Commit | Line | Data |
---|---|---|
2ba45a60 DM |
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" | |
28 | #include "libavutil/channel_layout.h" | |
29 | #include "libavutil/opt.h" | |
30 | #include "libavutil/samplefmt.h" | |
31 | #include "libavutil/avassert.h" | |
32 | #include "libswresample/swresample.h" | |
33 | #include "avfilter.h" | |
34 | #include "audio.h" | |
35 | #include "internal.h" | |
36 | ||
37 | typedef struct { | |
38 | const AVClass *class; | |
39 | int sample_rate_arg; | |
40 | double ratio; | |
41 | struct SwrContext *swr; | |
42 | int64_t next_pts; | |
43 | int req_fullfilled; | |
f6fa7814 | 44 | int more_data; |
2ba45a60 DM |
45 | } AResampleContext; |
46 | ||
47 | static av_cold int init_dict(AVFilterContext *ctx, AVDictionary **opts) | |
48 | { | |
49 | AResampleContext *aresample = ctx->priv; | |
50 | int ret = 0; | |
51 | ||
52 | aresample->next_pts = AV_NOPTS_VALUE; | |
53 | aresample->swr = swr_alloc(); | |
54 | if (!aresample->swr) { | |
55 | ret = AVERROR(ENOMEM); | |
56 | goto end; | |
57 | } | |
58 | ||
59 | if (opts) { | |
60 | AVDictionaryEntry *e = NULL; | |
61 | ||
62 | while ((e = av_dict_get(*opts, "", e, AV_DICT_IGNORE_SUFFIX))) { | |
63 | if ((ret = av_opt_set(aresample->swr, e->key, e->value, 0)) < 0) | |
64 | goto end; | |
65 | } | |
66 | av_dict_free(opts); | |
67 | } | |
68 | if (aresample->sample_rate_arg > 0) | |
69 | av_opt_set_int(aresample->swr, "osr", aresample->sample_rate_arg, 0); | |
70 | end: | |
71 | return ret; | |
72 | } | |
73 | ||
74 | static av_cold void uninit(AVFilterContext *ctx) | |
75 | { | |
76 | AResampleContext *aresample = ctx->priv; | |
77 | swr_free(&aresample->swr); | |
78 | } | |
79 | ||
80 | static int query_formats(AVFilterContext *ctx) | |
81 | { | |
82 | AResampleContext *aresample = ctx->priv; | |
83 | int out_rate = av_get_int(aresample->swr, "osr", NULL); | |
84 | uint64_t out_layout = av_get_int(aresample->swr, "ocl", NULL); | |
85 | enum AVSampleFormat out_format = av_get_int(aresample->swr, "osf", NULL); | |
86 | ||
87 | AVFilterLink *inlink = ctx->inputs[0]; | |
88 | AVFilterLink *outlink = ctx->outputs[0]; | |
89 | ||
90 | AVFilterFormats *in_formats = ff_all_formats(AVMEDIA_TYPE_AUDIO); | |
91 | AVFilterFormats *out_formats; | |
92 | AVFilterFormats *in_samplerates = ff_all_samplerates(); | |
93 | AVFilterFormats *out_samplerates; | |
94 | AVFilterChannelLayouts *in_layouts = ff_all_channel_counts(); | |
95 | AVFilterChannelLayouts *out_layouts; | |
96 | ||
97 | ff_formats_ref (in_formats, &inlink->out_formats); | |
98 | ff_formats_ref (in_samplerates, &inlink->out_samplerates); | |
99 | ff_channel_layouts_ref(in_layouts, &inlink->out_channel_layouts); | |
100 | ||
101 | if(out_rate > 0) { | |
102 | int ratelist[] = { out_rate, -1 }; | |
103 | out_samplerates = ff_make_format_list(ratelist); | |
104 | } else { | |
105 | out_samplerates = ff_all_samplerates(); | |
106 | } | |
f6fa7814 DM |
107 | if (!out_samplerates) { |
108 | av_log(ctx, AV_LOG_ERROR, "Cannot allocate output samplerates.\n"); | |
109 | return AVERROR(ENOMEM); | |
110 | } | |
111 | ||
2ba45a60 DM |
