2 * AMR Audio decoder stub
3 * Copyright (c) 2003 the ffmpeg project
5 * This file is part of FFmpeg.
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.
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.
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
23 * Adaptive Multi-Rate (AMR) Audio decoder stub.
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
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".
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".
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.
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".
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.
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.
73 #ifdef CONFIG_AMR_NB_FIXED
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"
85 #include "amr_float/interf_dec.h"
86 #include "amr_float/interf_enc.h"
89 /* Common code for fixed and float version*/
90 typedef struct AMR_bitrates
96 /* Match desired bitrate */
97 static int getBitrateMode(int bitrate
)
99 /* make the correspondance between bitrate and mode */
100 AMR_bitrates rates
[]={ {4750,MR475
},
113 if(rates
[i
].rate
==bitrate
)
115 return(rates
[i
].mode
);
118 /* no bitrate matching, return an error */
122 static void amr_decode_fix_avctx(AVCodecContext
* avctx
)
124 const int is_amr_wb
= 1 + (avctx
->codec_id
== CODEC_ID_AMR_WB
);
126 if(avctx
->sample_rate
== 0)
128 avctx
->sample_rate
= 8000 * is_amr_wb
;
131 if(avctx
->channels
== 0)
136 avctx
->frame_size
= 160 * is_amr_wb
;
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)
144 typedef struct AMRContext
{
146 Speech_Decode_FrameState
*speech_decoder_state
;
147 enum RXFrameType rx_type
;
150 Word16 reset_flag_old
;
153 Speech_Encode_FrameState
*enstate
;
154 sid_syncState
*sidstate
;
155 enum TXFrameType tx_frametype
;
158 static int amr_nb_decode_init(AVCodecContext
* avctx
)
160 AMRContext
*s
= avctx
->priv_data
;
163 s
->speech_decoder_state
=NULL
;
164 s
->rx_type
= (enum RXFrameType
)0;
165 s
->mode
= (enum Mode
)0;
169 if(Speech_Decode_Frame_init(&s
->speech_decoder_state
, "Decoder"))
171 av_log(avctx
, AV_LOG_ERROR
, "Speech_Decode_Frame_init error\n");
175 amr_decode_fix_avctx(avctx
);
177 if(avctx
->channels
> 1)
179 av_log(avctx
, AV_LOG_ERROR
, "amr_nb: multichannel decoding not supported\n");
186 static int amr_nb_encode_init(AVCodecContext
* avctx
)
188 AMRContext
*s
= avctx
->priv_data
;
191 s
->speech_decoder_state
=NULL
;
192 s
->rx_type
= (enum RXFrameType
)0;
193 s
->mode
= (enum Mode
)0;
197 if(avctx
->sample_rate
!=8000)
199 av_log(avctx
, AV_LOG_ERROR
, "Only 8000Hz sample rate supported\n");
203 if(avctx
->channels
!=1)
205 av_log(avctx
, AV_LOG_ERROR
, "Only mono supported\n");
209 avctx
->frame_size
=160;
210 avctx
->coded_frame
= avcodec_alloc_frame();
212 if(Speech_Encode_Frame_init(&s
->enstate
, 0, "encoder") || sid_sync_init (&s
->sidstate
))
214 av_log(avctx
, AV_LOG_ERROR
, "Speech_Encode_Frame_init error\n");
218 if((s
->enc_bitrate
=getBitrateMode(avctx
->bit_rate
))<0)
220 av_log(avctx
, AV_LOG_ERROR
, "bitrate not supported\n");
227 static int amr_nb_encode_close(AVCodecContext
* avctx
)
229 AMRContext
*s
= avctx
->priv_data
;
231 Speech_Encode_Frame_exit(&s
->enstate
);
232 sid_sync_exit (&s
->sidstate
);
233 av_freep(&avctx
->coded_frame
);
237 static int amr_nb_decode_close(AVCodecContext
* avctx
)
239 AMRContext
*s
= avctx
->priv_data
;
241 Speech_Decode_Frame_exit(&s
->speech_decoder_state
);
245 static int amr_nb_decode_frame(AVCodecContext
* avctx
,
246 void *data
, int *data_size
,
247 uint8_t * buf
, int buf_size
)
249 AMRContext
*s
= avctx
->priv_data
;
253 Word16 serial
[SERIAL_FRAMESIZE
]; /* coded bits */
