Remove commented-out while loop.
[libav.git] / libavcodec / amr.c
1 /*
2 * AMR Audio decoder stub
3 * Copyright (c) 2003 the ffmpeg project
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 /** @file
23 * Adaptive Multi-Rate (AMR) Audio decoder stub.
24 *
25 * This code implements both an AMR-NarrowBand (AMR-NB) and an AMR-WideBand
26 * (AMR-WB) audio encoder/decoder through external reference code from
27 * http://www.3gpp.org/. The license of the code from 3gpp is unclear so you
28 * have to download the code separately. Two versions exists: One fixed-point
29 * and one with floats. For some reason the float-encoder is significant faster
30 * at least on a P4 1.5GHz (0.9s instead of 9.9s on a 30s audio clip at MR102).
31 * Both float and fixed point are supported for AMR-NB, but only float for
32 * AMR-WB.
33 *
34 * \section AMR-NB
35 *
36 * \subsection Float
37 * The float version (default) can be downloaded from:
38 * http://www.3gpp.org/ftp/Specs/archive/26_series/26.104/26104-610.zip
39 * Extract the source into \c "ffmpeg/libavcodec/amr_float".
40 * Enable it by passing \c "--enable-amr-nb" to \c "./configure".
41 *
42 * \subsection Fixed-point
43 * The fixed-point (TS26.073) can be downloaded from:
44 * http://www.3gpp.org/ftp/Specs/archive/26_series/26.073/26073-510.zip.
45 * Extract the source into \c "ffmpeg/libavcodec/amr".
46 * Enable it by passing \c "--enable-amr-nb-fixed" to \c "./configure".
47 *
48 * \subsection Specification
49 * The specification for AMR-NB can be found in TS 26.071
50 * (http://www.3gpp.org/ftp/Specs/html-info/26071.htm) and some other
51 * info at http://www.3gpp.org/ftp/Specs/html-info/26-series.htm.
52 *
53 * \section AMR-WB
54 * \subsection Float
55 * The reference code can be downloaded from:
56 * http://www.3gpp.org/ftp/Specs/archive/26_series/26.204/26204-600.zip
57 * It should be extracted to \c "ffmpeg/libavcodec/amrwb_float".
58 * Enable it by passing \c "--enable-amr-wb" to \c "./configure".
59 *
60 * \subsection Fixed-point
61 * If someone wants to use the fixed point version it can be downloaded from:
62 * http://www.3gpp.org/ftp/Specs/archive/26_series/26.173/26173-571.zip.
63 *
64 * \subsection Specification
65 * The specification for AMR-WB can be found in TS 26.171
66 * (http://www.3gpp.org/ftp/Specs/html-info/26171.htm) and some other
67 * info at http://www.3gpp.org/ftp/Specs/html-info/26-series.htm.