112 | ff_formats_ref(out_samplerates, &outlink->in_samplerates); |
113 | ||
114 | if(out_format != AV_SAMPLE_FMT_NONE) { | |
115 | int formatlist[] = { out_format, -1 }; | |
116 | out_formats = ff_make_format_list(formatlist); | |
117 | } else | |
118 | out_formats = ff_all_formats(AVMEDIA_TYPE_AUDIO); | |
119 | ff_formats_ref(out_formats, &outlink->in_formats); | |
120 | ||
121 | if(out_layout) { | |
122 | int64_t layout_list[] = { out_layout, -1 }; | |
123 | out_layouts = avfilter_make_format64_list(layout_list); | |
124 | } else | |
125 | out_layouts = ff_all_channel_counts(); | |
126 | ff_channel_layouts_ref(out_layouts, &outlink->in_channel_layouts); | |
127 | ||
128 | return 0; | |
129 | } | |
130 | ||
131 | ||
132 | static int config_output(AVFilterLink *outlink) | |
133 | { | |
134 | int ret; | |
135 | AVFilterContext *ctx = outlink->src; | |
136 | AVFilterLink *inlink = ctx->inputs[0]; | |
137 | AResampleContext *aresample = ctx->priv; | |
138 | int out_rate; | |
139 | uint64_t out_layout; | |
140 | enum AVSampleFormat out_format; | |
141 | char inchl_buf[128], outchl_buf[128]; | |
142 | ||
143 | aresample->swr = swr_alloc_set_opts(aresample->swr, | |
144 | outlink->channel_layout, outlink->format, outlink->sample_rate, | |
145 | inlink->channel_layout, inlink->format, inlink->sample_rate, | |
146 | 0, ctx); | |
147 | if (!aresample->swr) | |
148 | return AVERROR(ENOMEM); | |
149 | if (!inlink->channel_layout) | |
150 | av_opt_set_int(aresample->swr, "ich", inlink->channels, 0); | |
151 | if (!outlink->channel_layout) | |
152 | av_opt_set_int(aresample->swr, "och", outlink->channels, 0); | |
153 | ||
154 | ret = swr_init(aresample->swr); | |
155 | if (ret < 0) | |
156 | return ret; | |
157 | ||
158 | out_rate = av_get_int(aresample->swr, "osr", NULL); | |
159 | out_layout = av_get_int(aresample->swr, "ocl", NULL); | |
160 | out_format = av_get_int(aresample->swr, "osf", NULL); | |
161 | outlink->time_base = (AVRational) {1, out_rate}; | |
162 | ||
163 | av_assert0(outlink->sample_rate == out_rate); | |
164 | av_assert0(outlink->channel_layout == out_layout || !outlink->channel_layout); | |
165 | av_assert0(outlink->format == out_format); | |
166 | ||
167 | aresample->ratio = (double)outlink->sample_rate / inlink->sample_rate; | |
168 | ||
169 | av_get_channel_layout_string(inchl_buf, sizeof(inchl_buf), inlink ->channels, inlink ->channel_layout); | |
170 | av_get_channel_layout_string(outchl_buf, sizeof(outchl_buf), outlink->channels, outlink->channel_layout); | |
171 | ||
172 | av_log(ctx, AV_LOG_VERBOSE, "ch:%d chl:%s fmt:%s r:%dHz -> ch:%d chl:%s fmt:%s r:%dHz\n", | |
173 | inlink ->channels, inchl_buf, av_get_sample_fmt_name(inlink->format), inlink->sample_rate, | |
174 | outlink->channels, outchl_buf, av_get_sample_fmt_name(outlink->format), outlink->sample_rate); | |
175 | return 0; | |
176 | } | |
177 | ||
178 | static int filter_frame(AVFilterLink *inlink, AVFrame *insamplesref) | |
179 | { | |
180 | AResampleContext *aresample = inlink->dst->priv; | |
181 | const int n_in = insamplesref->nb_samples; | |
182 | int64_t delay; | |
