aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorPaul B Mahol <onemda@gmail.com>2018-10-28 14:27:32 +0100
committerPaul B Mahol <onemda@gmail.com>2018-10-28 14:31:03 +0100
commitbb54c0ae719ef768463e78af82366e73c79e1e50 (patch)
treed2cb57de98b60fd95219f97f6562c3488ca103b5
parentbdfd2e3c79dd3d0a75033cad662a78fa0a04304f (diff)
downloadffmpeg-bb54c0ae719ef768463e78af82366e73c79e1e50.tar.gz
avfilter/af_afftdn: switch to activate
-rw-r--r--libavfilter/af_afftdn.c203
1 files changed, 116 insertions, 87 deletions
diff --git a/libavfilter/af_afftdn.c b/libavfilter/af_afftdn.c
index fbcb0f18d5..77e5406d1d 100644
--- a/libavfilter/af_afftdn.c
+++ b/libavfilter/af_afftdn.c
@@ -28,6 +28,7 @@
#include "avfilter.h"
#include "audio.h"
#include "formats.h"
+#include "filters.h"
#define C (M_LN10 * 0.1)
#define RATIO 0.98
@@ -1153,7 +1154,7 @@ static void get_auto_noise_levels(AudioFFTDeNoiseContext *s,
}
}
-static int filter_frame(AVFilterLink *inlink, AVFrame *frame)
+static int output_frame(AVFilterLink *inlink)
{
AVFilterContext *ctx = inlink->dst;
AVFilterLink *outlink = ctx->outputs[0];
@@ -1162,117 +1163,145 @@ static int filter_frame(AVFilterLink *inlink, AVFrame *frame)
ThreadData td;
int ret = 0;
- if (s->pts == AV_NOPTS_VALUE)
- s->pts = frame->pts;
+ if (!in) {
+ in = ff_get_audio_buffer(outlink, s->window_length);
+ if (!in)
+ return AVERROR(ENOMEM);
+ }
- ret = av_audio_fifo_write(s->fifo, (void **)frame->extended_data, frame->nb_samples);
- av_frame_free(&frame);
+ ret = av_audio_fifo_peek(s->fifo, (void **)in->extended_data, s->window_length);
if (ret < 0)
return ret;
- while (av_audio_fifo_size(s->fifo) >= s->window_length) {
- if (!in) {
- in = ff_get_audio_buffer(outlink, s->window_length);
- if (!in)
- return AVERROR(ENOMEM);
- }
+ if (s->track_noise) {
+ for (int ch = 0; ch < inlink->channels; ch++) {
+ DeNoiseChannel *dnch = &s->dnch[ch];
+ double levels[15];
- ret = av_audio_fifo_peek(s->fifo, (void **)in->extended_data, s->window_length);
- if (ret < 0)
- break;
+ get_auto_noise_levels(s, dnch, levels);
+ set_noise_profile(s, dnch, levels, 0);
+ }
- if (s->track_noise) {
- for (int ch = 0; ch < inlink->channels; ch++) {
- DeNoiseChannel *dnch = &s->dnch[ch];
- double levels[15];
+ if (s->noise_floor != s->last_noise_floor)
+ set_parameters(s);
+ }
- get_auto_noise_levels(s, dnch, levels);
- set_noise_profile(s, dnch, levels, 0);
- }
+ if (s->sample_noise_start) {
+ for (int ch = 0; ch < inlink->channels; ch++) {
+ DeNoiseChannel *dnch = &s->dnch[ch];
- if (s->noise_floor != s->last_noise_floor)
- set_parameters(s);
+ init_sample_noise(dnch);
}
+ s->sample_noise_start = 0;
+ s->sample_noise = 1;
+ }
- if (s->sample_noise_start) {
- for (int ch = 0; ch < inlink->channels; ch++) {
- DeNoiseChannel *dnch = &s->dnch[ch];
+ if (s->sample_noise) {
+ for (int ch = 0; ch < inlink->channels; ch++) {
+ DeNoiseChannel *dnch = &s->dnch[ch];
- init_sample_noise(dnch);
- }
- s->sample_noise_start = 0;
- s->sample_noise = 1;
+ sample_noise_block(s, dnch, in, ch);
}
+ }
- if (s->sample_noise) {
- for (int ch = 0; ch < inlink->channels; ch++) {
- DeNoiseChannel *dnch = &s->dnch[ch];
+ if (s->sample_noise_end) {
+ for (int ch = 0; ch < inlink->channels; ch++) {
+ DeNoiseChannel *dnch = &s->dnch[ch];
+ double sample_noise[15];
- sample_noise_block(s, dnch, in, ch);
- }
+ finish_sample_noise(s, dnch, sample_noise);
+ set_noise_profile(s, dnch, sample_noise, 1);
+ set_band_parameters(s, dnch);
}
+ s->sample_noise = 0;
+ s->sample_noise_end = 0;
+ }
- if (s->sample_noise_end) {
- for (int ch = 0; ch < inlink->channels; ch++) {
- DeNoiseChannel *dnch = &s->dnch[ch];
- double sample_noise[15];
-
- finish_sample_noise(s, dnch, sample_noise);
- set_noise_profile(s, dnch, sample_noise, 1);
- set_band_parameters(s, dnch);
- }
- s->sample_noise = 0;
