FFmpeg
 All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Macros Groups Pages
swresample.h
Go to the documentation of this file.
1 /*
2  * Copyright (C) 2011-2013 Michael Niedermayer (michaelni@gmx.at)
3  *
4  * This file is part of libswresample
5  *
6  * libswresample 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  * libswresample 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 libswresample; if not, write to the Free Software
18  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19  */
20 
21 #ifndef SWRESAMPLE_SWRESAMPLE_H
22 #define SWRESAMPLE_SWRESAMPLE_H
23 
24 /**
25  * @file
26  * @ingroup lswr
27  * libswresample public header
28  */
29 
30 /**
31  * @defgroup lswr Libswresample
32  * @{
33  *
34  * Libswresample (lswr) is a library that handles audio resampling, sample
35  * format conversion and mixing.
36  *
37  * Interaction with lswr is done through SwrContext, which is
38  * allocated with swr_alloc() or swr_alloc_set_opts(). It is opaque, so all parameters
39  * must be set with the @ref avoptions API.
40  *
41  * For example the following code will setup conversion from planar float sample
42  * format to interleaved signed 16-bit integer, downsampling from 48kHz to
43  * 44.1kHz and downmixing from 5.1 channels to stereo (using the default mixing
44  * matrix):
45  * @code
46  * SwrContext *swr = swr_alloc();
47  * av_opt_set_channel_layout(swr, "in_channel_layout", AV_CH_LAYOUT_5POINT1, 0);
48  * av_opt_set_channel_layout(swr, "out_channel_layout", AV_CH_LAYOUT_STEREO, 0);
49  * av_opt_set_int(swr, "in_sample_rate", 48000, 0);
50  * av_opt_set_int(swr, "out_sample_rate", 44100, 0);
51  * av_opt_set_sample_fmt(swr, "in_sample_fmt", AV_SAMPLE_FMT_FLTP, 0);
52  * av_opt_set_sample_fmt(swr, "out_sample_fmt", AV_SAMPLE_FMT_S16, 0);
53  * @endcode
54  *
55  * Once all values have been set, it must be initialized with swr_init(). If
56  * you need to change the conversion parameters, you can change the parameters
57  * as described above, or by using swr_alloc_set_opts(), then call swr_init()
58  * again.
59  *
60  * The conversion itself is done by repeatedly calling swr_convert().
61  * Note that the samples may get buffered in swr if you provide insufficient
62  * output space or if sample rate conversion is done, which requires "future"
63  * samples. Samples that do not require future input can be retrieved at any
64  * time by using swr_convert() (in_count can be set to 0).
65  * At the end of conversion the resampling buffer can be flushed by calling
66  * swr_convert() with NULL in and 0 in_count.
67  *
68  * The delay between input and output, can at any time be found by using
69  * swr_get_delay().
70  *
71  * The following code demonstrates the conversion loop assuming the parameters
72  * from above and caller-defined functions get_input() and handle_output():
73  * @code
74  * uint8_t **input;
75  * int in_samples;
76  *
77  * while (get_input(&input, &in_samples)) {
78  * uint8_t *output;
79  * int out_samples = av_rescale_rnd(swr_get_delay(swr, 48000) +
80  * in_samples, 44100, 48000, AV_ROUND_UP);
81  * av_samples_alloc(&output, NULL, 2, out_samples,
82  * AV_SAMPLE_FMT_S16, 0);
83  * out_samples = swr_convert(swr, &output, out_samples,
84  * input, in_samples);
85  * handle_output(output, out_samples);
86  * av_freep(&output);
87  * }
88  * @endcode
89  *
90  * When the conversion is finished, the conversion
91  * context and everything associated with it must be freed with swr_free().
92  * There will be no memory leak if the data is not completely flushed before
93  * swr_free().