68 *
69 */
70
71 #include "avcodec.h"
72
73 #ifdef CONFIG_AMR_NB_FIXED
74
75 #define MMS_IO
76
77 #include "amr/sp_dec.h"
78 #include "amr/d_homing.h"
79 #include "amr/typedef.h"
80 #include "amr/sp_enc.h"
81 #include "amr/sid_sync.h"
82 #include "amr/e_homing.h"
83
84 #else
85 #include "amr_float/interf_dec.h"
86 #include "amr_float/interf_enc.h"
87 #endif
88
89 /* Common code for fixed and float version*/
90 typedef struct AMR_bitrates
91 {
92 int rate;
93 enum Mode mode;
94 } AMR_bitrates;
95
96 /* Match desired bitrate */
97 static int getBitrateMode(int bitrate)
98 {
99 /* make the correspondance between bitrate and mode */
100 AMR_bitrates rates[]={ {4750,MR475},
101 {5150,MR515},
102 {5900,MR59},
103 {6700,MR67},
104 {7400,MR74},
105 {7950,MR795},
106 {10200,MR102},
107 {12200,MR122},
108 };
109 int i;
110
111 for(i=0;i<8;i++)
112 {
113 if(rates[i].rate==bitrate)
114 {
115 return(rates[i].mode);
116 }
117 }
118 /* no bitrate matching, return an error */
119 return -1;
120 }
121
122 static void amr_decode_fix_avctx(AVCodecContext * avctx)
123 {
124 const int is_amr_wb = 1 + (avctx->codec_id == CODEC_ID_AMR_WB);
125
126 if(avctx->sample_rate == 0)
127 {
128 avctx->sample_rate = 8000 * is_amr_wb;
129 }
130
131 if(avctx->channels == 0)
132 {
133 avctx->channels = 1;
134 }
135
136 avctx->frame_size = 160 * is_amr_wb;
137 }
138
139 #ifdef CONFIG_AMR_NB_FIXED
140 /* fixed point version*/
141 /* frame size in serial bitstream file (frame type + serial stream + flags) */
142 #define SERIAL_FRAMESIZE (1+MAX_SERIAL_SIZE+5)
143
144 typedef struct AMRContext {
145 int frameCount;
146 Speech_Decode_FrameState *speech_decoder_state;
147 enum RXFrameType rx_type;
148 enum Mode mode;
149 Word16 reset_flag;
150 Word16 reset_flag_old;
151
152 int enc_bitrate;
153 Speech_Encode_FrameState *enstate;
154 sid_syncState *sidstate;
155 enum TXFrameType tx_frametype;
156 } AMRContext;
157
158 static int amr_nb_decode_init(AVCodecContext * avctx)
159 {
160 AMRContext *s = avctx->priv_data;
161
162 s->frameCount=0;
163 s->speech_decoder_state=NULL;
164 s->rx_type = (enum RXFrameType)0;
165 s->mode= (enum Mode)0;
166 s->reset_flag=0;
167 s->reset_flag_old=1;
168
169 if(Speech_Decode_Frame_init(&s->speech_decoder_state, "Decoder"))
170 {
171 av_log(avctx, AV_LOG_ERROR, "Speech_Decode_Frame_init error\n");
172 return -1;
173 }
174
175 amr_decode_fix_avctx(avctx);
176
177 if(avctx->channels > 1)
178 {
179 av_log(avctx, AV_LOG_ERROR, "amr_nb: multichannel decoding not supported\n");
180 return -1;
181 }
182
183 return 0;
184 }
185
186 static int amr_nb_encode_init(AVCodecContext * avctx)
187 {
188 AMRContext *s = avctx->priv_data;
189
190 s->frameCount=0;
191 s->speech_decoder_state=NULL;
192 s->rx_type = (enum RXFrameType)0;
193 s->mode= (enum Mode)0;
194 s->reset_flag=0;
195 s->reset_flag_old=1;
196
197 if(avctx->sample_rate!=8000)
198 {
199 av_log(avctx, AV_LOG_ERROR, "Only 8000Hz sample rate supported\n");
200 return -1;
201 }
202
203 if(avctx->channels!=1)
204 {
