Signed-off-by: Paul B Mahol <onemda@gmail.com>
Paul B Mahol authored on 2013/04/22 21:38:24... | ... |
@@ -990,6 +990,51 @@ the data is treated as if all the planes were concatenated. |
990 | 990 |
A list of Adler-32 checksums for each data plane. |
991 | 991 |
@end table |
992 | 992 |
|
993 |
+@section astats |
|
994 |
+ |
|
995 |
+Display time domain statistical information about the audio channels. |
|
996 |
+Statistics are calculated and displayed for each audio channel and, |
|
997 |
+where applicable, an overall figure is also given. |
|
998 |
+ |
|
999 |
+The filter accepts the following option: |
|
1000 |
+@table @option |
|
1001 |
+@item length |
|
1002 |
+Short window length in seconds, used for peak and trough RMS measurement. |
|
1003 |
+Default is @code{0.05} (50 miliseconds). Allowed range is @code{[0.1 - 10]}. |
|
1004 |
+@end table |
|
1005 |
+ |
|
1006 |
+A description of each shown parameter follows: |
|
1007 |
+ |
|
1008 |
+@table @option |
|
1009 |
+@item DC offset |
|
1010 |
+Mean amplitude displacement from zero. |
|
1011 |
+ |
|
1012 |
+@item Min level |
|
1013 |
+Minimal sample level. |
|
1014 |
+ |
|
1015 |
+@item Max level |
|
1016 |
+Maximal sample level. |
|
1017 |
+ |
|
1018 |
+@item Peak level dB |
|
1019 |
+@item RMS level dB |
|
1020 |
+Standard peak and RMS level measured in dBFS. |
|
1021 |
+ |
|
1022 |
+@item RMS peak dB |
|
1023 |
+@item RMS trough dB |
|
1024 |
+Peak and trough values for RMS level measured over a short window. |
|
1025 |
+ |
|
1026 |
+@item Crest factor |
|
1027 |
+Standard ratio of peak to RMS level (note: not in dB). |
|
1028 |
+ |
|
1029 |
+@item Flat factor |
|
1030 |
+Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels |
|
1031 |
+(i.e. either @var{Min level} or @var{Max level}). |
|
1032 |
+ |
|
1033 |
+@item Peak count |
|
1034 |
+Number of occasions (not the number of samples) that the signal attained either |
|
1035 |
+@var{Min level} or @var{Max level}. |
|
1036 |
+@end table |
|
1037 |
+ |
|
993 | 1038 |
@section astreamsync |
994 | 1039 |
|
995 | 1040 |
Forward two audio streams and control the order the buffers are forwarded. |
... | ... |
@@ -69,6 +69,7 @@ OBJS-$(CONFIG_ASETRATE_FILTER) += af_asetrate.o |
69 | 69 |
OBJS-$(CONFIG_ASETTB_FILTER) += f_settb.o |
70 | 70 |
OBJS-$(CONFIG_ASHOWINFO_FILTER) += af_ashowinfo.o |
71 | 71 |
OBJS-$(CONFIG_ASPLIT_FILTER) += split.o |
72 |
+OBJS-$(CONFIG_ASTATS_FILTER) += af_astats.o |
|
72 | 73 |
OBJS-$(CONFIG_ASTREAMSYNC_FILTER) += af_astreamsync.o |
73 | 74 |
OBJS-$(CONFIG_ASYNCTS_FILTER) += af_asyncts.o |
74 | 75 |
OBJS-$(CONFIG_ATEMPO_FILTER) += af_atempo.o |
75 | 76 |
new file mode 100644 |
... | ... |
@@ -0,0 +1,274 @@ |
0 |
+/* |
|
1 |
+ * Copyright (c) 2009 Rob Sykes <robs@users.sourceforge.net> |
|
2 |
+ * Copyright (c) 2013 Paul B Mahol |
|
3 |
+ * |
|
4 |
+ * This file is part of FFmpeg. |
|
5 |
+ * |
|
6 |
