... | ... |
@@ -178,7 +178,7 @@ External library support: |
178 | 178 |
--enable-libopenjpeg enable JPEG 2000 decoding via OpenJPEG [no] |
179 | 179 |
--enable-librtmp enable RTMP[E] support via librtmp [no] |
180 | 180 |
--enable-libschroedinger enable Dirac support via libschroedinger [no] |
181 |
- --enable-libspeex enable Speex decoding via libspeex [no] |
|
181 |
+ --enable-libspeex enable Speex support via libspeex [no] |
|
182 | 182 |
--enable-libtheora enable Theora encoding via libtheora [no] |
183 | 183 |
--enable-libvo-aacenc enable AAC encoding via libvo-aacenc [no] |
184 | 184 |
--enable-libvo-amrwbenc enable AMR-WB encoding via libvo-amrwbenc [no] |
... | ... |
@@ -1406,6 +1406,7 @@ libopenjpeg_decoder_deps="libopenjpeg" |
1406 | 1406 |
libschroedinger_decoder_deps="libschroedinger" |
1407 | 1407 |
libschroedinger_encoder_deps="libschroedinger" |
1408 | 1408 |
libspeex_decoder_deps="libspeex" |
1409 |
+libspeex_encoder_deps="libspeex" |
|
1409 | 1410 |
libtheora_encoder_deps="libtheora" |
1410 | 1411 |
libvo_aacenc_encoder_deps="libvo_aacenc" |
1411 | 1412 |
libvo_amrwbenc_encoder_deps="libvo_amrwbenc" |
... | ... |
@@ -682,7 +682,7 @@ following image formats are supported: |
682 | 682 |
@tab Used in Sierra VMD files. |
683 | 683 |
@item Smacker audio @tab @tab X |
684 | 684 |
@item SMPTE 302M AES3 audio @tab @tab X |
685 |
-@item Speex @tab @tab E |
|
685 |
+@item Speex @tab E @tab E |
|
686 | 686 |
@tab supported through external library libspeex |
687 | 687 |
@item True Audio (TTA) @tab @tab X |
688 | 688 |
@item TrueHD @tab @tab X |
... | ... |
@@ -576,6 +576,7 @@ OBJS-$(CONFIG_LIBSCHROEDINGER_ENCODER) += libschroedingerenc.o \ |
576 | 576 |
libschroedinger.o \ |
577 | 577 |
libdirac_libschro.o |
578 | 578 |
OBJS-$(CONFIG_LIBSPEEX_DECODER) += libspeexdec.o |
579 |
+OBJS-$(CONFIG_LIBSPEEX_ENCODER) += libspeexenc.o |
|
579 | 580 |
OBJS-$(CONFIG_LIBTHEORA_ENCODER) += libtheoraenc.o |
580 | 581 |
OBJS-$(CONFIG_LIBVO_AACENC_ENCODER) += libvo-aacenc.o mpeg4audio.o |
581 | 582 |
OBJS-$(CONFIG_LIBVO_AMRWBENC_ENCODER) += libvo-amrwbenc.o |
... | ... |
@@ -370,7 +370,7 @@ void avcodec_register_all(void) |
370 | 370 |
REGISTER_DECODER (LIBOPENCORE_AMRWB, libopencore_amrwb); |
371 | 371 |
REGISTER_DECODER (LIBOPENJPEG, libopenjpeg); |
372 | 372 |
REGISTER_ENCDEC (LIBSCHROEDINGER, libschroedinger); |
373 |
- REGISTER_DECODER (LIBSPEEX, libspeex); |
|
373 |
+ REGISTER_ENCDEC (LIBSPEEX, libspeex); |
|
374 | 374 |
REGISTER_ENCODER (LIBTHEORA, libtheora); |
375 | 375 |
REGISTER_ENCODER (LIBVO_AACENC, libvo_aacenc); |
376 | 376 |
REGISTER_ENCODER (LIBVO_AMRWBENC, libvo_amrwbenc); |
377 | 377 |
new file mode 100644 |
... | ... |
@@ -0,0 +1,324 @@ |
0 |
+/* |
|
1 |
+ * Copyright (C) 2009 Justin Ruggles |
|
2 |