205 av_log(avctx, AV_LOG_ERROR, "Only mono supported\n");
206 return -1;
207 }
208
209 avctx->frame_size=160;
210 avctx->coded_frame= avcodec_alloc_frame();
211
212 if(Speech_Encode_Frame_init(&s->enstate, 0, "encoder") || sid_sync_init (&s->sidstate))
213 {
214 av_log(avctx, AV_LOG_ERROR, "Speech_Encode_Frame_init error\n");
215 return -1;
216 }
217
218 if((s->enc_bitrate=getBitrateMode(avctx->bit_rate))<0)
219 {
220 av_log(avctx, AV_LOG_ERROR, "bitrate not supported\n");
221 return -1;
222 }
223
224 return 0;
225 }
226
227 static int amr_nb_encode_close(AVCodecContext * avctx)
228 {
229 AMRContext *s = avctx->priv_data;
230
231 Speech_Encode_Frame_exit(&s->enstate);
232 sid_sync_exit (&s->sidstate);
233 av_freep(&avctx->coded_frame);
234 return 0;
235 }
236
237 static int amr_nb_decode_close(AVCodecContext * avctx)
238 {
239 AMRContext *s = avctx->priv_data;
240
241 Speech_Decode_Frame_exit(&s->speech_decoder_state);
242 return 0;
243 }
244
245 static int amr_nb_decode_frame(AVCodecContext * avctx,
246 void *data, int *data_size,
247 uint8_t * buf, int buf_size)
248 {
249 AMRContext *s = avctx->priv_data;
250 uint8_t*amrData=buf;
251 int offset=0;
252 UWord8 toc, q, ft;
253 Word16 serial[SERIAL_FRAMESIZE]; /* coded bits */
254 Word16 *synth;
255 UWord8 *packed_bits;
256 static Word16 packed_size[16] = {12, 13, 15, 17, 19, 20, 26, 31, 5, 0, 0, 0, 0, 0, 0, 0};
257 int i;
258
259 //printf("amr_decode_frame data_size=%i buf=0x%X buf_size=%d frameCount=%d!!\n",*data_size,buf,buf_size,s->frameCount);
260
261 synth=data;
262
263 toc=amrData[offset];
264 /* read rest of the frame based on ToC byte */
265 q = (toc >> 2) & 0x01;
266 ft = (toc >> 3) & 0x0F;
267
268 //printf("offset=%d, packet_size=%d amrData= 0x%X %X %X %X\n",offset,packed_size[ft],amrData[offset],amrData[offset+1],amrData[offset+2],amrData[offset+3]);
269
270 offset++;
271
272 packed_bits=amrData+offset;
273
274 offset+=packed_size[ft];
275
276 //Unsort and unpack bits
277 s->rx_type = UnpackBits(q, ft, packed_bits, &s->mode, &serial[1]);
278
279 //We have a new frame
280 s->frameCount++;
281
282 if (s->rx_type == RX_NO_DATA)
283 {
284 s->mode = s->speech_decoder_state->prev_mode;
285 }
286 else {
287 s->speech_decoder_state->prev_mode = s->mode;
288 }
289
290 /* if homed: check if this frame is another homing frame */
291 if (s->reset_flag_old == 1)
292 {
293 /* only check until end of first subframe */
294 s->reset_flag = decoder_homing_frame_test_first(&serial[1], s->mode);
295 }
296 /* produce encoder homing frame if homed & input=decoder homing frame */
297 if ((s->reset_flag != 0) && (s->reset_flag_old != 0))
298 {
299 for (i = 0; i < L_FRAME; i++)
300 {
301 synth[i] = EHF_MASK;
302 }
303 }
304 else
305 {
306 /* decode frame */
307 Speech_Decode_Frame(s->speech_decoder_state, s->mode, &serial[1], s->rx_type, synth);
308 }
309
310 //Each AMR-frame results in 160 16-bit samples
311 *data_size+=160*2;
312 synth+=160;
313
314 /* if not homed: check whether current frame is a homing frame */
315 if (s->reset_flag_old == 0)
316 {
