You've already forked FFmpeg
mirror of
https://github.com/FFmpeg/FFmpeg.git
synced 2025-08-10 06:10:52 +02:00
avfilter/af_dynaudnorm: use fmin/fmax for doubles
This commit is contained in:
@@ -385,13 +385,13 @@ static double find_peak_magnitude(AVFrame *frame, int channel)
|
|||||||
double *data_ptr = (double *)frame->extended_data[c];
|
double *data_ptr = (double *)frame->extended_data[c];
|
||||||
|
|
||||||
for (i = 0; i < frame->nb_samples; i++)
|
for (i = 0; i < frame->nb_samples; i++)
|
||||||
max = FFMAX(max, fabs(data_ptr[i]));
|
max = fmax(max, fabs(data_ptr[i]));
|
||||||
}
|
}
|
||||||
} else {
|
} else {
|
||||||
double *data_ptr = (double *)frame->extended_data[channel];
|
double *data_ptr = (double *)frame->extended_data[channel];
|
||||||
|
|
||||||
for (i = 0; i < frame->nb_samples; i++)
|
for (i = 0; i < frame->nb_samples; i++)
|
||||||
max = FFMAX(max, fabs(data_ptr[i]));
|
max = fmax(max, fabs(data_ptr[i]));
|
||||||
}
|
}
|
||||||
|
|
||||||
return max;
|
return max;
|
||||||
@@ -421,7 +421,7 @@ static double compute_frame_rms(AVFrame *frame, int channel)
|
|||||||
rms_value /= frame->nb_samples;
|
rms_value /= frame->nb_samples;
|
||||||
}
|
}
|
||||||
|
|
||||||
return FFMAX(sqrt(rms_value), DBL_EPSILON);
|
return fmax(sqrt(rms_value), DBL_EPSILON);
|
||||||
}
|
}
|
||||||
|
|
||||||
static local_gain get_max_local_gain(DynamicAudioNormalizerContext *s, AVFrame *frame,
|
static local_gain get_max_local_gain(DynamicAudioNormalizerContext *s, AVFrame *frame,
|
||||||
@@ -433,7 +433,7 @@ static local_gain get_max_local_gain(DynamicAudioNormalizerContext *s, AVFrame *
|
|||||||
local_gain gain;
|
local_gain gain;
|
||||||
|
|
||||||
gain.threshold = peak_magnitude > s->threshold;
|
gain.threshold = peak_magnitude > s->threshold;
|
||||||
gain.max_gain = bound(s->max_amplification, FFMIN(maximum_gain, rms_gain));
|
gain.max_gain = bound(s->max_amplification, fmin(maximum_gain, rms_gain));
|
||||||
|
|
||||||
return gain;
|
return gain;
|
||||||
}
|
}
|
||||||
@@ -444,7 +444,7 @@ static double minimum_filter(cqueue *q)
|
|||||||
int i;
|
int i;
|
||||||
|
|
||||||
for (i = 0; i < cqueue_size(q); i++) {
|
for (i = 0; i < cqueue_size(q); i++) {
|
||||||
min = FFMIN(min, cqueue_peek(q, i));
|
min = fmin(min, cqueue_peek(q, i));
|
||||||
}
|
}
|
||||||
|
|
||||||
return min;
|
return min;
|
||||||
@@ -475,7 +475,7 @@ static void update_gain_history(DynamicAudioNormalizerContext *s, int channel,
|
|||||||
{
|
{
|
||||||
if (cqueue_empty(s->gain_history_original[channel])) {
|
if (cqueue_empty(s->gain_history_original[channel])) {
|
||||||
const int pre_fill_size = s->filter_size / 2;
|
const int pre_fill_size = s->filter_size / 2;
|
||||||
const double initial_value = s->alt_boundary_mode ? gain.max_gain : FFMIN(1.0, gain.max_gain);
|
const double initial_value = s->alt_boundary_mode ? gain.max_gain : fmin(1.0, gain.max_gain);
|
||||||
|
|
||||||
s->prev_amplification_factor[channel] = initial_value;
|
s->prev_amplification_factor[channel] = initial_value;
|
||||||
|
|
||||||
@@ -497,7 +497,7 @@ static void update_gain_history(DynamicAudioNormalizerContext *s, int channel,
|
|||||||
|
|
||||||
while (cqueue_size(s->gain_history_minimum[channel]) < pre_fill_size) {
|
while (cqueue_size(s->gain_history_minimum[channel]) < pre_fill_size) {
|
||||||
input++;
|
input++;
|
||||||
initial_value = FFMIN(initial_value, cqueue_peek(s->gain_history_original[channel], input));
|
initial_value = fmin(initial_value, cqueue_peek(s->gain_history_original[channel], input));
|
||||||
cqueue_enqueue(s->gain_history_minimum[channel], initial_value);
|
cqueue_enqueue(s->gain_history_minimum[channel], initial_value);
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
@@ -516,7 +516,7 @@ static void update_gain_history(DynamicAudioNormalizerContext *s, int channel,