+ * Copyright (c) 2009 Xuggle Incorporated |
|
3 |
+ * |
|
4 |
+ * This file is part of Libav. |
|
5 |
+ * |
|
6 |
+ * Libav 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 |
+ * Libav 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 Libav; if not, write to the Free Software |
|
18 |
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
19 |
+ */ |
|
20 |
+ |
|
21 |
+/** |
|
22 |
+ * @file |
|
23 |
+ * libspeex Speex audio encoder |
|
24 |
+ * |
|
25 |
+ * Usage Guide |
|
26 |
+ * This explains the values that need to be set prior to initialization in |
|
27 |
+ * order to control various encoding parameters. |
|
28 |
+ * |
|
29 |
+ * Channels |
|
30 |
+ * Speex only supports mono or stereo, so avctx->channels must be set to |
|
31 |
+ * 1 or 2. |
|
32 |
+ * |
|
33 |
+ * Sample Rate / Encoding Mode |
|
34 |
+ * Speex has 3 modes, each of which uses a specific sample rate. |
|
35 |
+ * narrowband : 8 kHz |
|
36 |
+ * wideband : 16 kHz |
|
37 |
+ * ultra-wideband : 32 kHz |
|
38 |
+ * avctx->sample_rate must be set to one of these 3 values. This will be |
|
39 |
+ * used to set the encoding mode. |
|
40 |
+ * |
|
41 |
+ * Rate Control |
|
42 |
+ * VBR mode is turned on by setting CODEC_FLAG_QSCALE in avctx->flags. |
|
43 |
+ * avctx->global_quality is used to set the encoding quality. |
|
44 |
+ * For CBR mode, avctx->bit_rate can be used to set the constant bitrate. |
|
45 |
+ * Alternatively, the 'cbr_quality' option can be set from 0 to 10 to set |
|
46 |
+ * a constant bitrate based on quality. |
|
47 |
+ * For ABR mode, set avctx->bit_rate and set the 'abr' option to 1. |
|
48 |
+ * Approx. Bitrate Range: |
|
49 |
+ * narrowband : 2400 - 25600 bps |
|
50 |
+ * wideband : 4000 - 43200 bps |
|
51 |
+ * ultra-wideband : 4400 - 45200 bps |
|
52 |
+ * |
|
53 |
+ * Complexity |
|
54 |
+ * Encoding complexity is controlled by setting avctx->compression_level. |
|
55 |
+ * The valid range is 0 to 10. A higher setting gives generally better |
|
56 |
+ * quality at the expense of encoding speed. This does not affect the |
|
57 |
+ * bit rate. |
|
58 |
+ * |
|
59 |
+ * Frames-per-Packet |
|
60 |
+ * The encoder defaults to using 1 frame-per-packet. However, it is |
|
61 |
+ * sometimes desirable to use multiple frames-per-packet to reduce the |
|
62 |
+ * amount of container overhead. This can be done by setting the |
|
63 |
+ * 'frames_per_packet' option to a value 1 to 8. |
|
64 |
+ */ |
|
65 |
+ |
|
66 |
+#include <speex/speex.h> |
|
67 |
+#include <speex/speex_header.h> |
|
68 |
+#include <speex/speex_stereo.h> |
|
69 |
+#include "libavutil/mathematics.h" |
|
70 |
+#include "libavutil/opt.h" |
|
71 |
+#include "avcodec.h" |
|
72 |
+#include "internal.h" |
|
73 |
+ |
|
74 |
+typedef struct { |
|
75 |
+ AVClass *class; ///< AVClass for private options |