+ * FFmpeg is free software; you can redistribute it and/or |
|
7 |
+ * modify it under the terms of the GNU Lesser General Public |
|
8 |
+ * License as published by the Free Software Foundation; either |
|
9 |
+ * version 2.1 of the License, or (at your option) any later version. |
|
10 |
+ * |
|
11 |
+ * FFmpeg is distributed in the hope that it will be useful, |
|
12 |
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
13 |
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
|
14 |
+ * Lesser General Public License for more details. |
|
15 |
+ * |
|
16 |
+ * You should have received a copy of the GNU Lesser General Public |
|
17 |
+ * License along with FFmpeg; if not, write to the Free Software |
|
18 |
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
19 |
+ */ |
|
20 |
+ |
|
21 |
+#include <float.h> |
|
22 |
+ |
|
23 |
+#include "libavutil/opt.h" |
|
24 |
+#include "audio.h" |
|
25 |
+#include "avfilter.h" |
|
26 |
+#include "internal.h" |
|
27 |
+ |
|
28 |
+typedef struct ChannelStats { |
|
29 |
+ double last; |
|
30 |
+ double sigma_x, sigma_x2; |
|
31 |
+ double avg_sigma_x2, min_sigma_x2, max_sigma_x2; |
|
32 |
+ double min, max; |
|
33 |
+ double min_run, max_run; |
|
34 |
+ double min_runs, max_runs; |
|
35 |
+ uint64_t min_count, max_count; |
|
36 |
+ uint64_t nb_samples; |
|
37 |
+} ChannelStats; |
|
38 |
+ |
|
39 |
+typedef struct { |
|
40 |
+ const AVClass *class; |
|
41 |
+ ChannelStats *chstats; |
|
42 |
+ int nb_channels; |
|
43 |
+ uint64_t tc_samples; |
|
44 |
+ double time_constant; |
|
45 |
+ double mult; |
|
46 |
+} AudioStatsContext; |
|
47 |
+ |
|
48 |
+#define OFFSET(x) offsetof(AudioStatsContext, x) |
|
49 |
+#define FLAGS AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_FILTERING_PARAM |
|
50 |
+ |
|
51 |
+static const AVOption astats_options[] = { |
|
52 |
+ { "length", "set the window length", OFFSET(time_constant), AV_OPT_TYPE_DOUBLE, {.dbl=.05}, .01, 10, FLAGS }, |
|
53 |
+ {NULL}, |
|
54 |
+}; |
|
55 |
+ |
|
56 |
+AVFILTER_DEFINE_CLASS(astats); |
|
57 |
+ |
|
58 |
+static int query_formats(AVFilterContext *ctx) |
|
59 |
+{ |
|
60 |
+ AVFilterFormats *formats; |
|
61 |
+ AVFilterChannelLayouts *layouts; |
|
62 |
+ static const enum AVSampleFormat sample_fmts[] = { |
|
63 |
+ AV_SAMPLE_FMT_DBL, AV_SAMPLE_FMT_DBLP, |
|
64 |
+ AV_SAMPLE_FMT_NONE |
|
65 |
+ }; |
|
66 |
+ |
|
67 |
+ layouts = ff_all_channel_layouts(); |
|
68 |
+ if (!layouts) |
|
69 |
+ return AVERROR(ENOMEM); |
|
70 |
+ ff_set_common_channel_layouts(ctx, layouts); |
|
71 |
+ |
|
72 |
+ formats = ff_make_format_list(sample_fmts); |
|
73 |
+ if (!formats) |
|
74 |
+ return AVERROR(ENOMEM); |
|
75 |
+ ff_set_common_formats(ctx, formats); |
|
76 |
+ |
|
77 |
+ formats = ff_all_samplerates(); |
|
78 |
+ if (!formats) |
|
79 |
+ return AVERROR(ENOMEM); |
|
80 |
+ ff_set_common_samplerates(ctx, formats); |
|
81 |
+ |
|
82 |