317 /* check whole frame */
318 s->reset_flag = decoder_homing_frame_test(&serial[1], s->mode);
319 }
320 /* reset decoder if current frame is a homing frame */
321 if (s->reset_flag != 0)
322 {
323 Speech_Decode_Frame_reset(s->speech_decoder_state);
324 }
325 s->reset_flag_old = s->reset_flag;
326
327 return offset;
328 }
329
330
331 static int amr_nb_encode_frame(AVCodecContext *avctx,
332 unsigned char *frame/*out*/, int buf_size, void *data/*in*/)
333 {
334 short serial_data[250] = {0};
335 AMRContext *s = avctx->priv_data;
336 int written;
337
338 s->reset_flag = encoder_homing_frame_test(data);
339
340 Speech_Encode_Frame(s->enstate, s->enc_bitrate, data, &serial_data[1], &s->mode);
341
342 /* add frame type and mode */
343 sid_sync (s->sidstate, s->mode, &s->tx_frametype);
344
345 written = PackBits(s->mode, s->enc_bitrate, s->tx_frametype, &serial_data[1], frame);
346
347 if (s->reset_flag != 0)
348 {
349 Speech_Encode_Frame_reset(s->enstate);
350 sid_sync_reset(s->sidstate);
351 }
352 return written;
353 }
354
355
356 #elif defined(CONFIG_AMR_NB) /* Float point version*/
357
358 typedef struct AMRContext {
359 int frameCount;
360 void * decState;
361 int *enstate;
362 int enc_bitrate;
363 } AMRContext;
364
365 static int amr_nb_decode_init(AVCodecContext * avctx)
366 {
367 AMRContext *s = avctx->priv_data;
368
369 s->frameCount=0;
370 s->decState=Decoder_Interface_init();
371 if(!s->decState)
372 {
373 av_log(avctx, AV_LOG_ERROR, "Decoder_Interface_init error\r\n");
374 return -1;
375 }
376
377 amr_decode_fix_avctx(avctx);
378
379 if(avctx->channels > 1)
380 {
381 av_log(avctx, AV_LOG_ERROR, "amr_nb: multichannel decoding not supported\n");
382 return -1;
383 }
384
385 return 0;
386 }
387
388 static int amr_nb_encode_init(AVCodecContext * avctx)
389 {
390 AMRContext *s = avctx->priv_data;
391
392 s->frameCount=0;
393
394 if(avctx->sample_rate!=8000)
395 {
396 av_log(avctx, AV_LOG_ERROR, "Only 8000Hz sample rate supported\n");
397 return -1;
398 }
399
400 if(avctx->channels!=1)
401 {
402 av_log(avctx, AV_LOG_ERROR, "Only mono supported\n");
403 return -1;
404 }
405
406 avctx->frame_size=160;
407 avctx->coded_frame= avcodec_alloc_frame();
408
409 s->enstate=Encoder_Interface_init(0);
410 if(!s->enstate)
411 {
412 av_log(avctx, AV_LOG_ERROR, "Encoder_Interface_init error\n");
413 return -1;
414 }
415
416 if((s->enc_bitrate=getBitrateMode(avctx->bit_rate))<0)
417 {
418 av_log(avctx, AV_LOG_ERROR, "bitrate not supported\n");
419 return -1;
420 }
421
422 return 0;
423 }
424
425 static int amr_nb_decode_close(AVCodecContext * avctx)
426 {
427 AMRContext *s = avctx->priv_data;
428
429 Decoder_Interface_exit(s->decState);
430 return 0;
431 }
432
433 static int amr_nb_encode_close(AVCodecContext * avctx)
434 {
435 AMRContext *s = avctx->priv_data;
436
437 Encoder_Interface_exit(s->enstate);
438 av_freep(&avctx->coded_frame);
439 return 0;
440 }
441
442 static int amr_nb_decode_frame(AVCodecContext * avctx,
443 void *data, int *data_size,
444 uint8_t * buf, int buf_size)