|
76 |
+ SpeexBits bits; ///< libspeex bitwriter context |
|
77 |
+ SpeexHeader header; ///< libspeex header struct |
|
78 |
+ void *enc_state; ///< libspeex encoder state |
|
79 |
+ int frames_per_packet; ///< number of frames to encode in each packet |
|
80 |
+ float vbr_quality; ///< VBR quality 0.0 to 10.0 |
|
81 |
+ int cbr_quality; ///< CBR quality 0 to 10 |
|
82 |
+ int abr; ///< flag to enable ABR |
|
83 |
+ int pkt_frame_count; ///< frame count for the current packet |
|
84 |
+ int lookahead; ///< encoder delay |
|
85 |
+ int sample_count; ///< total sample count (used for pts) |
|
86 |
+} LibSpeexEncContext; |
|
87 |
+ |
|
88 |
+static av_cold void print_enc_params(AVCodecContext *avctx, |
|
89 |
+ LibSpeexEncContext *s) |
|
90 |
+{ |
|
91 |
+ const char *mode_str = "unknown"; |
|
92 |
+ |
|
93 |
+ av_log(avctx, AV_LOG_DEBUG, "channels: %d\n", avctx->channels); |
|
94 |
+ switch (s->header.mode) { |
|
95 |
+ case SPEEX_MODEID_NB: mode_str = "narrowband"; break; |
|
96 |
+ case SPEEX_MODEID_WB: mode_str = "wideband"; break; |
|
97 |
+ case SPEEX_MODEID_UWB: mode_str = "ultra-wideband"; break; |
|
98 |
+ } |
|
99 |
+ av_log(avctx, AV_LOG_DEBUG, "mode: %s\n", mode_str); |
|
100 |
+ if (s->header.vbr) { |
|
101 |
+ av_log(avctx, AV_LOG_DEBUG, "rate control: VBR\n"); |
|
102 |
+ av_log(avctx, AV_LOG_DEBUG, " quality: %f\n", s->vbr_quality); |
|
103 |
+ } else if (s->abr) { |
|
104 |
+ av_log(avctx, AV_LOG_DEBUG, "rate control: ABR\n"); |
|
105 |
+ av_log(avctx, AV_LOG_DEBUG, " bitrate: %d bps\n", avctx->bit_rate); |
|
106 |
+ } else { |
|
107 |
+ av_log(avctx, AV_LOG_DEBUG, "rate control: CBR\n"); |
|
108 |
+ av_log(avctx, AV_LOG_DEBUG, " bitrate: %d bps\n", avctx->bit_rate); |
|
109 |
+ } |
|
110 |
+ av_log(avctx, AV_LOG_DEBUG, "complexity: %d\n", |
|
111 |
+ avctx->compression_level); |
|
112 |
+ av_log(avctx, AV_LOG_DEBUG, "frame size: %d samples\n", |
|
113 |
+ avctx->frame_size); |
|
114 |
+ av_log(avctx, AV_LOG_DEBUG, "frames per packet: %d\n", |
|
115 |
+ s->frames_per_packet); |
|
116 |
+ av_log(avctx, AV_LOG_DEBUG, "packet size: %d\n", |
|
117 |
+ avctx->frame_size * s->frames_per_packet); |
|
118 |
+} |
|
119 |
+ |
|
120 |
+static av_cold int encode_init(AVCodecContext *avctx) |
|
121 |
+{ |
|
122 |
+ LibSpeexEncContext *s = avctx->priv_data; |
|
123 |
+ const SpeexMode *mode; |
|
124 |
+ uint8_t *header_data; |
|
125 |
+ int header_size; |
|
126 |
+ int32_t complexity; |
|
127 |
+ |
|
128 |
+ /* channels */ |
|
129 |
+ if (avctx->channels < 1 || avctx->channels > 2) { |
|
130 |
+ av_log(avctx, AV_LOG_ERROR, "Invalid channels (%d). Only stereo and " |
|
131 |
+ "mono are supported\n", avctx->channels); |
|
132 |
+ return AVERROR(EINVAL); |
|
133 |
+ } |
|
134 |
+ |
|
135 |
