2 * Copyright (C) 2013 nu774
3 * For conditions of distribution and use, see copyright notice in COPYING
12 # include <inttypes.h>
13 #elif defined(_MSC_VER)
14 # define SCNd64 "I64d"
32 #define WIN32_LEAN_AND_MEAN
36 #include "wav_reader.h"
43 #define PROGNAME "fdkaac"
45 static volatile int g_interrupted
= 0;
48 static void signal_handler(int signum
)
52 static void handle_signals(void)
54 int i
, sigs
[] = { SIGINT
, SIGHUP
, SIGTERM
};
55 for (i
= 0; i
< sizeof(sigs
)/sizeof(sigs
[0]); ++i
) {
56 struct sigaction sa
= { 0 };
57 sa
.sa_handler
= signal_handler
;
58 sa
.sa_flags
|= SA_RESTART
;
59 sigaction(sigs
[i
], &sa
, 0);
63 static BOOL WINAPI
signal_handler(DWORD type
)
69 static void handle_signals(void)
71 SetConsoleCtrlHandler(signal_handler
, TRUE
);
74 static void handle_signals(void)
80 int read_callback(void *cookie
, void *data
, uint32_t size
)
82 size_t rc
= fread(data
, 1, size
, (FILE*)cookie
);
83 return ferror((FILE*)cookie
) ? -1 : (int)rc
;
87 int write_callback(void *cookie
, const void *data
, uint32_t size
)
89 size_t rc
= fwrite(data
, 1, size
, (FILE*)cookie
);
90 return ferror((FILE*)cookie
) ? -1 : (int)rc
;
94 int seek_callback(void *cookie
, int64_t off
, int whence
)
96 return fseeko((FILE*)cookie
, off
, whence
);
100 int64_t tell_callback(void *cookie
)
102 return ftello((FILE*)cookie
);
110 "Usage: " PROGNAME
" [options] input_file\n"
112 " -h, --help Print this help message\n"
113 " -p, --profile <n> Profile (audio object type)\n"
114 " 2: MPEG-4 AAC LC (default)\n"
115 " 5: MPEG-4 HE-AAC (SBR)\n"
116 " 29: MPEG-4 HE-AAC v2 (SBR+PS)\n"
117 " 23: MPEG-4 AAC LD\n"
118 " 39: MPEG-4 AAC ELD\n"
119 " 129: MPEG-2 AAC LC\n"
120 " 132: MPEG-2 HE-AAC (SBR)\n"
121 " 156: MPEG-2 HE-AAC v2 (SBR+PS)\n"
122 " -b, --bitrate <n> Bitrate in bits per seconds (for CBR)\n"
123 " -m, --bitrate-mode <n> Bitrate configuration\n"
124 " 0: CBR (default)\n"
126 " (VBR mode is not officially supported, and\n"
127 " works only on a certain combination of\n"
128 " parameter settings, sample rate, and\n"
129 " channel configuration)\n"
130 " -w, --bandwidth <n> Frequency bandwidth in Hz (AAC LC only)\n"
131 " -a, --afterburner <n> Afterburner\n"
134 " -L, --lowdelay-sbr Enable ELD-SBR (AAC ELD only)\n"
135 " -s, --sbr-signaling <n> SBR signaling mode\n"
136 " 0: Implicit, backward compatible(default)\n"
137 " 1: Explicit SBR and implicit PS\n"
138 " 2: Explicit hierarchical signaling\n"
139 " -f, --transport-format <n> Transport format\n"
140 " 0: RAW (default, muxed into M4A)\n"
145 " 10: LOAS/LATM (LATM within LOAS)\n"
146 " -C, --adts-crc-check Add CRC protection on ADTS header\n"
147 " -h, --header-period <n> StreamMuxConfig/PCE repetition period in\n"
150 " -o <filename> Output filename\n"
151 " --ignorelength Ignore length of WAV header\n"
152 " -S, --silent Don't print progress messages\n"
154 "Options for raw (headerless) input:\n"
155 " -R, --raw Treat input as raw (by default WAV is\n"
157 " --raw-channels <n> Number of channels (default: 2)\n"
158 " --raw-rate <n> Sample rate (default: 44100)\n"