256 static Word16 packed_size
[16] = {12, 13, 15, 17, 19, 20, 26, 31, 5, 0, 0, 0, 0, 0, 0, 0};
259 //printf("amr_decode_frame data_size=%i buf=0x%X buf_size=%d frameCount=%d!!\n",*data_size,buf,buf_size,s->frameCount);
264 /* read rest of the frame based on ToC byte */
265 q
= (toc
>> 2) & 0x01;
266 ft
= (toc
>> 3) & 0x0F;
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]);
272 packed_bits
=amrData
+offset
;
274 offset
+=packed_size
[ft
];
276 //Unsort and unpack bits
277 s
->rx_type
= UnpackBits(q
, ft
, packed_bits
, &s
->mode
, &serial
[1]);
279 //We have a new frame
282 if (s
->rx_type
== RX_NO_DATA
)
284 s
->mode
= s
->speech_decoder_state
->prev_mode
;
287 s
->speech_decoder_state
->prev_mode
= s
->mode
;
290 /* if homed: check if this frame is another homing frame */
291 if (s
->reset_flag_old
== 1)
293 /* only check until end of first subframe */
294 s
->reset_flag
= decoder_homing_frame_test_first(&serial
[1], s
->mode
);
296 /* produce encoder homing frame if homed & input=decoder homing frame */
297 if ((s
->reset_flag
!= 0) && (s
->reset_flag_old
!= 0))
299 for (i
= 0; i
< L_FRAME
; i
++)
307 Speech_Decode_Frame(s
->speech_decoder_state
, s
->mode
, &serial
[1], s
->rx_type
, synth
);
310 //Each AMR-frame results in 160 16-bit samples
314 /* if not homed: check whether current frame is a homing frame */
315 if (s
->reset_flag_old
== 0)
317 /* check whole frame */
318 s
->reset_flag
= decoder_homing_frame_test(&serial
[1], s
->mode
);
320 /* reset decoder if current frame is a homing frame */
321 if (s
->reset_flag
!= 0)
323 Speech_Decode_Frame_reset(s
->speech_decoder_state
);
325 s
->reset_flag_old
= s
->reset_flag
;
331 static int amr_nb_encode_frame(AVCodecContext
*avctx
,
332 unsigned char *frame
/*out*/, int buf_size
, void *data
/*in*/)
334 short serial_data
[250] = {0};
335 AMRContext
*s
= avctx
->priv_data
;
338 s
->reset_flag
= encoder_homing_frame_test(data
);
340 Speech_Encode_Frame(s
->enstate
, s
->enc_bitrate
, data
, &serial_data
[1], &s
->mode
);
342 /* add frame type and mode */
343 sid_sync (s
->sidstate
, s
->mode
, &s
->tx_frametype
);
345 written
= PackBits(s
->mode
, s
->enc_bitrate
, s
->tx_frametype
, &serial_data
[1], frame
);
347 if (s
->reset_flag
!= 0)
349 Speech_Encode_Frame_reset(s
->enstate
);
350 sid_sync_reset(s
->sidstate
);
356 #elif defined(CONFIG_AMR_NB) /* Float point version*/
358 typedef struct AMRContext
{
365 static int amr_nb_decode_init(AVCodecContext
* avctx
)
367 AMRContext
*s
= avctx
->priv_data
;
370 s
->decState
=Decoder_Interface_init();
373 av_log(avctx
, AV_LOG_ERROR
, "Decoder_Interface_init error\r\n");
377 amr_decode_fix_avctx(avctx
);
379 if(avctx
->channels
> 1)
381 av_log(avctx
, AV_LOG_ERROR
, "amr_nb: multichannel decoding not supported\n");
388 static int amr_nb_encode_init(AVCodecContext
* avctx
)
390 AMRContext
*s
= avctx
->priv_data
;
394 if(avctx
->sample_rate
!=8000)
396 av_log(avctx
, AV_LOG_ERROR
, "Only 8000Hz sample rate supported\n");
400 if(avctx
->channels
!=1)
402 av_log(avctx
, AV_LOG_ERROR
, "Only mono supported\n");
406 avctx
->frame_size
=160;
407 avctx
->coded_frame
= avcodec_alloc_frame();
409 s
->enstate
=Encoder_Interface_init(0);
412 av_log(avctx
, AV_LOG_ERROR
, "Encoder_Interface_init error\n");
416 if((s
->enc_bitrate
=getBitrateMode(avctx
->bit_rate
))<0)
418 av_log(avctx
, AV_LOG_ERROR
, "bitrate not supported\n");
425 static int amr_nb_decode_close(AVCodecContext
* avctx
)
427 AMRContext
*s
= avctx
->priv_data
;
429 Decoder_Interface_exit(s
->decState
);
433 static int amr_nb_encode_close(AVCodecContext
* avctx
)
435 AMRContext
*s
= avctx
->priv_data
;
437 Encoder_Interface_exit(s
->enstate
);
438 av_freep(&avctx