183 | int n_out = n_in * aresample->ratio + 32; | |
184 | AVFilterLink *const outlink = inlink->dst->outputs[0]; | |
185 | AVFrame *outsamplesref; | |
186 | int ret; | |
187 | ||
188 | delay = swr_get_delay(aresample->swr, outlink->sample_rate); | |
189 | if (delay > 0) | |
f6fa7814 | 190 | n_out += FFMIN(delay, FFMAX(4096, n_out)); |
2ba45a60 DM |
191 | |
192 | outsamplesref = ff_get_audio_buffer(outlink, n_out); | |
193 | ||
194 | if(!outsamplesref) | |
195 | return AVERROR(ENOMEM); | |
196 | ||
197 | av_frame_copy_props(outsamplesref, insamplesref); | |
198 | outsamplesref->format = outlink->format; | |
199 | av_frame_set_channels(outsamplesref, outlink->channels); | |
200 | outsamplesref->channel_layout = outlink->channel_layout; | |
201 | outsamplesref->sample_rate = outlink->sample_rate; | |
202 | ||
203 | if(insamplesref->pts != AV_NOPTS_VALUE) { | |
204 | int64_t inpts = av_rescale(insamplesref->pts, inlink->time_base.num * (int64_t)outlink->sample_rate * inlink->sample_rate, inlink->time_base.den); | |
205 | int64_t outpts= swr_next_pts(aresample->swr, inpts); | |
206 | aresample->next_pts = | |
207 | outsamplesref->pts = ROUNDED_DIV(outpts, inlink->sample_rate); | |
208 | } else { | |
209 | outsamplesref->pts = AV_NOPTS_VALUE; | |
210 | } | |
211 | n_out = swr_convert(aresample->swr, outsamplesref->extended_data, n_out, | |
212 | (void *)insamplesref->extended_data, n_in); | |
213 | if (n_out <= 0) { | |
214 | av_frame_free(&outsamplesref); | |
215 | av_frame_free(&insamplesref); | |
216 | return 0; | |
217 | } | |
218 | ||
f6fa7814 DM |
219 | aresample->more_data = outsamplesref->nb_samples == n_out; // Indicate that there is probably more data in our buffers |
220 | ||
2ba45a60 DM |
221 | outsamplesref->nb_samples = n_out; |
222 | ||
223 | ret = ff_filter_frame(outlink, outsamplesref); | |
224 | aresample->req_fullfilled= 1; | |
225 | av_frame_free(&insamplesref); | |
226 | return ret; | |
227 | } | |
228 | ||
f6fa7814 | 229 | static int flush_frame(AVFilterLink *outlink, int final, AVFrame **outsamplesref_ret) |
2ba45a60 DM |
230 | { |
231 | AVFilterContext *ctx = outlink->src; | |
232 | AResampleContext *aresample = ctx->priv; | |
233 | AVFilterLink *const inlink = outlink->src->inputs[0]; | |
f6fa7814 DM |
234 | AVFrame *outsamplesref; |
235 | int n_out = 4096; | |
236 | int64_t pts; | |
237 | ||
238 | outsamplesref = ff_get_audio_buffer(outlink, n_out); | |
239 | *outsamplesref_ret = outsamplesref; | |
240 | if (!outsamplesref) | |
241 | return AVERROR(ENOMEM); | |
242 | ||
243 | pts = swr_next_pts(aresample->swr, INT64_MIN); | |
244 | pts = ROUNDED_DIV(pts, inlink->sample_rate); | |
245 | ||
246 | n_out = swr_convert(aresample->swr, outsamplesref->extended_data, n_out, final ? NULL : (void*)outsamplesref->extended_data, 0); | |
247 | if (n_out <= 0) { | |
248 | av_frame_free(&outsamplesref); | |
249 | return (n_out == 0) ? AVERROR_EOF : n_out; | |
250 | } | |
251 | ||
252 | outsamplesref->sample_rate = outlink->sample_rate; | |
253 | outsamplesref->nb_samples = n_out; | |
254 | ||
255 | outsamplesref->pts = pts; | |
256 | ||
257 | return 0; | |
258 | } | |
259 | ||