+ return 0; |
|
83 |
+} |
|
84 |
+ |
|
85 |
+static int config_output(AVFilterLink *outlink) |
|
86 |
+{ |
|
87 |
+ AudioStatsContext *s = outlink->src->priv; |
|
88 |
+ int c; |
|
89 |
+ |
|
90 |
+ s->chstats = av_calloc(sizeof(*s->chstats), outlink->channels); |
|
91 |
+ if (!s->chstats) |
|
92 |
+ return AVERROR(ENOMEM); |
|
93 |
+ s->nb_channels = outlink->channels; |
|
94 |
+ s->mult = exp((-1 / s->time_constant / outlink->sample_rate)); |
|
95 |
+ s->tc_samples = 5 * s->time_constant * outlink->sample_rate + .5; |
|
96 |
+ |
|
97 |
+ for (c = 0; c < s->nb_channels; c++) { |
|
98 |
+ ChannelStats *p = &s->chstats[c]; |
|
99 |
+ |
|
100 |
+ p->min = p->min_sigma_x2 = DBL_MAX; |
|
101 |
+ p->max = p->max_sigma_x2 = DBL_MIN; |
|
102 |
+ } |
|
103 |
+ |
|
104 |
+ return 0; |
|
105 |
+} |
|
106 |
+ |
|
107 |
+static inline void stat(AudioStatsContext *s, ChannelStats *p, double d) |
|
108 |
+{ |
|
109 |
+ if (d < p->min) { |
|
110 |
+ p->min = d; |
|
111 |
+ p->min_run = 1; |
|
112 |
+ p->min_runs = 0; |
|
113 |
+ p->min_count = 1; |
|
114 |
+ } else if (d == p->min) { |
|
115 |
+ p->min_count++; |
|
116 |
+ p->min_run = d == p->last ? p->min_run + 1 : 1; |
|
117 |
+ } else if (p->last == p->min) { |
|
118 |
+ p->min_runs += p->min_run * p->min_run; |
|
119 |
+ } |
|
120 |
+ |
|
121 |
+ if (d > p->max) { |
|
122 |
+ p->max = d; |
|
123 |
+ p->max_run = 1; |
|
124 |
+ p->max_runs = 0; |
|
125 |
+ p->max_count = 1; |
|
126 |
+ } else if (d == p->max) { |
|
127 |
+ p->max_count++; |
|
128 |
+ p->max_run = d == p->last ? p->max_run + 1 : 1; |
|
129 |
+ } else if (p->last == p->max) { |
|
130 |
+ p->max_runs += p->max_run * p->max_run; |
|
131 |
+ } |
|
132 |
+ |
|
133 |
+ p->sigma_x += d; |
|
134 |
+ p->sigma_x2 += d * d; |
|
135 |
+ p->avg_sigma_x2 = p->avg_sigma_x2 * s->mult + (1.0 - s->mult) * d * d; |
|
136 |
+ p->last = d; |
|
137 |
+ |
|
138 |
+ if (p->nb_samples >= s->tc_samples) { |
|
139 |
+ p->max_sigma_x2 = FFMAX(p->max_sigma_x2, p->avg_sigma_x2); |
|
140 |
+ p->min_sigma_x2 = FFMIN(p->min_sigma_x2, p->avg_sigma_x2); |
|
141 |
+ } |
|
142 |
+ p->nb_samples++; |
|
143 |
+} |
|
144 |
+ |
|
145 |
+static int filter_frame(AVFilterLink *inlink, AVFrame *buf) |
|
146 |
+{ |
|
147 |
+ AudioStatsContext *s = inlink->dst->priv; |
|
148 |
+ const int channels = s->nb_channels; |
|
149 |
+ const double *src; |
|
150 |
+ int i, c; |
|
151 |
+ |
|
152 |
+ switch (inlink->format) { |
|
153 |
+ case AV_SAMPLE_FMT_DBLP: |
|
154 |
+ for (c = 0; c < channels; c++) { |
|
155 |
+ ChannelStats *p = &s->chstats[c]; |
|
156 |
+ src = (const double *)buf->extended_data[c]; |
|
157 |
+ |
|
158 |
+ for (i = 0; i < buf->nb_samples; i++, src++) |
|
159 |
+ stat(s, p, *src); |
|
160 |
+ } |
|
161 |
+ break; |
|
162 |
+ case AV_SAMPLE_FMT_DBL: |
|
163 |
+ src = (const double *)buf->extended_data[0]; |
|
164 |
+ |
|
165 |
+ for (i = 0; i < buf->nb_samples; i++) { |