- s->sample_noise_end = 0;
- }
+ s->block_count++;
+ td.in = in;
+ ctx->internal->execute(ctx, filter_channel, &td, NULL,
+ FFMIN(outlink->channels, ff_filter_get_nb_threads(ctx)));
- s->block_count++;
- td.in = in;
- ctx->internal->execute(ctx, filter_channel, &td, NULL,
- FFMIN(outlink->channels, ff_filter_get_nb_threads(ctx)));
+ out = ff_get_audio_buffer(outlink, s->sample_advance);
+ if (!out) {
+ ret = AVERROR(ENOMEM);
+ goto end;
+ }
- out = ff_get_audio_buffer(outlink, s->sample_advance);
- if (!out) {
- ret = AVERROR(ENOMEM);
+ for (int ch = 0; ch < inlink->channels; ch++) {
+ DeNoiseChannel *dnch = &s->dnch[ch];
+ double *src = dnch->out_samples;
+ float *orig = (float *)in->extended_data[ch];
+ float *dst = (float *)out->extended_data[ch];
+
+ switch (s->output_mode) {
+ case IN_MODE:
+ for (int m = 0; m < s->sample_advance; m++)
+ dst[m] = orig[m];
+ break;
+ case OUT_MODE:
+ for (int m = 0; m < s->sample_advance; m++)
+ dst[m] = src[m];
+ break;
+ case NOISE_MODE:
+ for (int m = 0; m < s->sample_advance; m++)
+ dst[m] = orig[m] - src[m];
break;
+ default:
+ return AVERROR_BUG;
}
+ memmove(src, src + s->sample_advance, (s->window_length - s->sample_advance) * sizeof(*src));
+ memset(src + (s->window_length - s->sample_advance), 0, s->sample_advance * sizeof(*src));
+ }
- for (int ch = 0; ch < inlink->channels; ch++) {
- DeNoiseChannel *dnch = &s->dnch[ch];
- double *src = dnch->out_samples;
- float *orig = (float *)in->extended_data[ch];
- float *dst = (float *)out->extended_data[ch];
-
- switch (s->output_mode) {
- case IN_MODE:
- for (int m = 0; m < s->sample_advance; m++)
- dst[m] = orig[m];
- break;
- case OUT_MODE:
- for (int m = 0; m < s->sample_advance; m++)
- dst[m] = src[m];
- break;
- case NOISE_MODE:
- for (int m = 0; m < s->sample_advance; m++)
- dst[m] = orig[m] - src[m];
- break;
- default:
- return AVERROR_BUG;
- }
- memmove(src, src + s->sample_advance, (s->window_length - s->sample_advance) * sizeof(*src));
- memset(src + (s->window_length - s->sample_advance), 0, s->sample_advance * sizeof(*src));
- }
+ av_audio_fifo_drain(s->fifo, s->sample_advance);
+
+ out->pts = s->pts;
+ ret = ff_filter_frame(outlink, out);
+ if (ret < 0)
+ goto end;
+ s->pts += s->sample_advance;
+end:
+ av_frame_free(&in);
- av_audio_fifo_drain(s->fifo, s->sample_advance);
+ return ret;
+}
- out->pts = s->pts;
- ret = ff_filter_frame(outlink, out);
+static int activate(AVFilterContext *ctx)
+{
+ AVFilterLink *inlink = ctx->inputs[0];
+ AVFilterLink *outlink = ctx->outputs[0];
+ AudioFFTDeNoiseContext *s = ctx->priv;
+ AVFrame *frame = NULL;
+ int ret;
+
+ FF_FILTER_FORWARD_STATUS_BACK(outlink, inlink);
+
+ ret = ff_inlink_consume_frame(inlink, &frame);
+ if (ret < 0)
+ return ret;
+
+ if (ret > 0) {
+ if (s->pts == AV_NOPTS_VALUE)
+ s->pts = frame->pts;
+
+ ret = av_audio_fifo_write(s->fifo, (void **)frame->extended_data, frame->nb_samples);
+ av_frame_free(&frame);
if (ret < 0)
- break;
- s->pts += s->sample_advance;
+ return ret;
}
- av_frame_free(&in);
- return ret;
+ if (av_audio_fifo_size(s->fifo) >= s->window_length)
+ return output_frame(inlink);
+
+ FF_FILTER_FORWARD_STATUS(inlink, outlink);
+ if (ff_outlink_frame_wanted(outlink) &&
+ av_audio_fifo_size(s->fifo) < s->window_length) {
+ ff_inlink_request_frame(inlink);
+ return 0;
+ }
+
+ return FFERROR_NOT_READY;
}
static av_cold void uninit(AVFilterContext *ctx)
@@ -1393,7 +1422,6 @@ static const AVFilterPad inputs[] = {
{
.name = "default",
.type = AVMEDIA_TYPE_AUDIO,
- .filter_frame = filter_frame,
.config_props = config_input,
},
{ NULL }
@@ -1413,6 +1441,7 @@ AVFilter ff_af_afftdn = {
.query_formats = query_formats,
.priv_size = sizeof(AudioFFTDeNoiseContext),
.priv_class = &afftdn_class,
+ .activate = activate,
.uninit = uninit,
.inputs = inputs,
.outputs = outputs,