1
0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2025-01-13 21:28:01 +02:00

avfilter/af_speechnorm: speed up filtering code

Reduce some asserts by default.
This commit is contained in:
Paul B Mahol 2022-02-16 19:31:31 +01:00
parent 52218c0322
commit 698de27f25

View File

@ -128,7 +128,7 @@ static int get_pi_samples(PeriodItem *pi, int start, int end, int remain)
start = 0;
if (pi[start].type == 0)
break;
av_assert0(pi[start].size > 0);
av_assert1(pi[start].size > 0);
sum += pi[start].size;
}
@ -156,7 +156,7 @@ static void consume_pi(ChannelContext *cc, int nb_samples)
if (cc->pi_size >= nb_samples) {
cc->pi_size -= nb_samples;
} else {
av_assert0(0);
av_assert1(0);
}
}
@ -178,16 +178,16 @@ static double next_gain(AVFilterContext *ctx, double pi_max_peak, int bypass, do
static void next_pi(AVFilterContext *ctx, ChannelContext *cc, int bypass)
{
av_assert0(cc->pi_size >= 0);
av_assert1(cc->pi_size >= 0);
if (cc->pi_size == 0) {
SpeechNormalizerContext *s = ctx->priv;
int start = cc->pi_start;
av_assert0(cc->pi[start].size > 0);
av_assert1(cc->pi[start].size > 0);
av_assert0(cc->pi[start].type > 0 || s->eof);
cc->pi_size = cc->pi[start].size;
cc->pi_max_peak = cc->pi[start].max_peak;
av_assert0(cc->pi_start != cc->pi_end || s->eof);
av_assert1(cc->pi_start != cc->pi_end || s->eof);
start++;
if (start >= MAX_ITEMS)
start = 0;
@ -219,58 +219,71 @@ static double min_gain(AVFilterContext *ctx, ChannelContext *cc, int max_size)
return min_gain;
}
#define ANALYZE_CHANNEL(name, ptype, zero, min_peak) \
static void analyze_channel_## name (AVFilterContext *ctx, ChannelContext *cc, \
const uint8_t *srcp, int nb_samples) \
{ \
SpeechNormalizerContext *s = ctx->priv; \
const ptype *src = (const ptype *)srcp; \
int n = 0; \
\
if (cc->state < 0) \
cc->state = src[0] >= zero; \
\
while (n < nb_samples) { \
if ((cc->state != (src[n] >= zero)) || \
(cc->pi[cc->pi_end].size > s->max_period)) { \
ptype max_peak = cc->pi[cc->pi_end].max_peak; \
int state = cc->state; \
cc->state = src[n] >= zero; \
av_assert0(cc->pi[cc->pi_end].size > 0); \
if (max_peak >= min_peak || \
cc->pi[cc->pi_end].size > s->max_period) { \
cc->pi[cc->pi_end].type = 1; \
cc->pi_end++; \
if (cc->pi_end >= MAX_ITEMS) \
cc->pi_end = 0; \
if (cc->state != state) \
cc->pi[cc->pi_end].max_peak = DBL_MIN; \
else \
cc->pi[cc->pi_end].max_peak = max_peak; \
cc->pi[cc->pi_end].type = 0; \
cc->pi[cc->pi_end].size = 0; \
av_assert0(cc->pi_end != cc->pi_start); \
} \
} \
\
if (cc->state) { \
while (src[n] >= zero) { \
cc->pi[cc->pi_end].max_peak = FFMAX(cc->pi[cc->pi_end].max_peak, src[n]); \
cc->pi[cc->pi_end].size++; \
n++; \
if (n >= nb_samples) \
break; \
} \
} else { \
while (src[n] < zero) { \
cc->pi[cc->pi_end].max_peak = FFMAX(cc->pi[cc->pi_end].max_peak, -src[n]); \
cc->pi[cc->pi_end].size++; \
n++; \
if (n >= nb_samples) \
break; \
} \
} \
} \
#define ANALYZE_CHANNEL(name, ptype, zero, min_peak) \
static void analyze_channel_## name (AVFilterContext *ctx, ChannelContext *cc, \
const uint8_t *srcp, int nb_samples) \
{ \
SpeechNormalizerContext *s = ctx->priv; \
const ptype *src = (const ptype *)srcp; \
const int max_period = s->max_period; \
PeriodItem *pi = (PeriodItem *)&cc->pi; \
int pi_end = cc->pi_end; \
int n = 0; \
\
if (cc->state < 0) \
cc->state = src[0] >= zero; \
\
while (n < nb_samples) { \
ptype new_max_peak; \
int new_size; \
\
if ((cc->state != (src[n] >= zero)) || \
(pi[pi_end].size > max_period)) { \
ptype max_peak = pi[pi_end].max_peak; \
int state = cc->state; \
\
cc->state = src[n] >= zero; \
av_assert1(pi[pi_end].size > 0); \
if (max_peak >= min_peak || \
pi[pi_end].size > max_period) { \
pi[pi_end].type = 1; \
pi_end++; \
if (pi_end >= MAX_ITEMS) \
pi_end = 0; \
if (cc->state != state) \
pi[pi_end].max_peak = DBL_MIN; \
else \
pi[pi_end].max_peak = max_peak; \
pi[pi_end].type = 0; \
pi[pi_end].size = 0; \
av_assert1(pi_end != cc->pi_start); \
} \
} \
\
new_max_peak = pi[pi_end].max_peak; \
new_size = pi[pi_end].size; \
if (cc->state) { \
while (src[n] >= zero) { \
new_max_peak = FFMAX(new_max_peak, src[n]); \
new_size++; \
n++; \
if (n >= nb_samples) \
break; \
} \
} else { \
while (src[n] < zero) { \
new_max_peak = FFMAX(new_max_peak, -src[n]); \
new_size++; \
n++; \
if (n >= nb_samples) \
break; \
} \
} \
\
pi[pi_end].max_peak = new_max_peak; \
pi[pi_end].size = new_size; \
} \
cc->pi_end = pi_end; \
}
ANALYZE_CHANNEL(dbl, double, 0.0, MIN_PEAK)
@ -296,7 +309,7 @@ static void filter_channels_## name (AVFilterContext *ctx,
\
next_pi(ctx, cc, bypass); \
size = FFMIN(nb_samples - n, cc->pi_size); \
av_assert0(size > 0); \
av_assert1(size > 0); \
gain = cc->gain_state; \
consume_pi(cc, size); \
for (int i = n; !ctx->is_disabled && i < n + size; i++) \
@ -343,7 +356,7 @@ static void filter_link_channels_## name (AVFilterContext *ctx,
max_size = FFMAX(max_size, cc->pi_size); \
} \
\
av_assert0(min_size > 0); \
av_assert1(min_size > 0); \
for (int ch = 0; ch < inlink->channels; ch++) { \
ChannelContext *cc = &s->cc[ch]; \
\
@ -509,7 +522,7 @@ static int config_input(AVFilterLink *inlink)
s->filter_channels[1] = filter_link_channels_dbl;
break;
default:
av_assert0(0);
av_assert1(0);
}
return 0;