);
// Allocate audio frame
self->audio_target_frame =
allocate_output_audio_frame(retval, self->audio_codec_ctx, self->audio_codec_ctx->sample_fmt);
self->audio_packed_frame =
allocate_output_audio_frame(retval, self->audio_codec_ctx, av_get_packed_sample_fmt(self->audio_codec_ctx->sample_fmt));
// Initialise audio buffer
ringbuffer_init(&self->audio_buffer, 1024);
};
if (scm_is_true(scm_debug)) av_dump_format(self->fmt_ctx, 0, file_name, 1);
// Open the output file if needed
if (!(self->fmt_ctx->oformat->flags & AVFMT_NOFILE)) {
int err = avio_open(&self->fmt_ctx->pb, file_name, AVIO_FLAG_WRITE);
if (err < 0) {
ffmpeg_destroy(retval);
scm_misc_error("make-ffmpeg-output", "Could not open '~a': ~a",
scm_list_2(scm_file_name, get_error_text(err)));
}
self->output_file = 1;
}
// Write video file header
err = avformat_write_header(self->fmt_ctx, NULL);
if (err < 0) {
ffmpeg_destroy(retval);
scm_misc_error("make-ffmpeg-output", "Error writing header of video '~a': ~a",
scm_list_2(scm_file_name, get_error_text(err)));
};
self->header_written = 1;
scm_dynwind_end();
return retval;
}
