FFmpeg
af_aresample.c
Go to the documentation of this file.
1 /*
2  * Copyright (c) 2011 Stefano Sabatini
3  * Copyright (c) 2011 Mina Nagy Zaki
4  *
5  * This file is part of FFmpeg.
6  *
7  * FFmpeg is free software; you can redistribute it and/or
8  * modify it under the terms of the GNU Lesser General Public
9  * License as published by the Free Software Foundation; either
10  * version 2.1 of the License, or (at your option) any later version.
11  *
12  * FFmpeg is distributed in the hope that it will be useful,
13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15  * Lesser General Public License for more details.
16  *
17  * You should have received a copy of the GNU Lesser General Public
18  * License along with FFmpeg; if not, write to the Free Software
19  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20  */
21 
22 /**
23  * @file
24  * resampling audio filter
25  */
26 
27 #include "libavutil/avstring.h"
29 #include "libavutil/opt.h"
30 #include "libavutil/samplefmt.h"
31 #include "libavutil/avassert.h"
33 #include "avfilter.h"
34 #include "audio.h"
35 #include "internal.h"
36 
37 typedef struct AResampleContext {
38  const AVClass *class;
40  double ratio;
41  struct SwrContext *swr;
42  int64_t next_pts;
43  int more_data;
45 
47 {
48  AResampleContext *aresample = ctx->priv;
49  int ret = 0;
50 
51  aresample->next_pts = AV_NOPTS_VALUE;
52  aresample->swr = swr_alloc();
53  if (!aresample->swr) {
54  ret = AVERROR(ENOMEM);
55  goto end;
56  }
57 
58  if (opts) {
60 
61  while ((e = av_dict_get(*opts, "", e, AV_DICT_IGNORE_SUFFIX))) {
62  if ((ret = av_opt_set(aresample->swr, e->key, e->value, 0)) < 0)
63  goto end;
64  }
65  av_dict_free(opts);
66  }
67  if (aresample->sample_rate_arg > 0)
68  av_opt_set_int(aresample->swr, "osr", aresample->sample_rate_arg, 0);
69 end:
70  return ret;
71 }
72 
74 {
75  AResampleContext *aresample = ctx->priv;
76  swr_free(&aresample->swr);
77 }
78 
80 {
81  AResampleContext *aresample = ctx->priv;
82  enum AVSampleFormat out_format;
83  int64_t out_rate, out_layout;
84 
85  AVFilterLink *inlink = ctx->inputs[0];
86  AVFilterLink *outlink = ctx->outputs[0];
87 
88  AVFilterFormats *in_formats, *out_formats;
89  AVFilterFormats *in_samplerates, *out_samplerates;
90  AVFilterChannelLayouts *in_layouts, *out_layouts;
91  int ret;
92 
93  av_opt_get_sample_fmt(aresample->swr, "osf", 0, &out_format);
94  av_opt_get_int(aresample->swr, "osr", 0, &out_rate);
95  av_opt_get_int(aresample->swr, "ocl", 0, &out_layout);
96 
97  in_formats = ff_all_formats(AVMEDIA_TYPE_AUDIO);
98  if ((ret = ff_formats_ref(in_formats, &inlink->out_formats)) < 0)
99  return ret;
100 
101  in_samplerates = ff_all_samplerates();
102  if ((ret = ff_formats_ref(in_samplerates, &inlink->out_samplerates)) < 0)
103  return ret;
104 
105  in_layouts = ff_all_channel_counts();
106  if ((ret = ff_channel_layouts_ref(in_layouts, &inlink->out_channel_layouts)) < 0)
107  return ret;
108 
109  if(out_rate > 0) {
110  int ratelist[] = { out_rate, -1 };
111  out_samplerates = ff_make_format_list(ratelist);
112  } else {
113  out_samplerates = ff_all_samplerates();
114  }
115 
116  if ((ret = ff_formats_ref(out_samplerates, &outlink->in_samplerates)) < 0)
117  return ret;
118 
119  if(out_format != AV_SAMPLE_FMT_NONE) {
120  int formatlist[] = { out_format, -1 };
