1
0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2024-11-21 10:55:51 +02:00

avfilter/avf_showcwt: always consume some available samples

Stop waiting for big cache of samples.
This commit is contained in:
Paul B Mahol 2022-11-29 00:27:14 +01:00
parent 39dc9a7421
commit 634e29f3c0

View File

@ -73,7 +73,6 @@ typedef struct ShowCWTContext {
int ifft_in_size;
int ifft_out_size;
int pos;
int in_nb_samples;
int64_t in_pts;
int64_t old_pts;
float *frequency_band;
@ -95,6 +94,7 @@ typedef struct ShowCWTContext {
int slide;
int direction;
int hop_size;
int hop_index;
int ihop_size;
int ihop_index;
int input_padding_size;
@ -250,7 +250,7 @@ static float remap_log(float value, float log_factor)
return 1.f - av_clipf(value, 0.f, 1.f);
}
static int run_channel_cwt_prepare(AVFilterContext *ctx, void *arg, int ch)
static int run_channel_cwt_prepare(AVFilterContext *ctx, void *arg, int ch, int eof)
{
ShowCWTContext *s = ctx->priv;
AVFrame *fin = arg;
@ -260,14 +260,19 @@ static int run_channel_cwt_prepare(AVFilterContext *ctx, void *arg, int ch)
AVComplexFloat *dst = (AVComplexFloat *)s->fft_out->extended_data[ch];
const int nb_consumed_samples = s->nb_consumed_samples;
const int input_padding_size = s->input_padding_size;
const int hop_size = s->hop_size;
const int hop_size = eof ? s->hop_size : fin->nb_samples;
const int offset = input_padding_size - hop_size;
memmove(overlap, &overlap[hop_size], offset * sizeof(float));
memcpy(&overlap[offset], input,
fin->nb_samples * sizeof(float));
memset(&overlap[offset + fin->nb_samples], 0,
(hop_size - fin->nb_samples) * sizeof(float));
if (eof) {
memset(&overlap[offset + fin->nb_samples], 0,
(hop_size - fin->nb_samples) * sizeof(float));
} else if (s->hop_index + fin->nb_samples < s->hop_size) {
return 0;
}
for (int n = 0; n < nb_consumed_samples; n++) {
src[n].re = overlap[n];
@ -665,22 +670,26 @@ static int activate(AVFilterContext *ctx)
AVFrame *fin;
if (s->ihop_index == 0) {
ret = ff_inlink_consume_samples(inlink, s->hop_size, s->hop_size, &fin);
ret = ff_inlink_consume_samples(inlink, 1, s->hop_size - s->hop_index, &fin);
if (ret < 0)
return ret;
if (ret > 0) {
for (int ch = 0; ch < s->nb_channels; ch++)
run_channel_cwt_prepare(ctx, fin, ch);
run_channel_cwt_prepare(ctx, fin, ch,
ff_outlink_get_status(inlink));
s->in_pts = fin->pts;
s->in_nb_samples = fin->nb_samples;
if (s->hop_index == 0)
s->in_pts = fin->pts;
s->hop_index += fin->nb_samples;
av_frame_free(&fin);
}
}
if (ret > 0 || s->ihop_index > 0) {
if (s->hop_index >= s->hop_size || s->ihop_index > 0) {
int64_t pts_offset;
s->hop_index = 0;
switch (s->slide) {
case SLIDE_SCROLL:
switch (s->direction) {
@ -717,7 +726,7 @@ static int activate(AVFilterContext *ctx)
ff_filter_execute(ctx, draw, NULL, NULL, s->nb_threads);
pts_offset = av_rescale_q(s->ihop_index, av_make_q(1, s->ihop_size), av_make_q(1, s->in_nb_samples));
pts_offset = av_rescale(s->ihop_index, s->hop_size, s->ihop_size);
s->outpicref->pts = av_rescale_q(s->in_pts + pts_offset, inlink->time_base, outlink->time_base);
s->ihop_index++;
@ -795,7 +804,7 @@ fail:
}
}
if (ff_inlink_queued_samples(inlink) >= s->hop_size || s->ihop_index) {
if (ff_inlink_queued_samples(inlink) > 0 || s->ihop_index || s->hop_index >= s->hop_size) {
ff_filter_set_ready(ctx, 10);
return 0;
}