94  */
95 
96 #include <stdint.h>
97 #include "libavutil/samplefmt.h"
98 
99 #include "libswresample/version.h"
100 
101 #if LIBSWRESAMPLE_VERSION_MAJOR < 1
102 #define SWR_CH_MAX 32 ///< Maximum number of channels
103 #endif
104 
105 #define SWR_FLAG_RESAMPLE 1 ///< Force resampling even if equal sample rate
106 //TODO use int resample ?
107 //long term TODO can we enable this dynamically?
108 
114 
115  SWR_DITHER_NS = 64, ///< not part of API/ABI
123  SWR_DITHER_NB, ///< not part of API/ABI
124 };
125 
126 /** Resampling Engines */
127 enum SwrEngine {
128  SWR_ENGINE_SWR, /**< SW Resampler */
129  SWR_ENGINE_SOXR, /**< SoX Resampler */
130  SWR_ENGINE_NB, ///< not part of API/ABI
131 };
132 
133 /** Resampling Filter Types */
135  SWR_FILTER_TYPE_CUBIC, /**< Cubic */
136  SWR_FILTER_TYPE_BLACKMAN_NUTTALL, /**< Blackman Nuttall Windowed Sinc */
137  SWR_FILTER_TYPE_KAISER, /**< Kaiser Windowed Sinc */
138 };
139 
140 typedef struct SwrContext SwrContext;
141 
142 /**
143  * Get the AVClass for swrContext. It can be used in combination with
144  * AV_OPT_SEARCH_FAKE_OBJ for examining options.
145  *
146  * @see av_opt_find().
147  */
148 const AVClass *swr_get_class(void);
149 
150 /**
151  * Allocate SwrContext.
152  *
153  * If you use this function you will need to set the parameters (manually or
154  * with swr_alloc_set_opts()) before calling swr_init().
155  *
156  * @see swr_alloc_set_opts(), swr_init(), swr_free()
157  * @return NULL on error, allocated context otherwise
158  */
159 struct SwrContext *swr_alloc(void);
160 
161 /**
162  * Initialize context after user parameters have been set.
163  *
164  * @return AVERROR error code in case of failure.
165  */
166 int swr_init(struct SwrContext *s);
167 
168 /**
169  * Check whether an swr context has been initialized or not.
170  *
171  * @return positive if it has been initialized, 0 if not initialized
172  */
173 int swr_is_initialized(struct SwrContext *s);
174 
175 /**
176  * Allocate SwrContext if needed and set/reset common parameters.
177  *
178  * This function does not require s to be allocated with swr_alloc(). On the
179  * other hand, swr_alloc() can use swr_alloc_set_opts() to set the parameters
180  * on the allocated context.
181  *
182  * @param s Swr context, can be NULL
183  * @param out_ch_layout output channel layout (AV_CH_LAYOUT_*)
184  * @param out_sample_fmt output sample format (AV_SAMPLE_FMT_*).
185  * @param out_sample_rate output sample rate (frequency in Hz)
186  * @param in_ch_layout input channel layout (AV_CH_LAYOUT_*)
187  * @param in_sample_fmt input sample format (AV_SAMPLE_FMT_*).
188  * @param in_sample_rate input sample rate (frequency in Hz)
189  * @param log_offset logging level offset
190  * @param log_ctx parent logging context, can be NULL
191  *
192  * @see swr_init(), swr_free()
193  * @return NULL on error, allocated context otherwise
194  */
195 struct SwrContext *swr_alloc_set_opts(struct SwrContext *s,
198  int log_offset, void *log_ctx);
199 
200 /**
201  * Free the given SwrContext and set the pointer to NULL.
202  */
203 void swr_free(struct SwrContext **s);
204 
205 /**
206  * Convert audio.
207  *
208  * in and in_count can be set to 0 to flush the last few samples out at the
209  * end.
210  *
211  * If more input is provided than output space then the input will be buffered.
212  * You can avoid this buffering by providing more output space than input.
213  * Convertion will run directly without copying whenever possible.