121  out_formats = ff_make_format_list(formatlist);
122  } else
123  out_formats = ff_all_formats(AVMEDIA_TYPE_AUDIO);
124  if ((ret = ff_formats_ref(out_formats, &outlink->in_formats)) < 0)
125  return ret;
126 
127  if(out_layout) {
128  int64_t layout_list[] = { out_layout, -1 };
129  out_layouts = avfilter_make_format64_list(layout_list);
130  } else
131  out_layouts = ff_all_channel_counts();
132 
133  return ff_channel_layouts_ref(out_layouts, &outlink->in_channel_layouts);
134 }
135 
136 
137 static int config_output(AVFilterLink *outlink)
138 {
139  int ret;
140  AVFilterContext *ctx = outlink->src;
141  AVFilterLink *inlink = ctx->inputs[0];
142  AResampleContext *aresample = ctx->priv;
143  int64_t out_rate, out_layout;
144  enum AVSampleFormat out_format;
145  char inchl_buf[128], outchl_buf[128];
146 
147  aresample->swr = swr_alloc_set_opts(aresample->swr,
148  outlink->channel_layout, outlink->format, outlink->sample_rate,
149  inlink->channel_layout, inlink->format, inlink->sample_rate,
150  0, ctx);
151  if (!aresample->swr)
152  return AVERROR(ENOMEM);
153  if (!inlink->channel_layout)
154  av_opt_set_int(aresample->swr, "ich", inlink->channels, 0);
155  if (!outlink->channel_layout)
156  av_opt_set_int(aresample->swr, "och", outlink->channels, 0);
157 
158  ret = swr_init(aresample->swr);
159  if (ret < 0)
160  return ret;
161 
162  av_opt_get_int(aresample->swr, "osr", 0, &out_rate);
163  av_opt_get_int(aresample->swr, "ocl", 0, &out_layout);
164  av_opt_get_sample_fmt(aresample->swr, "osf", 0, &out_format);
165  outlink->time_base = (AVRational) {1, out_rate};
166 
167  av_assert0(outlink->sample_rate == out_rate);
168  av_assert0(outlink->channel_layout == out_layout || !outlink->channel_layout);
169  av_assert0(outlink->format == out_format);
170 
171  aresample->ratio = (double)outlink->sample_rate / inlink->sample_rate;
172 
173  av_get_channel_layout_string(inchl_buf, sizeof(inchl_buf), inlink ->channels, inlink ->channel_layout);
174  av_get_channel_layout_string(outchl_buf, sizeof(outchl_buf), outlink->channels, outlink->channel_layout);
175 
176  av_log(ctx, AV_LOG_VERBOSE, "ch:%d chl:%s fmt:%s r:%dHz -> ch:%d chl:%s fmt:%s r:%dHz\n",
177  inlink ->channels, inchl_buf, av_get_sample_fmt_name(inlink->format), inlink->sample_rate,
178  outlink->channels, outchl_buf, av_get_sample_fmt_name(outlink->format), outlink->sample_rate);
179  return 0;
180 }
181 
182 static int filter_frame(AVFilterLink *inlink, AVFrame *insamplesref)
183 {
184  AResampleContext *aresample = inlink->dst->priv;
185  const int n_in = insamplesref->nb_samples;
186  int64_t delay;
187  int n_out = n_in * aresample->ratio + 32;
188  AVFilterLink *const outlink = inlink->dst->outputs[0];
189  AVFrame *outsamplesref;
190  int ret;
191 
192  delay = swr_get_delay(aresample->swr, outlink->sample_rate);
193  if (delay > 0)
194  n_out += FFMIN(delay, FFMAX(4096, n_out));
195 
196  outsamplesref = ff_get_audio_buffer(outlink, n_out);
197 
198  if(!outsamplesref) {
199  av_frame_free(&insamplesref);
200  return AVERROR(ENOMEM);
201  }
202 
203  av_frame_copy_props(outsamplesref, insamplesref);
204  outsamplesref->format = outlink->format;
205  outsamplesref->channels = outlink->channels;
206  outsamplesref->channel_layout = outlink->channel_layout;
207  outsamplesref->sample_rate = outlink->sample_rate;
208 