445 {
446 AMRContext *s = avctx->priv_data;
447 uint8_t*amrData=buf;
448 static short block_size[16]={ 12, 13, 15, 17, 19, 20, 26, 31, 5, 0, 0, 0, 0, 0, 0, 0 };
449 enum Mode dec_mode;
450 int packet_size;
451
452 /* av_log(NULL,AV_LOG_DEBUG,"amr_decode_frame buf=%p buf_size=%d frameCount=%d!!\n",buf,buf_size,s->frameCount); */
453
454 dec_mode = (buf[0] >> 3) & 0x000F;
455 packet_size = block_size[dec_mode]+1;
456
457 if(packet_size > buf_size) {
458 av_log(avctx, AV_LOG_ERROR, "amr frame too short (%u, should be %u)\n", buf_size, packet_size);
459 return -1;
460 }
461
462 s->frameCount++;
463 /* av_log(NULL,AV_LOG_DEBUG,"packet_size=%d amrData= 0x%X %X %X %X\n",packet_size,amrData[0],amrData[1],amrData[2],amrData[3]); */
464 /* call decoder */
465 Decoder_Interface_Decode(s->decState, amrData, data, 0);
466 *data_size=160*2;
467
468 return packet_size;
469 }
470
471 static int amr_nb_encode_frame(AVCodecContext *avctx,
472 unsigned char *frame/*out*/, int buf_size, void *data/*in*/)
473 {
474 AMRContext *s = avctx->priv_data;
475 int written;
476
477 if((s->enc_bitrate=getBitrateMode(avctx->bit_rate))<0)
478 {
479 av_log(avctx, AV_LOG_ERROR, "bitrate not supported\n");
480 return -1;
481 }
482
483 written = Encoder_Interface_Encode(s->enstate,
484 s->enc_bitrate,
485 data,
486 frame,
487 0);
488 /* av_log(NULL,AV_LOG_DEBUG,"amr_nb_encode_frame encoded %u bytes, bitrate %u, first byte was %#02x\n",written, s->enc_bitrate, frame[0] ); */
489
490 return written;
491 }
492
493 #endif
494
495 #if defined(CONFIG_AMR_NB) || defined(CONFIG_AMR_NB_FIXED)
496
497 AVCodec amr_nb_decoder =
498 {
499 "amr_nb",
500 CODEC_TYPE_AUDIO,
501 CODEC_ID_AMR_NB,
502 sizeof(AMRContext),
503 amr_nb_decode_init,
504 NULL,
505 amr_nb_decode_close,
506 amr_nb_decode_frame,
507 };
508
509 AVCodec amr_nb_encoder =
510 {
511 "amr_nb",
512 CODEC_TYPE_AUDIO,
513 CODEC_ID_AMR_NB,
514 sizeof(AMRContext),
515 amr_nb_encode_init,
516 amr_nb_encode_frame,
517 amr_nb_encode_close,
518 NULL,
519 };
520
521 #endif
522
523 /* -----------AMR wideband ------------*/
524 #ifdef CONFIG_AMR_WB
525
526 #ifdef _TYPEDEF_H
527 //To avoid duplicate typedefs from typdef in amr-nb
528 #define typedef_h
529 #endif
530
531 #include "amrwb_float/enc_if.h"
532 #include "amrwb_float/dec_if.h"
533
534 /* Common code for fixed and float version*/
535 typedef struct AMRWB_bitrates
536 {
537 int rate;
538 int mode;
539 } AMRWB_bitrates;
540
541 static int getWBBitrateMode(int bitrate)
542 {
543 /* make the correspondance between bitrate and mode */
544 AMRWB_bitrates rates[]={ {6600,0},
545 {8850,1},
546 {12650,2},
547 {14250,3},
548 {15850,4},
549 {18250,5},
550 {19850,6},
551 {23050,7},
552 {23850,8},
553 };
554 int i;
555
556 for(i=0;i<9;i++)
557 {
558 if(rates[i].rate==bitrate)
559 {
560 return(rates[i].mode);
561 }
562 }
563 /* no bitrate matching, return an error */
564 return -1;
565 }
566
567
568 typedef struct AMRWBContext {
569 int frameCount;
570 void *state;
571 int mode;
572 Word16 allow_dtx;
573 } AMRWBContext;