|
166 |
+ for (c = 0; c < channels; c++, src++) |
|
167 |
+ stat(s, &s->chstats[c], *src); |
|
168 |
+ } |
|
169 |
+ break; |
|
170 |
+ } |
|
171 |
+ |
|
172 |
+ return ff_filter_frame(inlink->dst->outputs[0], buf); |
|
173 |
+} |
|
174 |
+ |
|
175 |
+#define LINEAR_TO_DB(x) (log10(x) * 20) |
|
176 |
+ |
|
177 |
+static void print_stats(AVFilterContext *ctx) |
|
178 |
+{ |
|
179 |
+ AudioStatsContext *s = ctx->priv; |
|
180 |
+ uint64_t min_count = 0, max_count = 0, nb_samples = 0; |
|
181 |
+ double min_runs = 0, max_runs = 0, |
|
182 |
+ min = DBL_MAX, max = DBL_MIN, |
|
183 |
+ max_sigma_x = 0, |
|
184 |
+ sigma_x = 0, |
|
185 |
+ sigma_x2 = 0, |
|
186 |
+ min_sigma_x2 = DBL_MAX, |
|
187 |
+ max_sigma_x2 = DBL_MIN; |
|
188 |
+ int c; |
|
189 |
+ |
|
190 |
+ for (c = 0; c < s->nb_channels; c++) { |
|
191 |
+ ChannelStats *p = &s->chstats[c]; |
|
192 |
+ |
|
193 |
+ if (p->nb_samples < s->tc_samples) |
|
194 |
+ p->min_sigma_x2 = p->max_sigma_x2 = p->sigma_x2 / p->nb_samples; |
|
195 |
+ |
|
196 |
+ min = FFMIN(min, p->min); |
|
197 |
+ max = FFMAX(max, p->max); |
|
198 |
+ min_sigma_x2 = FFMIN(min_sigma_x2, p->min_sigma_x2); |
|
199 |
+ max_sigma_x2 = FFMAX(max_sigma_x2, p->max_sigma_x2); |
|
200 |
+ sigma_x += p->sigma_x; |
|
201 |
+ sigma_x2 += p->sigma_x2; |
|
202 |
+ min_count += p->min_count; |
|
203 |
+ max_count += p->max_count; |
|
204 |
+ min_runs += p->min_runs; |
|
205 |
+ max_runs += p->max_runs; |
|
206 |
+ nb_samples += p->nb_samples; |
|
207 |
+ if (fabs(p->sigma_x) > fabs(max_sigma_x)) |
|
208 |
+ max_sigma_x = p->sigma_x; |
|
209 |
+ |
|
210 |
+ av_log(ctx, AV_LOG_INFO, "Channel: %d\n", c + 1); |
|
211 |
+ av_log(ctx, AV_LOG_INFO, "DC offset: %f\n", p->sigma_x / p->nb_samples); |
|
212 |
+ av_log(ctx, AV_LOG_INFO, "Min level: %f\n", p->min); |
|
213 |
+ av_log(ctx, AV_LOG_INFO, "Max level: %f\n", p->max); |
|
214 |
+ av_log(ctx, AV_LOG_INFO, "Peak level dB: %f\n", LINEAR_TO_DB(FFMAX(-p->min, p->max))); |
|
215 |
+ av_log(ctx, AV_LOG_INFO, "RMS level dB: %f\n", LINEAR_TO_DB(sqrt(p->sigma_x2 / p->nb_samples))); |
|
216 |
+ av_log(ctx, AV_LOG_INFO, "RMS peak dB: %f\n", LINEAR_TO_DB(sqrt(p->max_sigma_x2))); |
|
217 |
+ if (p->min_sigma_x2 != 1) |
|
218 |
+ av_log(ctx, AV_LOG_INFO, "RMS trough dB: %f\n",LINEAR_TO_DB(sqrt(p->min_sigma_x2))); |
|
219 |
+ av_log(ctx, AV_LOG_INFO, "Crest factor: %f\n", p->sigma_x2 ? FFMAX(-p->min, p->max) / sqrt(p->sigma_x2 / p->nb_samples) : 1); |
|
220 |
+ av_log(ctx, AV_LOG_INFO, "Flat factor: %f\n", LINEAR_TO_DB((p->min_runs + p->max_runs) / (p->min_count + p->max_count))); |
|
221 |
+ av_log(ctx, AV_LOG_INFO, "Peak count: %lld\n", p->min_count + p->max_count); |
|
222 |
+ } |
|
223 |
+ |
|
224 |
+ av_log(ctx, AV_LOG_INFO, "Overall\n"); |
|
225 |
+ av_log(ctx, AV_LOG_INFO, "DC offset: %f\n", max_sigma_x / (nb_samples / s->nb_channels)); |
|
226 |
+ av_log(ctx, AV_LOG_INFO, "Min level: %f\n", min); |