214  *
215  * @param s allocated Swr context, with parameters set
216  * @param out output buffers, only the first one need be set in case of packed audio
217  * @param out_count amount of space available for output in samples per channel
218  * @param in input buffers, only the first one need to be set in case of packed audio
219  * @param in_count number of input samples available in one channel
220  *
221  * @return number of samples output per channel, negative value on error
222  */
223 int swr_convert(struct SwrContext *s, uint8_t **out, int out_count,
224  const uint8_t **in , int in_count);
225 
226 /**
227  * Convert the next timestamp from input to output
228  * timestamps are in 1/(in_sample_rate * out_sample_rate) units.
229  *
230  * @note There are 2 slightly differently behaving modes.
231  * First is when automatic timestamp compensation is not used, (min_compensation >= FLT_MAX)
232  * in this case timestamps will be passed through with delays compensated
233  * Second is when automatic timestamp compensation is used, (min_compensation < FLT_MAX)
234  * in this case the output timestamps will match output sample numbers
235  *
236  * @param pts timestamp for the next input sample, INT64_MIN if unknown
237  * @return the output timestamp for the next output sample
238  */
239 int64_t swr_next_pts(struct SwrContext *s, int64_t pts);
240 
241 /**
242  * Activate resampling compensation.
243  */
244 int swr_set_compensation(struct SwrContext *s, int sample_delta, int compensation_distance);
245 
246 /**
247  * Set a customized input channel mapping.
248  *
249  * @param s allocated Swr context, not yet initialized
250  * @param channel_map customized input channel mapping (array of channel
251  * indexes, -1 for a muted channel)
252  * @return AVERROR error code in case of failure.
253  */
254 int swr_set_channel_mapping(struct SwrContext *s, const int *channel_map);
255 
256 /**
257  * Set a customized remix matrix.
258  *
259  * @param s allocated Swr context, not yet initialized
260  * @param matrix remix coefficients; matrix[i + stride * o] is
261  * the weight of input channel i in output channel o
262  * @param stride offset between lines of the matrix
263  * @return AVERROR error code in case of failure.
264  */
265 int swr_set_matrix(struct SwrContext *s, const double *matrix, int stride);
266 
267 /**
268  * Drops the specified number of output samples.
269  */
270 int swr_drop_output(struct SwrContext *s, int count);
271 
272 /**
273  * Injects the specified number of silence samples.
274  */
275 int swr_inject_silence(struct SwrContext *s, int count);
276 
277 /**
278  * Gets the delay the next input sample will experience relative to the next output sample.
279  *
280  * Swresample can buffer data if more input has been provided than available
281  * output space, also converting between sample rates needs a delay.
282  * This function returns the sum of all such delays.
283  * The exact delay is not necessarily an integer value in either input or
284  * output sample rate. Especially when downsampling by a large value, the
285  * output sample rate may be a poor choice to represent the delay, similarly
286  * for upsampling and the input sample rate.
287  *
288  * @param s swr context
289  * @param base timebase in which the returned delay will be
290  * if its set to 1 the returned delay is in seconds
291  * if its set to 1000 the returned delay is in milli seconds
292  * if its set to the input sample rate then the returned delay is in input samples
293  * if its set to the output sample rate then the returned delay is in output samples
294  * an exact rounding free delay can be found by using LCM(in_sample_rate, out_sample_rate)
295  * @returns the delay in 1/base units.
296  */
297 int64_t swr_get_delay(struct SwrContext *s, int64_t base);
298 
299 /**
300  * Return the LIBSWRESAMPLE_VERSION_INT constant.
301  */
302 unsigned swresample_version(void);
303 
304 /**
305  * Return the swr build-time configuration.
306  */
307 const char *swresample_configuration(void);
308 
309 /**
310  * Return the swr license.
311  */
312 const char *swresample_license(void);
313 
314 /**
315  * @}
316  */
317 
318 #endif /* SWRESAMPLE_SWRESAMPLE_H */