+ /* sample rate and encoding mode */ |
|
136 |
+ switch (avctx->sample_rate) { |
|
137 |
+ case 8000: mode = &speex_nb_mode; break; |
|
138 |
+ case 16000: mode = &speex_wb_mode; break; |
|
139 |
+ case 32000: mode = &speex_uwb_mode; break; |
|
140 |
+ default: |
|
141 |
+ av_log(avctx, AV_LOG_ERROR, "Sample rate of %d Hz is not supported. " |
|
142 |
+ "Resample to 8, 16, or 32 kHz.\n", avctx->sample_rate); |
|
143 |
+ return AVERROR(EINVAL); |
|
144 |
+ } |
|
145 |
+ |
|
146 |
+ /* initialize libspeex */ |
|
147 |
+ s->enc_state = speex_encoder_init(mode); |
|
148 |
+ if (!s->enc_state) { |
|
149 |
+ av_log(avctx, AV_LOG_ERROR, "Error initializing libspeex\n"); |
|
150 |
+ return -1; |
|
151 |
+ } |
|
152 |
+ speex_init_header(&s->header, avctx->sample_rate, avctx->channels, mode); |
|
153 |
+ |
|
154 |
+ /* rate control method and parameters */ |
|
155 |
+ if (avctx->flags & CODEC_FLAG_QSCALE) { |
|
156 |
+ /* VBR */ |
|
157 |
+ s->header.vbr = 1; |
|
158 |
+ speex_encoder_ctl(s->enc_state, SPEEX_SET_VBR, &s->header.vbr); |
|
159 |
+ s->vbr_quality = av_clipf(avctx->global_quality / (float)FF_QP2LAMBDA, |
|
160 |
+ 0.0f, 10.0f); |
|
161 |
+ speex_encoder_ctl(s->enc_state, SPEEX_SET_VBR_QUALITY, &s->vbr_quality); |
|
162 |
+ } else { |
|
163 |
+ s->header.bitrate = avctx->bit_rate; |
|
164 |
+ if (avctx->bit_rate > 0) { |
|
165 |
+ /* CBR or ABR by bitrate */ |
|
166 |
+ if (s->abr) { |
|
167 |
+ speex_encoder_ctl(s->enc_state, SPEEX_SET_ABR, |
|
168 |
+ &s->header.bitrate); |
|
169 |
+ speex_encoder_ctl(s->enc_state, SPEEX_GET_ABR, |
|
170 |
+ &s->header.bitrate); |
|
171 |
+ } else { |
|
172 |
+ speex_encoder_ctl(s->enc_state, SPEEX_SET_BITRATE, |
|
173 |
+ &s->header.bitrate); |
|
174 |
+ speex_encoder_ctl(s->enc_state, SPEEX_GET_BITRATE, |
|
175 |
+ &s->header.bitrate); |
|
176 |
+ } |
|
177 |
+ } else { |
|
178 |
+ /* CBR by quality */ |
|
179 |
+ speex_encoder_ctl(s->enc_state, SPEEX_SET_QUALITY, |
|
180 |
+ &s->cbr_quality); |
|
181 |
+ speex_encoder_ctl(s->enc_state, SPEEX_GET_BITRATE, |
|
182 |
+ &s->header.bitrate); |
|
183 |
+ } |
|
184 |
+ /* stereo side information adds about 800 bps to the base bitrate */ |
|
185 |
+ /* TODO: this should be calculated exactly */ |
|
186 |
+ avctx->bit_rate = s->header.bitrate + (avctx->channels == 2 ? 800 : 0); |
|
187 |
+ } |
|
188 |
+ |
|
189 |
+ /* set encoding complexity */ |
|
190 |
+ if (avctx->compression_level > FF_COMPRESSION_DEFAULT) { |
|
191 |
+ complexity = av_clip(avctx->compression_level, 0, 10); |
|
192 |
+ speex_encoder_ctl(s->enc_state, SPEEX_SET_COMPLEXITY, &complexity); |
|
193 |
+ } |
|
194 |
+ speex_encoder_ctl(s->enc_state, SPEEX_GET_COMPLEXITY, &complexity); |
|
195 |
+ avctx->compression_level = complexity; |
|
196 |
+ |
|
197 |
+ /* set packet size */ |
|