209  if(insamplesref->pts != AV_NOPTS_VALUE) {
210  int64_t inpts = av_rescale(insamplesref->pts, inlink->time_base.num * (int64_t)outlink->sample_rate * inlink->sample_rate, inlink->time_base.den);
211  int64_t outpts= swr_next_pts(aresample->swr, inpts);
212  aresample->next_pts =
213  outsamplesref->pts = ROUNDED_DIV(outpts, inlink->sample_rate);
214  } else {
215  outsamplesref->pts = AV_NOPTS_VALUE;
216  }
217  n_out = swr_convert(aresample->swr, outsamplesref->extended_data, n_out,
218  (void *)insamplesref->extended_data, n_in);
219  if (n_out <= 0) {
220  av_frame_free(&outsamplesref);
221  av_frame_free(&insamplesref);
222  return 0;
223  }
224 
225  aresample->more_data = outsamplesref->nb_samples == n_out; // Indicate that there is probably more data in our buffers
226 
227  outsamplesref->nb_samples = n_out;
228 
229  ret = ff_filter_frame(outlink, outsamplesref);
230  av_frame_free(&insamplesref);
231  return ret;
232 }
233 
234 static int flush_frame(AVFilterLink *outlink, int final, AVFrame **outsamplesref_ret)
235 {
236  AVFilterContext *ctx = outlink->src;
237  AResampleContext *aresample = ctx->priv;
238  AVFilterLink *const inlink = outlink->src->inputs[0];
239  AVFrame *outsamplesref;
240  int n_out = 4096;
241  int64_t pts;
242 
243  outsamplesref = ff_get_audio_buffer(outlink, n_out);
244  *outsamplesref_ret = outsamplesref;
245  if (!outsamplesref)
246  return AVERROR(ENOMEM);
247 
248  pts = swr_next_pts(aresample->swr, INT64_MIN);
249  pts = ROUNDED_DIV(pts, inlink->sample_rate);
250 
251  n_out = swr_convert(aresample->swr, outsamplesref->extended_data, n_out, final ? NULL : (void*)outsamplesref->extended_data, 0);
252  if (n_out <= 0) {
253  av_frame_free(&outsamplesref);
254  return (n_out == 0) ? AVERROR_EOF : n_out;
255  }
256 
257  outsamplesref->sample_rate = outlink->sample_rate;
258  outsamplesref->nb_samples = n_out;
259 
260  outsamplesref->pts = pts;
261 
262  return 0;
263 }
264 
265 static int request_frame(AVFilterLink *outlink)
266 {
267  AVFilterContext *ctx = outlink->src;
268  AResampleContext *aresample = ctx->priv;
269  int ret;
270 
271  // First try to get data from the internal buffers
272  if (aresample->more_data) {
273  AVFrame *outsamplesref;
274 
275  if (flush_frame(outlink, 0, &outsamplesref) >= 0) {
276  return ff_filter_frame(outlink, outsamplesref);
277  }
278  }
279  aresample->more_data = 0;
280 
281  // Second request more data from the input
282  ret = ff_request_frame(ctx->inputs[0]);
283 
284  // Third if we hit the end flush
285  if (ret == AVERROR_EOF) {
286  AVFrame *outsamplesref;
287 
288  if ((ret = flush_frame(outlink, 1, &outsamplesref)) < 0)
289  return ret;
290 
291  return ff_filter_frame(outlink, outsamplesref);
292  }
293  return ret;
294 }
295 
296 static const AVClass *resample_child_class_next(const AVClass *prev)
297 {
298  return prev ? NULL : swr_get_class();
299 }
300 
301 static void *resample_child_next(void *obj, void *prev)
302 {
303  AResampleContext *s = obj;
304  return prev ? NULL : s->swr;
305 }
306 
307 #define OFFSET(x) offsetof(AResampleContext, x)
308 #define FLAGS AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_FILTERING_PARAM
309 
310 static const AVOption options[] = {
311  {"sample_rate", NULL, OFFSET(sample_rate_arg), AV_OPT_TYPE_INT, {.i64=0}, 0, INT_MAX, FLAGS },
312  {NULL}