574
575 static int amr_wb_encode_init(AVCodecContext * avctx)
576 {
577 AMRWBContext *s = avctx->priv_data;
578
579 s->frameCount=0;
580
581 if(avctx->sample_rate!=16000)
582 {
583 av_log(avctx, AV_LOG_ERROR, "Only 16000Hz sample rate supported\n");
584 return -1;
585 }
586
587 if(avctx->channels!=1)
588 {
589 av_log(avctx, AV_LOG_ERROR, "Only mono supported\n");
590 return -1;
591 }
592
593 if((s->mode=getWBBitrateMode(avctx->bit_rate))<0)
594 {
595 av_log(avctx, AV_LOG_ERROR, "bitrate not supported\n");
596 return -1;
597 }
598
599 avctx->frame_size=320;
600 avctx->coded_frame= avcodec_alloc_frame();
601
602 s->state = E_IF_init();
603 s->allow_dtx=0;
604
605 return 0;
606 }
607
608 static int amr_wb_encode_close(AVCodecContext * avctx)
609 {
610 AMRWBContext *s = avctx->priv_data;
611
612 E_IF_exit(s->state);
613 av_freep(&avctx->coded_frame);
614 s->frameCount++;
615 return 0;
616 }
617
618 static int amr_wb_encode_frame(AVCodecContext *avctx,
619 unsigned char *frame/*out*/, int buf_size, void *data/*in*/)
620 {
621 AMRWBContext *s = avctx->priv_data;
622 int size;
623
624 if((s->mode=getWBBitrateMode(avctx->bit_rate))<0)
625 {
626 av_log(avctx, AV_LOG_ERROR, "bitrate not supported\n");
627 return -1;
628 }
629 size = E_IF_encode(s->state, s->mode, data, frame, s->allow_dtx);
630 return size;
631 }
632
633 static int amr_wb_decode_init(AVCodecContext * avctx)
634 {
635 AMRWBContext *s = avctx->priv_data;
636
637 s->frameCount=0;
638 s->state = D_IF_init();
639
640 amr_decode_fix_avctx(avctx);
641
642 if(avctx->channels > 1)
643 {
644 av_log(avctx, AV_LOG_ERROR, "amr_wb: multichannel decoding not supported\n");
645 return -1;
646 }
647
648 return 0;
649 }
650
651 extern const UWord8 block_size[];
652
653 static int amr_wb_decode_frame(AVCodecContext * avctx,
654 void *data, int *data_size,
655 uint8_t * buf, int buf_size)
656 {
657 AMRWBContext *s = avctx->priv_data;
658 uint8_t*amrData=buf;
659 int mode;
660 int packet_size;
661
662 if(buf_size==0) {
663 /* nothing to do */
664 return 0;
665 }
666
667 mode = (amrData[0] >> 3) & 0x000F;
668 packet_size = block_size[mode];
669
670 if(packet_size > buf_size) {
671 av_log(avctx, AV_LOG_ERROR, "amr frame too short (%u, should be %u)\n", buf_size, packet_size+1);
672 return -1;
673 }
674
675 s->frameCount++;
676 D_IF_decode( s->state, amrData, data, _good_frame);
677 *data_size=320*2;
678 return packet_size;
679 }
680
681 static int amr_wb_decode_close(AVCodecContext * avctx)
682 {
683 AMRWBContext *s = avctx->priv_data;
684
685 D_IF_exit(s->state);
686 return 0;
687 }
688
689 AVCodec amr_wb_decoder =
690 {
691 "amr_wb",
692 CODEC_TYPE_AUDIO,
693 CODEC_ID_AMR_WB,
694 sizeof(AMRWBContext),
695 amr_wb_decode_init,
696 NULL,
697 amr_wb_decode_close,
698 amr_wb_decode_frame,
699 };
700
701 AVCodec amr_wb_encoder =
702 {
703 "amr_wb",
704 CODEC_TYPE_AUDIO,
705 CODEC_ID_AMR_WB,
706 sizeof(AMRWBContext),
707 amr_wb_encode_init,
708 amr_wb_encode_frame,
709 amr_wb_encode_close,
710 NULL,
711 };
712
713 #endif //CONFIG_AMR_WB