->coded_frame
);
442 static int amr_nb_decode_frame(AVCodecContext
* avctx
,
443 void *data
, int *data_size
,
444 uint8_t * buf
, int buf_size
)
446 AMRContext
*s
= avctx
->priv_data
;
448 static short block_size
[16]={ 12, 13, 15, 17, 19, 20, 26, 31, 5, 0, 0, 0, 0, 0, 0, 0 };
452 /* av_log(NULL,AV_LOG_DEBUG,"amr_decode_frame buf=%p buf_size=%d frameCount=%d!!\n",buf,buf_size,s->frameCount); */
454 dec_mode
= (buf
[0] >> 3) & 0x000F;
455 packet_size
= block_size
[dec_mode
]+1;
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
);
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]); */
465 Decoder_Interface_Decode(s
->decState
, amrData
, data
, 0);
471 static int amr_nb_encode_frame(AVCodecContext
*avctx
,
472 unsigned char *frame
/*out*/, int buf_size
, void *data
/*in*/)
474 AMRContext
*s
= avctx
->priv_data
;
477 if((s
->enc_bitrate
=getBitrateMode(avctx
->bit_rate
))<0)
479 av_log(avctx
, AV_LOG_ERROR
, "bitrate not supported\n");
483 written
= Encoder_Interface_Encode(s
->enstate
,
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] ); */
495 #if defined(CONFIG_AMR_NB) || defined(CONFIG_AMR_NB_FIXED)
497 AVCodec amr_nb_decoder
=
509 AVCodec amr_nb_encoder
=
523 /* -----------AMR wideband ------------*/
527 //To avoid duplicate typedefs from typdef in amr-nb
531 #include "amrwb_float/enc_if.h"
532 #include "amrwb_float/dec_if.h"
534 /* Common code for fixed and float version*/
535 typedef struct AMRWB_bitrates
541 static int getWBBitrateMode(int bitrate
)
543 /* make the correspondance between bitrate and mode */
544 AMRWB_bitrates rates
[]={ {6600,0},
558 if(rates
[i
].rate
==bitrate
)
560 return(rates
[i
].mode
);
563 /* no bitrate matching, return an error */
568 typedef struct AMRWBContext
{
575 static int amr_wb_encode_init(AVCodecContext
* avctx
)
577 AMRWBContext
*s
= avctx
->priv_data
;
581 if(avctx
->sample_rate
!=16000)
583 av_log(avctx
, AV_LOG_ERROR
, "Only 16000Hz sample rate supported\n");
587 if(avctx
->channels
!=1)
589 av_log(avctx
, AV_LOG_ERROR
, "Only mono supported\n");
593 if((s
->mode
=getWBBitrateMode(avctx
->bit_rate
))<0)
595 av_log(avctx
, AV_LOG_ERROR
, "bitrate not supported\n");
599 avctx
->frame_size
=320;
600 avctx
->coded_frame
= avcodec_alloc_frame();
602 s
->state
= E_IF_init();
608 static int amr_wb_encode_close(AVCodecContext
* avctx
)
610 AMRWBContext
*s
= avctx
->priv_data
;
613 av_freep(&avctx
->coded_frame
);
618 static int amr_wb_encode_frame(AVCodecContext
*avctx
,
619 unsigned char *frame
/*out*/, int buf_size
, void *data
/*in*/)
621 AMRWBContext
*s
= avctx
->priv_data
;
624 if((s
->mode
=getWBBitrateMode(avctx
->bit_rate
))<0)
626 av_log(avctx
, AV_LOG_ERROR
, "bitrate not supported\n");
629 size
= E_IF_encode(s
->state
, s
->mode
, data
, frame
, s
->allow_dtx
);
633 static int amr_wb_decode_init(AVCodecContext
* avctx
)
635 AMRWBContext
*s
= avctx
->priv_data
;
638 s
->state
= D_IF_init();
640 amr_decode_fix_avctx(avctx
);
642 if(avctx
->channels
> 1)
644 av_log(avctx
, AV_LOG_ERROR
, "amr_wb: multichannel decoding not supported\n");
651 extern const UWord8 block_size
[];
653 static int amr_wb_decode_frame(AVCodecContext
* avctx
,
654 void *data
, int *data_size
,
655 uint8_t * buf
, int buf_size
)
657 AMRWBContext
*s
= avctx
->priv_data
;
667 mode
= (amrData
[0] >> 3) & 0x000F;
668 packet_size
= block_size
[mode
];
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);
676 D_IF_decode( s
->state
, amrData
, data
, _good_frame
);
681 static int amr_wb_decode_close(AVCodecContext
* avctx
)
683 AMRWBContext
*s
= avctx
->priv_data
;
689 AVCodec amr_wb_decoder
=
694 sizeof(AMRWBContext
),
701 AVCodec amr_wb_encoder
=
706 sizeof(AMRWBContext
),
713 #endif //CONFIG_AMR_WB