260 | static int request_frame(AVFilterLink *outlink) | |
261 | { | |
262 | AVFilterContext *ctx = outlink->src; | |
263 | AResampleContext *aresample = ctx->priv; | |
2ba45a60 DM |
264 | int ret; |
265 | ||
f6fa7814 DM |
266 | // First try to get data from the internal buffers |
267 | if (aresample->more_data) { | |
268 | AVFrame *outsamplesref; | |
269 | ||
270 | if (flush_frame(outlink, 0, &outsamplesref) >= 0) { | |
271 | return ff_filter_frame(outlink, outsamplesref); | |
272 | } | |
273 | } | |
274 | aresample->more_data = 0; | |
275 | ||
276 | // Second request more data from the input | |
2ba45a60 DM |
277 | aresample->req_fullfilled = 0; |
278 | do{ | |
279 | ret = ff_request_frame(ctx->inputs[0]); | |
280 | }while(!aresample->req_fullfilled && ret>=0); | |
281 | ||
f6fa7814 | 282 | // Third if we hit the end flush |
2ba45a60 DM |
283 | if (ret == AVERROR_EOF) { |
284 | AVFrame *outsamplesref; | |
2ba45a60 | 285 | |
f6fa7814 DM |
286 | if ((ret = flush_frame(outlink, 1, &outsamplesref)) < 0) |
287 | return ret; | |
2ba45a60 DM |
288 | |
289 | return ff_filter_frame(outlink, outsamplesref); | |
290 | } | |
291 | return ret; | |
292 | } | |
293 | ||
294 | static const AVClass *resample_child_class_next(const AVClass *prev) | |
295 | { | |
296 | return prev ? NULL : swr_get_class(); | |
297 | } | |
298 | ||
299 | static void *resample_child_next(void *obj, void *prev) | |
300 | { | |
301 | AResampleContext *s = obj; | |
302 | return prev ? NULL : s->swr; | |
303 | } | |
304 | ||
305 | #define OFFSET(x) offsetof(AResampleContext, x) | |
306 | #define FLAGS AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_FILTERING_PARAM | |
307 | ||
308 | static const AVOption options[] = { | |
309 | {"sample_rate", NULL, OFFSET(sample_rate_arg), AV_OPT_TYPE_INT, {.i64=0}, 0, INT_MAX, FLAGS }, | |
310 | {NULL} | |
311 | }; | |
312 | ||
313 | static const AVClass aresample_class = { | |
314 | .class_name = "aresample", | |
315 | .item_name = av_default_item_name, | |
316 | .option = options, | |
317 | .version = LIBAVUTIL_VERSION_INT, | |
318 | .child_class_next = resample_child_class_next, | |
319 | .child_next = resample_child_next, | |
320 | }; | |
321 | ||
322 | static const AVFilterPad aresample_inputs[] = { | |
323 | { | |
324 | .name = "default", | |
325 | .type = AVMEDIA_TYPE_AUDIO, | |
326 | .filter_frame = filter_frame, | |
327 | }, | |
328 | { NULL } | |
329 | }; | |
330 | ||
331 | static const AVFilterPad aresample_outputs[] = { | |
332 | { | |
333 | .name = "default", | |
334 | .config_props = config_output, | |
335 | .request_frame = request_frame, | |
336 | .type = AVMEDIA_TYPE_AUDIO, | |
337 | }, | |
338 | { NULL } | |
339 | }; | |
340 | ||
341 | AVFilter ff_af_aresample = { | |
342 | .name = "aresample", | |
343 | .description = NULL_IF_CONFIG_SMALL("Resample audio data."), | |
344 | .init_dict = init_dict, | |
345 | .uninit = uninit, | |
346 | .query_formats = query_formats, | |
347 | .priv_size = sizeof(AResampleContext), | |
348 | .priv_class = &aresample_class, | |
349 | .inputs = aresample_inputs, | |
350 | .outputs = aresample_outputs, | |
351 | }; |