|
|||||||
|
|
||||||
smoothed = gaussian_filter(s, s->gain_history_minimum[channel], s->threshold_history[channel]);
|
smoothed = gaussian_filter(s, s->gain_history_minimum[channel], s->threshold_history[channel]);
|
||||||
limit = cqueue_peek(s->gain_history_original[channel], 0);
|
limit = cqueue_peek(s->gain_history_original[channel], 0);
|
||||||
smoothed = FFMIN(smoothed, limit);
|
smoothed = fmin(smoothed, limit);
|
||||||
|
|
||||||
cqueue_enqueue(s->gain_history_smoothed[channel], smoothed);
|
cqueue_enqueue(s->gain_history_smoothed[channel], smoothed);
|
||||||
|
|
||||||
@@ -606,7 +606,7 @@ static double compute_frame_std_dev(DynamicAudioNormalizerContext *s,
|
|||||||
variance /= frame->nb_samples - 1;
|
variance /= frame->nb_samples - 1;
|
||||||
}
|
}
|
||||||
|
|
||||||
return FFMAX(sqrt(variance), DBL_EPSILON);
|
return fmax(sqrt(variance), DBL_EPSILON);
|
||||||
}
|
}
|
||||||
|
|
||||||
static void perform_compression(DynamicAudioNormalizerContext *s, AVFrame *frame)
|
static void perform_compression(DynamicAudioNormalizerContext *s, AVFrame *frame)
|
||||||
@@ -616,7 +616,7 @@ static void perform_compression(DynamicAudioNormalizerContext *s, AVFrame *frame
|
|||||||
|
|
||||||
if (s->channels_coupled) {
|
if (s->channels_coupled) {
|
||||||
const double standard_deviation = compute_frame_std_dev(s, frame, -1);
|
const double standard_deviation = compute_frame_std_dev(s, frame, -1);
|
||||||
const double current_threshold = FFMIN(1.0, s->compress_factor * standard_deviation);
|
const double current_threshold = fmin(1.0, s->compress_factor * standard_deviation);
|
||||||
|
|
||||||
const double prev_value = is_first_frame ? current_threshold : s->compress_threshold[0];
|
const double prev_value = is_first_frame ? current_threshold : s->compress_threshold[0];
|
||||||
double prev_actual_thresh, curr_actual_thresh;
|
double prev_actual_thresh, curr_actual_thresh;
|
||||||
@@ -641,7 +641,7 @@ static void perform_compression(DynamicAudioNormalizerContext *s, AVFrame *frame
|
|||||||
for (c = 0; c < s->channels; c++) {
|
for (c = 0; c < s->channels; c++) {
|
||||||
const int bypass = bypass_channel(s, frame, c);
|
const int bypass = bypass_channel(s, frame, c);
|
||||||
const double standard_deviation = compute_frame_std_dev(s, frame, c);
|
const double standard_deviation = compute_frame_std_dev(s, frame, c);
|
||||||
const double current_threshold = setup_compress_thresh(FFMIN(1.0, s->compress_factor * standard_deviation));
|
const double current_threshold = setup_compress_thresh(fmin(1.0, s->compress_factor * standard_deviation));
|
||||||
const double prev_value = is_first_frame ? current_threshold : s->compress_threshold[c];
|
const double prev_value = is_first_frame ? current_threshold : s->compress_threshold[c];
|
||||||
double prev_actual_thresh, curr_actual_thresh;
|
double prev_actual_thresh, curr_actual_thresh;
|
||||||
double *dst_ptr;
|
double *dst_ptr;
|
||||||
@@ -820,7 +820,7 @@ static int flush_buffer(DynamicAudioNormalizerContext *s, AVFilterLink *inlink,
|
|||||||
double *dst_ptr = (double *)out->extended_data[c];
|
double *dst_ptr = (double *)out->extended_data[c];
|
||||||
|
|
||||||
for (i = 0; i < out->nb_samples; i++) {
|
for (i = 0; i < out->nb_samples; i++) {
|
||||||
dst_ptr[i] = s->alt_boundary_mode ? DBL_EPSILON : ((s->target_rms > DBL_EPSILON) ? FFMIN(s->peak_value, s->target_rms) : s->peak_value);
|
dst_ptr[i] = s->alt_boundary_mode ? DBL_EPSILON : ((s->target_rms > DBL_EPSILON) ? fmin(s->peak_value, s->target_rms) : s->peak_value);
|
||||||
if (s->dc_correction) {
|
if (s->dc_correction) {
|
||||||
dst_ptr[i] *= ((i % 2) == 1) ? -1 : 1;
|
dst_ptr[i] *= ((i % 2) == 1) ? -1 : 1;
|
||||||
dst_ptr[i] += s->dc_correction_value[c];
|
dst_ptr[i] += s->dc_correction_value[c];
|
||||||
|
Reference in New Issue
Block a user