198 |
+ avctx->frame_size = s->header.frame_size; |
|
199 |
+ s->header.frames_per_packet = s->frames_per_packet; |
|
200 |
+ |
|
201 |
+ /* set encoding delay */ |
|
202 |
+ speex_encoder_ctl(s->enc_state, SPEEX_GET_LOOKAHEAD, &s->lookahead); |
|
203 |
+ s->sample_count = -s->lookahead; |
|
204 |
+ |
|
205 |
+ /* create header packet bytes from header struct */ |
|
206 |
+ /* note: libspeex allocates the memory for header_data, which is freed |
|
207 |
+ below with speex_header_free() */ |
|
208 |
+ header_data = speex_header_to_packet(&s->header, &header_size); |
|
209 |
+ |
|
210 |
+ /* allocate extradata and coded_frame */ |
|
211 |
+ avctx->extradata = av_malloc(header_size + FF_INPUT_BUFFER_PADDING_SIZE); |
|
212 |
+ avctx->coded_frame = avcodec_alloc_frame(); |
|
213 |
+ if (!avctx->extradata || !avctx->coded_frame) { |
|
214 |
+ speex_header_free(header_data); |
|
215 |
+ speex_encoder_destroy(s->enc_state); |
|
216 |
+ av_log(avctx, AV_LOG_ERROR, "memory allocation error\n"); |
|
217 |
+ return AVERROR(ENOMEM); |
|
218 |
+ } |
|
219 |
+ |
|
220 |
+ /* copy header packet to extradata */ |
|
221 |
+ memcpy(avctx->extradata, header_data, header_size); |
|
222 |
+ avctx->extradata_size = header_size; |
|
223 |
+ speex_header_free(header_data); |
|
224 |
+ |
|
225 |
+ /* init libspeex bitwriter */ |
|
226 |
+ speex_bits_init(&s->bits); |
|
227 |
+ |
|
228 |
+ print_enc_params(avctx, s); |
|
229 |
+ return 0; |
|
230 |
+} |
|
231 |
+ |
|
232 |
+static int encode_frame(AVCodecContext *avctx, uint8_t *frame, int buf_size, |
|
233 |
+ void *data) |
|
234 |
+{ |
|
235 |
+ LibSpeexEncContext *s = avctx->priv_data; |
|
236 |
+ int16_t *samples = data; |
|
237 |
+ int sample_count = s->sample_count; |
|
238 |
+ |
|
239 |
+ if (data) { |
|
240 |
+ /* encode Speex frame */ |
|
241 |
+ if (avctx->channels == 2) |
|
242 |
+ speex_encode_stereo_int(samples, s->header.frame_size, &s->bits); |
|
243 |
+ speex_encode_int(s->enc_state, samples, &s->bits); |
|
244 |
+ s->pkt_frame_count++; |
|
245 |
+ s->sample_count += avctx->frame_size; |
|
246 |
+ } else { |
|
247 |
+ /* handle end-of-stream */ |
|
248 |
+ if (!s->pkt_frame_count) |
|
249 |
+ return 0; |
|
250 |
+ /* add extra terminator codes for unused frames in last packet */ |
|
251 |
+ while (s->pkt_frame_count < s->frames_per_packet) { |
|
252 |
+ speex_bits_pack(&s->bits, 15, 5); |
|
253 |
+ s->pkt_frame_count++; |
|
254 |
+ } |
|
255 |
+ } |
|
256 |
+ |
|
257 |
+ /* write output if all frames for the packet have been encoded */ |
|
258 |
+ if (s->pkt_frame_count == s->frames_per_packet) { |
|
259 |
+ s->pkt_frame_count = 0; |
|
260 |
+ avctx->coded_frame->pts = |
|
261 |
+ av_rescale_q(sample_count, (AVRational){ 1, avctx->sample_rate }, |
|
262 |
+ avctx->time_base); |
|
263 |