313 };
314 
315 static const AVClass aresample_class = {
316  .class_name = "aresample",
317  .item_name = av_default_item_name,
318  .option = options,
319  .version = LIBAVUTIL_VERSION_INT,
320  .child_class_next = resample_child_class_next,
322 };
323 
324 static const AVFilterPad aresample_inputs[] = {
325  {
326  .name = "default",
327  .type = AVMEDIA_TYPE_AUDIO,
328  .filter_frame = filter_frame,
329  },
330  { NULL }
331 };
332 
333 static const AVFilterPad aresample_outputs[] = {
334  {
335  .name = "default",
336  .config_props = config_output,
337  .request_frame = request_frame,
338  .type = AVMEDIA_TYPE_AUDIO,
339  },
340  { NULL }
341 };
342 
344  .name = "aresample",
345  .description = NULL_IF_CONFIG_SMALL("Resample audio data."),
346  .init_dict = init_dict,
347  .uninit = uninit,
348  .query_formats = query_formats,
349  .priv_size = sizeof(AResampleContext),
350  .priv_class = &aresample_class,
351  .inputs = aresample_inputs,
352  .outputs = aresample_outputs,
353 };
#define NULL
Definition: coverity.c:32
This structure describes decoded (raw) audio or video data.
Definition: frame.h:295
AVOption.
Definition: opt.h:246
static const AVFilterPad aresample_inputs[]
Definition: af_aresample.c:324
#define LIBAVUTIL_VERSION_INT
Definition: version.h:85
Main libavfilter public API header.
channels
Definition: aptx.c:30
int num
Numerator.
Definition: rational.h:59
const char * av_default_item_name(void *ptr)
Return the context name.
Definition: log.c:191
int64_t swr_next_pts(struct SwrContext *s, int64_t pts)
Convert the next timestamp from input to output timestamps are in 1/(in_sample_rate * out_sample_rate...
Definition: swresample.c:917
static const AVOption options[]
Definition: af_aresample.c:310
AVFilterFormats * ff_make_format_list(const int *fmts)
Create a list of supported formats.
Definition: formats.c:283
const char * name
Pad name.
Definition: internal.h:60
const char * class_name
The name of the class; usually it is the same name as the context structure type to which the AVClass...
Definition: log.h:72
AVFilterLink ** inputs
array of pointers to input links
Definition: avfilter.h:346
#define av_assert0(cond)
assert() equivalent, that is always enabled.
Definition: avassert.h:37
int ff_channel_layouts_ref(AVFilterChannelLayouts *f, AVFilterChannelLayouts **ref)
Add *ref as a new reference to f.
Definition: formats.c:435
int ff_filter_frame(AVFilterLink *link, AVFrame *frame)
Send a frame of data to the next filter.
Definition: avfilter.c:1080
av_cold struct SwrContext * swr_alloc(void)
Allocate SwrContext.
Definition: options.c:149
#define av_cold
Definition: attributes.h:82
AVOptions.
#define FLAGS
Definition: af_aresample.c:308
static av_cold int end(AVCodecContext *avctx)
Definition: avrndec.c:90
int64_t pts
Presentation timestamp in time_base units (time when frame should be shown to user).
Definition: frame.h:388
AVDictionaryEntry * av_dict_get(const AVDictionary *m, const char *key, const AVDictionaryEntry *prev, int flags)
Get a dictionary entry with matching key.
Definition: dict.c:40
#define AVERROR_EOF
End of file.
Definition: error.h:55
#define AV_LOG_VERBOSE
Detailed information.
Definition: log.h:192
#define av_log(a,...)
AVFilterFormats * ff_all_formats(enum AVMediaType type)
Return a list of all formats supported by FFmpeg for the given media type.
