summaryrefslogtreecommitdiffstats
path: root/libhb
diff options
context:
space:
mode:
authorJohn Stebbins <[email protected]>2016-12-08 18:09:06 -0800
committerJohn Stebbins <[email protected]>2016-12-17 07:28:51 -0800
commit9e39762cce248ee1356742663c11b1b53d0e2ba1 (patch)
tree0b463d45b4f555135fe7f2f073a7ddc5d4654afd /libhb
parente159ab1662ba8588e97262936df9dcccf6b07ad1 (diff)
encavcodecaudio: fix use of deprecated libav interfaces
avcodec_encode_audio2 is deprecated, use avcodec_send_frame and avcodec_receive_packet
Diffstat (limited to 'libhb')
-rw-r--r--libhb/encavcodecaudio.c211
1 files changed, 100 insertions, 111 deletions
diff --git a/libhb/encavcodecaudio.c b/libhb/encavcodecaudio.c
index b79aada6f..f6779fe31 100644
--- a/libhb/encavcodecaudio.c
+++ b/libhb/encavcodecaudio.c
@@ -285,20 +285,6 @@ static void Finalize(hb_work_object_t *w)
{
hb_work_private_t *pv = w->private_data;
- // Finalize with NULL input needed by FLAC to generate md5sum
- // in context extradata
-
- // Prepare output packet
- AVPacket pkt;
- int got_packet;
- hb_buffer_t *buf = hb_buffer_init(pv->max_output_bytes);
- av_init_packet(&pkt);
- pkt.data = buf->data;
- pkt.size = buf->alloc;
-
- avcodec_encode_audio2(pv->context, &pkt, NULL, &got_packet);
- hb_buffer_close(&buf);
-
// Then we need to recopy the header since it was modified
if (pv->context->extradata != NULL)
{
@@ -317,7 +303,7 @@ static void encavcodecaClose(hb_work_object_t * w)
if (pv->context != NULL)
{
Finalize(w);
- hb_deep_log(2, "encavcodeca: closing libavcodec");
+ hb_deep_log(2, "encavcodecaudio: closing libavcodec");
if (pv->context->codec != NULL)
avcodec_flush_buffers(pv->context);
hb_avcodec_close(pv->context);
@@ -349,76 +335,35 @@ static void encavcodecaClose(hb_work_object_t * w)
}
}
-static hb_buffer_t* Encode(hb_work_object_t *w)
+static void get_packets( hb_work_object_t * w, hb_buffer_list_t * list )
{
- hb_work_private_t *pv = w->private_data;
- hb_audio_t *audio = w->audio;
- uint64_t pts, pos;
-
- if (hb_list_bytes(pv->list) < pv->input_samples * sizeof(float))
- {
- return NULL;
- }
-
- hb_list_getbytes(pv->list, pv->input_buf, pv->input_samples * sizeof(float),
- &pts, &pos);
+ hb_work_private_t * pv = w->private_data;
+ hb_audio_t * audio = w->audio;
- // Prepare input frame
- int out_linesize;
- int out_size = av_samples_get_buffer_size(&out_linesize,
- pv->context->channels,
- pv->samples_per_frame,
- pv->context->sample_fmt, 1);
- AVFrame frame = { .nb_samples = pv->samples_per_frame, };
- avcodec_fill_audio_frame(&frame,
- pv->context->channels, pv->context->sample_fmt,
- pv->output_buf, out_size, 1);
- if (pv->avresample != NULL)
+ while (1)
{
- int in_linesize;
- av_samples_get_buffer_size(&in_linesize, pv->context->channels,
- frame.nb_samples, AV_SAMPLE_FMT_FLT, 1);
- int out_samples = avresample_convert(pv->avresample,
- frame.extended_data, out_linesize,
- frame.nb_samples,
- &pv->input_buf, in_linesize,
- frame.nb_samples);
- if (out_samples != pv->samples_per_frame)
+ // Prepare output packet
+ int ret;
+ AVPacket pkt;
+ hb_buffer_t * out;
+
+ av_init_packet(&pkt);
+ ret = avcodec_receive_packet(pv->context, &pkt);
+ if (ret == AVERROR(EAGAIN) || ret == AVERROR_EOF)
{
- // we're not doing sample rate conversion, so this shouldn't happen
- hb_log("encavcodecaWork: avresample_convert() failed");
- return NULL;
+ break;
+ }
+ if (ret < 0)
+ {
+ hb_log("encavcodecaudio: avcodec_receive_packet failed");
+ break;
}
- }
- // Libav requires that timebase of audio frames be in sample_rate units
- frame.pts = pts + (90000 * pos / (sizeof(float) *
- pv->out_discrete_channels *
- audio->config.out.samplerate));
- frame.pts = av_rescale(frame.pts, pv->context->sample_rate, 90000);
-
- // Prepare output packet
- AVPacket pkt;
- int got_packet;
- hb_buffer_t *out = hb_buffer_init(pv->max_output_bytes);
- av_init_packet(&pkt);
- pkt.data = out->data;
- pkt.size = out->alloc;
-
- // Encode
- int ret = avcodec_encode_audio2(pv->context, &pkt, &frame, &got_packet);
- if (ret < 0)
- {
- hb_log("encavcodeca: avcodec_encode_audio failed");
- hb_buffer_close(&out);
- return NULL;
- }
+ out = hb_buffer_init(pkt.size);
+ memcpy(out->data, pkt.data, out->size);
- if (got_packet && pkt.size)
- {
- out->size = pkt.size;
- // The output pts from libav is in context->time_base. Convert it back
- // to our timebase.