|
227 |
+ av_log(ctx, AV_LOG_INFO, "Max level: %f\n", max); |
|
228 |
+ av_log(ctx, AV_LOG_INFO, "Peak level dB: %f\n", LINEAR_TO_DB(FFMAX(-min, max))); |
|
229 |
+ av_log(ctx, AV_LOG_INFO, "RMS level dB: %f\n", LINEAR_TO_DB(sqrt(sigma_x2 / nb_samples))); |
|
230 |
+ av_log(ctx, AV_LOG_INFO, "RMS peak dB: %f\n", LINEAR_TO_DB(sqrt(max_sigma_x2))); |
|
231 |
+ if (min_sigma_x2 != 1) |
|
232 |
+ av_log(ctx, AV_LOG_INFO, "RMS trough dB: %f\n", LINEAR_TO_DB(sqrt(min_sigma_x2))); |
|
233 |
+ av_log(ctx, AV_LOG_INFO, "Flat factor: %f\n", LINEAR_TO_DB((min_runs + max_runs) / (min_count + max_count))); |
|
234 |
+ av_log(ctx, AV_LOG_INFO, "Peak count: %f\n", (min_count + max_count) / (double)s->nb_channels); |
|
235 |
+ av_log(ctx, AV_LOG_INFO, "Number of samples: %lld\n", nb_samples / s->nb_channels); |
|
236 |
+} |
|
237 |
+ |
|
238 |
+static void uninit(AVFilterContext *ctx) |
|
239 |
+{ |
|
240 |
+ AudioStatsContext *s = ctx->priv; |
|
241 |
+ |
|
242 |
+ print_stats(ctx); |
|
243 |
+ av_freep(&s->chstats); |
|
244 |
+} |
|
245 |
+ |
|
246 |
+static const AVFilterPad astats_inputs[] = { |
|
247 |
+ { |
|
248 |
+ .name = "default", |
|
249 |
+ .type = AVMEDIA_TYPE_AUDIO, |
|
250 |
+ .filter_frame = filter_frame, |
|
251 |
+ }, |
|
252 |
+ { NULL } |
|
253 |
+}; |
|
254 |
+ |
|
255 |
+static const AVFilterPad astats_outputs[] = { |
|
256 |
+ { |
|
257 |
+ .name = "default", |
|
258 |
+ .type = AVMEDIA_TYPE_AUDIO, |
|
259 |
+ .config_props = config_output, |
|
260 |
+ }, |
|
261 |
+ { NULL } |
|
262 |
+}; |
|
263 |
+ |
|
264 |
+AVFilter avfilter_af_astats = { |
|
265 |
+ .name = "astats", |
|
266 |
+ .description = NULL_IF_CONFIG_SMALL("Show time domain statistics about audio frames."), |
|
267 |
+ .query_formats = query_formats, |
|
268 |
+ .priv_size = sizeof(AudioStatsContext), |
|
269 |
+ .priv_class = &astats_class, |
|
270 |
+ .uninit = uninit, |
|
271 |
+ .inputs = astats_inputs, |
|
272 |
+ .outputs = astats_outputs, |
|
273 |
+}; |
... | ... |
@@ -67,6 +67,7 @@ void avfilter_register_all(void) |
67 | 67 |
REGISTER_FILTER(ASETTB, asettb, af); |
68 | 68 |
REGISTER_FILTER(ASHOWINFO, ashowinfo, af); |
69 | 69 |
REGISTER_FILTER(ASPLIT, asplit, af); |
70 |
+ REGISTER_FILTER(ASTATS, astats, af); |
|
70 | 71 |
REGISTER_FILTER(ASTREAMSYNC, astreamsync, af); |
71 | 72 |
REGISTER_FILTER(ASYNCTS, asyncts, af); |
72 | 73 |
REGISTER_FILTER(ATEMPO, atempo, af); |
... | ... |
@@ -29,8 +29,8 @@ |
29 | 29 |
#include "libavutil/avutil.h" |
30 | 30 |
|
31 | 31 |
#define LIBAVFILTER_VERSION_MAJOR 3 |
32 |
-#define LIBAVFILTER_VERSION_MINOR 60 |
|
33 |
-#define LIBAVFILTER_VERSION_MICRO 102 |
|
32 |
+#define LIBAVFILTER_VERSION_MINOR 61 |
|
33 |
+#define LIBAVFILTER_VERSION_MICRO 100 |
|
34 | 34 |
|
35 | 35 |
#define LIBAVFILTER_VERSION_INT AV_VERSION_INT(LIBAVFILTER_VERSION_MAJOR, \ |
36 | 36 |
LIBAVFILTER_VERSION_MINOR, \ |