FFmpeg: how to encode last audio frame? - ffmpeg

Question

Considering the following snippet: ( from https://ffmpeg.org/doxygen/trunk/encode_audio_8c-example.html )
for (i = 0; i < 200; i++) {
/* make sure the frame is writable -- makes a copy if the encoder
* kept a reference internally */
ret = av_frame_make_writable(frame);
if (ret < 0)
exit(1);
samples = (uint16_t*)frame->data[0];
for (j = 0; j < c->frame_size; j++) {
samples[2*j] = (int)(sin(t) * 10000);
for (k = 1; k < c->channels; k++)
samples[2*j + k] = samples[2*j];
t += tincr;
}
encode(c, frame, pkt, f);
}
If I understand the example correctly, the generated audio stream consists exactly of 200 frames of size c->frame_size which are encoded and saved to disk.
However, if I want to encode a generic stream of data of size soundsize, I will have a certain number of frames of fixed size c->frame_size, i.e. size_t nframes = soundsize / c->frame_size; plus one last frame of size: size_t rem_lastframe = soundsize % c->frame_size;
Can you explain me how to process this last frame? The frame_size seems to be fixed and chosen by the codec.

Answer 1

This is what ffmpeg does.
if (src->nb_samples < avctx->frame_size) {
ret = pad_last_frame(avctx, dst, src);
...
You can use apad filter or mimic what libavcodec does
/**
* Pad last frame with silence.
*/
static int pad_last_frame(AVCodecContext *s, AVFrame *frame, const AVFrame *src)
{
int ret;
frame->format = src->format;
frame->channel_layout = src->channel_layout;
frame->channels = src->channels;
frame->nb_samples = s->frame_size;
ret = av_frame_get_buffer(frame, 0);
if (ret < 0)
goto fail;
ret = av_frame_copy_props(frame, src);
if (ret < 0)
goto fail;
if ((ret = av_samples_copy(frame->extended_data, src->extended_data, 0, 0,
src->nb_samples, s->channels, s->sample_fmt)) < 0)
goto fail;
if ((ret = av_samples_set_silence(frame->extended_data, src->nb_samples,
frame->nb_samples - src->nb_samples,
s->channels, s->sample_fmt)) < 0)
goto fail;
return 0;
fail:
av_frame_unref(frame);
return ret;
}