+ // The output pts from libav is in context->time_base. Convert it
+ // back to our timebase.
out->s.start = av_rescale_q(pkt.pts, pv->context->time_base,
(AVRational){1, 90000});
out->s.duration = (double)90000 * pv->samples_per_frame /
@@ -426,31 +371,78 @@ static hb_buffer_t* Encode(hb_work_object_t *w)
out->s.stop = out->s.start + out->s.duration;
out->s.type = AUDIO_BUF;
out->s.frametype = HB_FRAME_AUDIO;
- }
- else
- {
- hb_buffer_close(&out);
- return Encode(w);
- }
- return out;
+ hb_buffer_list_append(list, out);
+ }
}
-static hb_buffer_t * Flush( hb_work_object_t * w )
+static void Encode(hb_work_object_t *w, hb_buffer_list_t *list)
{
- hb_buffer_list_t list;
- hb_buffer_t *buf;
+ hb_work_private_t * pv = w->private_data;
+ hb_audio_t * audio = w->audio;
+ uint64_t pts, pos;
- hb_buffer_list_clear(&list);
- buf = Encode( w );
- while (buf != NULL)
+ while (hb_list_bytes(pv->list) > pv->input_samples * sizeof(float))
{
- hb_buffer_list_append(&list, buf);
- buf = Encode( w );
+ int ret;
+
+ hb_list_getbytes(pv->list, pv->input_buf,
+ pv->input_samples * sizeof(float), &pts, &pos);
+
+ // Prepare input frame
+ int out_linesize, out_size;
+ AVFrame frame = { .nb_samples = pv->samples_per_frame, };
+
+ out_size = av_samples_get_buffer_size(&out_linesize,
+ pv->context->channels,
+ pv->samples_per_frame,
+ pv->context->sample_fmt, 1);
+ avcodec_fill_audio_frame(&frame,
+ pv->context->channels, pv->context->sample_fmt,
+ pv->output_buf, out_size, 1);
+ if (pv->avresample != NULL)
+ {
+ int in_linesize, out_samples;
+
+ av_samples_get_buffer_size(&in_linesize, pv->context->channels,
+ frame.nb_samples, AV_SAMPLE_FMT_FLT, 1);
+ out_samples = avresample_convert(pv->avresample,
+ frame.extended_data, out_linesize,
+ frame.nb_samples, &pv->input_buf,
+ in_linesize, frame.nb_samples);
+ if (out_samples != pv->samples_per_frame)
+ {
+ // we're not doing sample rate conversion,
+ // so this shouldn't happen
+ hb_log("encavcodecaWork: avresample_convert() failed");
+ continue;
+ }
+ }
+
+ // Libav requires that timebase of audio frames be in sample_rate units
+ frame.pts = pts + (90000 * pos / (sizeof(float) *
+ pv->out_discrete_channels *
+ audio->config.out.samplerate));
+ frame.pts = av_rescale(frame.pts, pv->context->sample_rate, 90000);
+
+
+ // Encode
+ ret = avcodec_send_frame(pv->context, &frame);
+ if (ret < 0)
+ {
+ hb_log("encavcodecaudio: avcodec_send_frame failed");
+ return;
+ }
+ get_packets(w, list);
}
+}
- hb_buffer_list_append(&list, hb_buffer_eof_init());
- return hb_buffer_list_clear(&list);
+static void Flush( hb_work_object_t * w, hb_buffer_list_t * list )
+{
+ hb_work_private_t * pv = w->private_data;
+
+ avcodec_send_frame(pv->context, NULL);
+ get_packets(w, list);
}
/***********************************************************************
@@ -462,34 +454,31 @@ static int encavcodecaWork( hb_work_object_t * w, hb_buffer_t ** buf_in,
hb_buffer_t ** buf_out )
{
hb_work_private_t * pv = w->private_data;
- hb_buffer_t * in = *buf_in, * buf;
- hb_buffer_list_t list;
+ hb_buffer_t * in = *buf_in;
+ hb_buffer_list_t list;
- if (in->s.flags & HB_BUF_FLAG_EOF)
+ if (pv->context == NULL || pv->context->codec == NULL)
{
- /* EOF on input - send it downstream & say we're done */
- *buf_out = Flush( w );
+ hb_error("encavcodecaudio: codec context is uninitialized");
return HB_WORK_DONE;
}
- if ( pv->context == NULL || pv->context->codec == NULL )
+ hb_buffer_list_clear(&list);
+ if (in->s.flags & HB_BUF_FLAG_EOF)
{
- // No encoder context. Nothing we can do.
- return HB_WORK_OK;
+ /* EOF on input - send it downstream & say we're done */
+ Flush(w, &list);
+ hb_buffer_list_append(&list, hb_buffer_eof_init());
+ *buf_out = hb_buffer_list_clear(&list);
+ return HB_WORK_DONE;
}
hb_list_add( pv->list, in );
*buf_in = NULL;
- hb_buffer_list_clear(&list);
- buf = Encode( w );
- while (buf != NULL)
- {
- hb_buffer_list_append(&list, buf);
- buf = Encode( w );
- }
-
+ Encode(w, &list);
*buf_out = hb_buffer_list_clear(&list);
+
return HB_WORK_OK;
}