Definition: formats.c:350
#define ROUNDED_DIV(a, b)
Definition: common.h:56
A filter pad used for either input or output.
Definition: internal.h:54
void *(* child_next)(void *obj, void *prev)
Return next AVOptions-enabled child or NULL.
Definition: log.h:113
libswresample public header
static void * resample_child_next(void *obj, void *prev)
Definition: af_aresample.c:301
AVFrame * ff_get_audio_buffer(AVFilterLink *link, int nb_samples)
Request an audio samples buffer with a specific set of permissions.
Definition: audio.c:86
void av_frame_free(AVFrame **frame)
Free the frame and any dynamically allocated objects in it, e.g.
Definition: frame.c:202
The libswresample context.
#define NULL_IF_CONFIG_SMALL(x)
Return NULL if CONFIG_SMALL is true, otherwise the argument without modification. ...
Definition: internal.h:186
void * priv
private data for use by the filter
Definition: avfilter.h:353
void av_dict_free(AVDictionary **pm)
Free all the memory allocated for an AVDictionary struct and all keys and values. ...
Definition: dict.c:203
int av_opt_set_int(void *obj, const char *name, int64_t val, int search_flags)
Definition: opt.c:568
simple assert() macros that are a bit more flexible than ISO C assert().
const char * av_get_sample_fmt_name(enum AVSampleFormat sample_fmt)
Return the name of sample_fmt, or NULL if sample_fmt is not recognized.
Definition: samplefmt.c:49
int64_t swr_get_delay(struct SwrContext *s, int64_t base)
Gets the delay the next input sample will experience relative to the next output sample.
Definition: swresample.c:867
#define FFMAX(a, b)
Definition: common.h:94
uint64_t channel_layout
Channel layout of the audio data.
Definition: frame.h:472
AVDictionary * opts
Definition: movenc.c:50
int av_opt_get_sample_fmt(void *obj, const char *name, int search_flags, enum AVSampleFormat *out_fmt)
Definition: opt.c:985
int channels
number of audio channels, only used for audio.
Definition: frame.h:601
audio channel layout utility functions
int64_t av_rescale(int64_t a, int64_t b, int64_t c)
Rescale a 64-bit integer with rounding to nearest.
Definition: mathematics.c:129
#define FFMIN(a, b)
Definition: common.h:96
int ff_formats_ref(AVFilterFormats *f, AVFilterFormats **ref)
Add *ref as a new reference to formats.
Definition: formats.c:440
static int flush_frame(AVFilterLink *outlink, int final, AVFrame **outsamplesref_ret)
Definition: af_aresample.c:234
static const AVClass aresample_class
Definition: af_aresample.c:315
static av_cold int init_dict(AVFilterContext *ctx, AVDictionary **opts)
Definition: af_aresample.c:46
AVFormatContext * ctx
Definition: movenc.c:48
struct SwrContext * swr_alloc_set_opts(struct SwrContext *s, int64_t out_ch_layout, enum AVSampleFormat out_sample_fmt, int out_sample_rate, int64_t in_ch_layout, enum AVSampleFormat in_sample_fmt, int in_sample_rate, int log_offset, void *log_ctx)
Allocate SwrContext if needed and set/reset common parameters.
Definition: swresample.c:59
#define s(width, name)
Definition: cbs_vp9.c:257
const AVClass * swr_get_class(void)
Get the AVClass for SwrContext.
Definition: options.c:144
static const AVFilterPad outputs[]
Definition: af_acontrast.c:203
A list of supported channel layouts.
Definition: formats.h:85
void av_get_channel_layout_string(char *buf, int buf_size, int nb_channels, uint64_t channel_layout)
Return a description of a channel layout.
static av_cold void uninit(AVFilterContext *ctx)
Definition: af_aresample.c:73
int format
format of the frame, -1 if unknown or unset Values correspond to enum AVPixelFormat for video frames...