+ if (buf_size > speex_bits_nbytes(&s->bits)) { |
|
264 |
+ int ret = speex_bits_write(&s->bits, frame, buf_size); |
|
265 |
+ speex_bits_reset(&s->bits); |
|
266 |
+ return ret; |
|
267 |
+ } else { |
|
268 |
+ av_log(avctx, AV_LOG_ERROR, "output buffer too small"); |
|
269 |
+ return AVERROR(EINVAL); |
|
270 |
+ } |
|
271 |
+ } |
|
272 |
+ return 0; |
|
273 |
+} |
|
274 |
+ |
|
275 |
+static av_cold int encode_close(AVCodecContext *avctx) |
|
276 |
+{ |
|
277 |
+ LibSpeexEncContext *s = avctx->priv_data; |
|
278 |
+ |
|
279 |
+ speex_bits_destroy(&s->bits); |
|
280 |
+ speex_encoder_destroy(s->enc_state); |
|
281 |
+ |
|
282 |
+ av_freep(&avctx->coded_frame); |
|
283 |
+ av_freep(&avctx->extradata); |
|
284 |
+ |
|
285 |
+ return 0; |
|
286 |
+} |
|
287 |
+ |
|
288 |
+#define OFFSET(x) offsetof(LibSpeexEncContext, x) |
|
289 |
+#define AE AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_ENCODING_PARAM |
|
290 |
+static const AVOption options[] = { |
|
291 |
+ { "abr", "Use average bit rate", OFFSET(abr), AV_OPT_TYPE_INT, { 0 }, 0, 1, AE }, |
|
292 |
+ { "cbr_quality", "Set quality value (0 to 10) for CBR", OFFSET(cbr_quality), AV_OPT_TYPE_INT, { 8 }, 0, 10, AE }, |
|
293 |
+ { "frames_per_packet", "Number of frames to encode in each packet", OFFSET(frames_per_packet), AV_OPT_TYPE_INT, { 1 }, 1, 8, AE }, |
|
294 |
+ { NULL }, |
|
295 |
+}; |
|
296 |
+ |
|
297 |
+static const AVClass class = { |
|
298 |
+ .class_name = "libspeex", |
|
299 |
+ .item_name = av_default_item_name, |
|
300 |
+ .option = options, |
|
301 |
+ .version = LIBAVUTIL_VERSION_INT, |
|
302 |
+}; |
|
303 |
+ |
|
304 |
+static const AVCodecDefault defaults[] = { |
|
305 |
+ { "b", "0" }, |
|
306 |
+ { "compression_level", "3" }, |
|
307 |
+ { NULL }, |
|
308 |
+}; |
|
309 |
+ |
|
310 |
+AVCodec ff_libspeex_encoder = { |
|
311 |
+ .name = "libspeex", |
|
312 |
+ .type = AVMEDIA_TYPE_AUDIO, |
|
313 |
+ .id = CODEC_ID_SPEEX, |
|
314 |
+ .priv_data_size = sizeof(LibSpeexEncContext), |
|
315 |
+ .init = encode_init, |
|
316 |
+ .encode = encode_frame, |
|
317 |
+ .close = encode_close, |
|
318 |
+ .capabilities = CODEC_CAP_DELAY, |
|
319 |
+ .sample_fmts = (const enum SampleFormat[]){ AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_NONE }, |
|
320 |
+ .long_name = NULL_IF_CONFIG_SMALL("libspeex Speex"), |
|
321 |
+ .priv_class = &class, |
|
322 |
+ .defaults = defaults, |
|
323 |
+}; |
... | ... |
@@ -21,7 +21,7 @@ |
21 | 21 |
#define AVCODEC_VERSION_H |
22 | 22 |
|
23 | 23 |
#define LIBAVCODEC_VERSION_MAJOR 53 |
24 |
-#define LIBAVCODEC_VERSION_MINOR 13 |
|
24 |
+#define LIBAVCODEC_VERSION_MINOR 14 |
|
25 | 25 |
#define LIBAVCODEC_VERSION_MICRO 0 |
26 | 26 |
|
27 | 27 |
#define LIBAVCODEC_VERSION_INT AV_VERSION_INT(LIBAVCODEC_VERSION_MAJOR, \ |