Hi,
I have problem with audio_resample procedure. I decode AMR (over RTP) stream.
Problem one: otput of ffmpeg is after calling avcodec_open(avctx, codec) set
to SAMPLE_FMT_FLT (however I set avctx->sample_fmt to SAMPLE_FMT_S16). But OK I
try another way:
Problem two: Because I need to convert output of avcodec_decode_audio3 to
SAMPLE_FMT_S16 I try this:
if ( dec->context->sample_fmt != AUDIO_FORMAT_OUT) {
// create convertor
ReSampleContext *ctx = av_audio_resample_init(1, dec->context->channels,
A_SAMPLE_RATE_OUT, dec->context->sample_rate,
AUDIO_FORMAT_OUT, dec->context->sample_fmt,
16, 10, 0, 0.8);
if (ctx == NULL) {
av_log(dec->context, AV_LOG_ERROR, "Resampling context cannot be
created.");
return -1;
}
// size of one sample on input buffer and output buffer
int isamp = av_get_bits_per_sample_format( dec->context->sample_fmt )
>> 3,
osamp = av_get_bits_per_sample_format( AUDIO_FORMAT_OUT ) >> 3;
// space necessary in samples output buffer
int extra_samples = ( dec->output_size * osamp ) / isamp;
if (extra_samples + dec->samples_length >= AUDIO_BUFFER_SIZE) {
av_log(NULL, AV_LOG_ERROR, "not enough memory");
return -1;
}
extra_samples = audio_resample(ctx, dec->samples+dec->samples_length,
dec->ffmpeg_output, dec->output_size/isamp);
dec->samples_length += extra_samples*osamp;
audio_resample_close( ctx );
}
in spite of putting 160 samples to audio_resample and output frequency is the
same as input frequency I obtain extra_samples = 150 and number of processed
samples (inside audio_resample value consumed) is 141.
Problem three: The sound is only noise not sound.
Any suggestion?
Note: I use last git version of ffmpeg
Thanks for any help
gimli
_______________________________________________
libav-user mailing list
[email protected]
https://lists.mplayerhq.hu/mailman/listinfo/libav-user
