avfilter/af_asetnsamples: use lavfi internal queue

This commit is contained in:
Paul B Mahol 2018-10-03 12:20:34 +02:00
parent 7d65fe87e9
commit ef3babb2c7

View File

@ -24,20 +24,18 @@
* Filter that changes number of samples on single output operation
*/
#include "libavutil/audio_fifo.h"
#include "libavutil/avassert.h"
#include "libavutil/channel_layout.h"
#include "libavutil/opt.h"
#include "avfilter.h"
#include "audio.h"
#include "filters.h"
#include "internal.h"
#include "formats.h"
typedef struct ASNSContext {
const AVClass *class;
int nb_out_samples; ///< how many samples to output
AVAudioFifo *fifo; ///< samples are queued here
int64_t next_out_pts;
int pad;
} ASNSContext;
@ -54,134 +52,55 @@ static const AVOption asetnsamples_options[] = {
AVFILTER_DEFINE_CLASS(asetnsamples);
static av_cold int init(AVFilterContext *ctx)
static int activate(AVFilterContext *ctx)
{
ASNSContext *asns = ctx->priv;
asns->next_out_pts = AV_NOPTS_VALUE;
av_log(ctx, AV_LOG_VERBOSE, "nb_out_samples:%d pad:%d\n", asns->nb_out_samples, asns->pad);
return 0;
}
static av_cold void uninit(AVFilterContext *ctx)
{
ASNSContext *asns = ctx->priv;
av_audio_fifo_free(asns->fifo);
}
static int config_props_output(AVFilterLink *outlink)
{
ASNSContext *asns = outlink->src->priv;
asns->fifo = av_audio_fifo_alloc(outlink->format, outlink->channels, asns->nb_out_samples);
if (!asns->fifo)
return AVERROR(ENOMEM);
return 0;
}
static int push_samples(AVFilterLink *outlink)
{
ASNSContext *asns = outlink->src->priv;
AVFrame *outsamples = NULL;
int ret, nb_out_samples, nb_pad_samples;
if (asns->pad) {
nb_out_samples = av_audio_fifo_size(asns->fifo) ? asns->nb_out_samples : 0;
nb_pad_samples = nb_out_samples - FFMIN(nb_out_samples, av_audio_fifo_size(asns->fifo));
} else {
nb_out_samples = FFMIN(asns->nb_out_samples, av_audio_fifo_size(asns->fifo));
nb_pad_samples = 0;
}
if (!nb_out_samples)
return 0;
outsamples = ff_get_audio_buffer(outlink, nb_out_samples);
if (!outsamples)
return AVERROR(ENOMEM);
av_audio_fifo_read(asns->fifo,
(void **)outsamples->extended_data, nb_out_samples);
if (nb_pad_samples)
av_samples_set_silence(outsamples->extended_data, nb_out_samples - nb_pad_samples,
nb_pad_samples, outlink->channels,
outlink->format);
outsamples->nb_samples = nb_out_samples;
outsamples->channel_layout = outlink->channel_layout;
outsamples->sample_rate = outlink->sample_rate;
outsamples->pts = asns->next_out_pts;
if (asns->next_out_pts != AV_NOPTS_VALUE)
asns->next_out_pts += av_rescale_q(nb_out_samples, (AVRational){1, outlink->sample_rate}, outlink->time_base);
ret = ff_filter_frame(outlink, outsamples);
if (ret < 0)
return ret;
return nb_out_samples;
}
static int filter_frame(AVFilterLink *inlink, AVFrame *insamples)
{
AVFilterContext *ctx = inlink->dst;
ASNSContext *asns = ctx->priv;
AVFilterLink *inlink = ctx->inputs[0];
AVFilterLink *outlink = ctx->outputs[0];
ASNSContext *s = ctx->priv;
AVFrame *frame = NULL, *pad_frame;
int ret;
int nb_samples = insamples->nb_samples;
if (av_audio_fifo_space(asns->fifo) < nb_samples) {
av_log(ctx, AV_LOG_DEBUG, "No space for %d samples, stretching audio fifo\n", nb_samples);
ret = av_audio_fifo_realloc(asns->fifo, av_audio_fifo_size(asns->fifo) + nb_samples);
if (ret < 0) {
av_log(ctx, AV_LOG_ERROR,
"Stretching audio fifo failed, discarded %d samples\n", nb_samples);
return -1;
}
}
ret = av_audio_fifo_write(asns->fifo, (void **)insamples->extended_data, nb_samples);
if (ret > 0 && asns->next_out_pts == AV_NOPTS_VALUE)
asns->next_out_pts = insamples->pts;
av_frame_free(&insamples);
FF_FILTER_FORWARD_STATUS_BACK(outlink, inlink);
ret = ff_inlink_consume_samples(inlink, s->nb_out_samples, s->nb_out_samples, &frame);
if (ret < 0)
return ret;
while (av_audio_fifo_size(asns->fifo) >= asns->nb_out_samples)
push_samples(outlink);
return 0;
}
if (ret > 0) {
if ((!s->pad || (s->pad && frame->nb_samples == s->nb_out_samples)))
return ff_filter_frame(outlink, frame);
static int request_frame(AVFilterLink *outlink)
{
AVFilterLink *inlink = outlink->src->inputs[0];
int ret;
pad_frame = ff_get_audio_buffer(outlink, s->nb_out_samples);
if (!pad_frame)
return AVERROR(ENOMEM);
ret = ff_request_frame(inlink);
if (ret == AVERROR_EOF) {
ret = push_samples(outlink);
return ret < 0 ? ret : ret > 0 ? 0 : AVERROR_EOF;
av_samples_copy(pad_frame->extended_data, frame->extended_data,
0, 0, frame->nb_samples, frame->channels, frame->format);
av_samples_set_silence(pad_frame->extended_data, frame->nb_samples,
s->nb_out_samples - frame->nb_samples, frame->channels,
frame->format);
av_frame_free(&frame);
return ff_filter_frame(outlink, pad_frame);
}
return ret;
FF_FILTER_FORWARD_STATUS(inlink, outlink);
FF_FILTER_FORWARD_WANTED(outlink, inlink);
return FFERROR_NOT_READY;
}
static const AVFilterPad asetnsamples_inputs[] = {
{
.name = "default",
.type = AVMEDIA_TYPE_AUDIO,
.filter_frame = filter_frame,
.name = "default",
.type = AVMEDIA_TYPE_AUDIO,
},
{ NULL }
};
static const AVFilterPad asetnsamples_outputs[] = {
{
.name = "default",
.type = AVMEDIA_TYPE_AUDIO,
.request_frame = request_frame,
.config_props = config_props_output,
.name = "default",
.type = AVMEDIA_TYPE_AUDIO,
},
{ NULL }
};
@ -191,8 +110,7 @@ AVFilter ff_af_asetnsamples = {
.description = NULL_IF_CONFIG_SMALL("Set the number of samples for each output audio frames."),
.priv_size = sizeof(ASNSContext),
.priv_class = &asetnsamples_class,
.init = init,
.uninit = uninit,
.inputs = asetnsamples_inputs,
.outputs = asetnsamples_outputs,
.activate = activate,
};