Definition: frame.h:368
int av_opt_get_int(void *obj, const char *name, int search_flags, int64_t *out_val)
Definition: opt.c:888
int64_t outpts
output PTS
AVFilterChannelLayouts * avfilter_make_format64_list(const int64_t *fmts)
Definition: formats.c:303
AVFilter ff_af_aresample
Definition: af_aresample.c:343
AVSampleFormat
Audio sample formats.
Definition: samplefmt.h:58
static const AVFilterPad aresample_outputs[]
Definition: af_aresample.c:333
av_cold void swr_free(SwrContext **ss)
Free the given SwrContext and set the pointer to NULL.
Definition: swresample.c:137
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several inputs
Describe the class of an AVClass context structure.
Definition: log.h:67
int sample_rate
Sample rate of the audio data.
Definition: frame.h:467
Filter definition.
Definition: avfilter.h:144
Rational number (pair of numerator and denominator).
Definition: rational.h:58
struct SwrContext * swr
Definition: af_aresample.c:41
const char * name
Filter name.
Definition: avfilter.h:148
int attribute_align_arg swr_convert(struct SwrContext *s, uint8_t *out_arg[SWR_CH_MAX], int out_count, const uint8_t *in_arg[SWR_CH_MAX], int in_count)
Definition: swresample.c:714
AVFilterLink ** outputs
array of pointers to output links
Definition: avfilter.h:350
AVFilterFormats * ff_all_samplerates(void)
Definition: formats.c:395
static int64_t pts
The exact code depends on how similar the blocks are and how related they are to the and needs to apply these operations to the correct inlink or outlink if there are several Macros are available to factor that when no extra processing is inlink
static int filter_frame(AVFilterLink *inlink, AVFrame *insamplesref)
Definition: af_aresample.c:182
static const AVClass * resample_child_class_next(const AVClass *prev)
Definition: af_aresample.c:296
char * key
Definition: dict.h:86
int den
Denominator.
Definition: rational.h:60
#define OFFSET(x)
Definition: af_aresample.c:307
char * value
Definition: dict.h:87
A list of supported formats for one end of a filter link.
Definition: formats.h:64
An instance of a filter.
Definition: avfilter.h:338
#define AV_DICT_IGNORE_SUFFIX
Return first entry in a dictionary whose first part corresponds to the search key, ignoring the suffix of the found key string.
Definition: dict.h:70
int ff_request_frame(AVFilterLink *link)
Request an input frame from the filter at the other end of the link.
Definition: avfilter.c:407
internal API functions
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel sample they are references to shared objects When the negotiation mechanism computes the intersection of the formats supported at each end of a all references to both lists are replaced with a reference to the intersection And when a single format is eventually chosen for a link amongst the remaining all references to the list are updated That means that if a filter requires that its input and output have the same format amongst a supported all it has to do is use a reference to the same list of formats query_formats can leave some formats unset and return AVERROR(EAGAIN) to cause the negotiation mechanism toagain later.That can be used by filters with complex requirements to use the format negotiated on one link to set the formats supported on another.Frame references ownership and permissions
AVFilterChannelLayouts * ff_all_channel_counts(void)
Construct an AVFilterChannelLayouts coding for any channel layout, with known or unknown disposition...
Definition: formats.c:410
static int query_formats(AVFilterContext *ctx)
Definition: af_aresample.c:79
uint8_t ** extended_data
pointers to the data planes/channels.
Definition: frame.h:342
int av_opt_set(void *obj, const char *name, const char *val, int search_flags)
Definition: opt.c:449
int nb_samples
number of audio samples (per channel) described by this frame
Definition: frame.h:361
int av_frame_copy_props(AVFrame *dst, const AVFrame *src)
Copy only "metadata" fields from src to dst.
Definition: frame.c:654
#define AV_NOPTS_VALUE
Undefined timestamp value.
Definition: avutil.h:248
static int request_frame(AVFilterLink *outlink)
Definition: af_aresample.c:265
av_cold int swr_init(struct SwrContext *s)
Initialize context after user parameters have been set.
Definition: swresample.c:152
static int config_output(AVFilterLink *outlink)
Definition: af_aresample.c:137