159 " --raw-format <spec> Sample format, default is \"S16L\".\n"
160 " Spec is as follows:\n"
161 " 1st char: S(igned)|U(nsigned)|F(loat)\n"
162 " 2nd part: bits per channel\n"
163 " Last char: L(ittle)|B(ig)\n"
164 " Last char can be omitted, in which case L is\n"
165 " assumed. Spec is case insensitive, therefore\n"
166 " \"u16b\" is same as \"U16B\".\n"
169 " --title <string>\n"
170 " --artist <string>\n"
171 " --album <string>\n"
172 " --genre <string>\n"
174 " --composer <string>\n"
175 " --grouping <string>\n"
176 " --comment <string>\n"
177 " --album-artist <string>\n"
178 " --track <number[/total]>\n"
179 " --disk <number[/total]>\n"
181 " --tag <fcc>:<value> Set iTunes predefined tag with four char code.\n"
182 " --tag-from-file <fcc>:<filename>\n"
183 " Same as above, but value is read from file.\n"
184 " --long-tag <name>:<value> Set arbitrary tag as iTunes custom metadata.\n"
185 " --tag-from-json <filename[?dot_notation]>\n"
186 " Read tags from JSON. By default, tags are\n"
187 " assumed to be direct children of the root\n"
188 " object(dictionary).\n"
189 " Optionally, position of the dictionary\n"
190 " that contains tags can be specified with\n"
191 " dotted notation.\n"
193 " --tag-from-json /path/to/json?format.tags\n"
197 typedef struct aacenc_param_ex_t
{
200 char *input_filename
;
201 char *output_filename
;
202 unsigned ignore_length
;
206 unsigned raw_channels
;
208 const char *raw_format
;
210 aacenc_tag_param_t tags
;
216 int parse_options(int argc
, char **argv
, aacenc_param_ex_t
*params
)
221 #define OPT_RAW_CHANNELS M4AF_FOURCC('r','c','h','n')
222 #define OPT_RAW_RATE M4AF_FOURCC('r','r','a','t')
223 #define OPT_RAW_FORMAT M4AF_FOURCC('r','f','m','t')
224 #define OPT_SHORT_TAG M4AF_FOURCC('s','t','a','g')
225 #define OPT_SHORT_TAG_FILE M4AF_FOURCC('s','t','g','f')
226 #define OPT_LONG_TAG M4AF_FOURCC('l','t','a','g')
227 #define OPT_TAG_FROM_JSON M4AF_FOURCC('t','f','j','s')
229 static struct option long_options
[] = {
230 { "help", no_argument
, 0, 'h' },
231 { "profile", required_argument
, 0, 'p' },
232 { "bitrate", required_argument
, 0, 'b' },
233 { "bitrate-mode", required_argument
, 0, 'm' },
234 { "bandwidth", required_argument
, 0, 'w' },
235 { "afterburner", required_argument
, 0, 'a' },
236 { "lowdelay-sbr", no_argument
, 0, 'L' },
237 { "sbr-signaling", required_argument
, 0, 's' },
238 { "transport-format", required_argument
, 0, 'f' },
239 { "adts-crc-check", no_argument
, 0, 'C' },
240 { "header-period", required_argument
, 0, 'P' },
242 { "ignorelength", no_argument
, 0, 'I' },
243 { "silent", no_argument
, 0, 'S' },
245 { "raw", no_argument
, 0, 'R' },
246 { "raw-channels", required_argument
, 0, OPT_RAW_CHANNELS
},
247 { "raw-rate", required_argument
, 0, OPT_RAW_RATE
},
248 { "raw-format", required_argument
, 0, OPT_RAW_FORMAT
},
250 { "title", required_argument
, 0, M4AF_TAG_TITLE
},
251 { "artist", required_argument
, 0, M4AF_TAG_ARTIST
},
252 { "album", required_argument
, 0, M4AF_TAG_ALBUM
},
253 { "genre", required_argument
, 0, M4AF_TAG_GENRE
},
254 { "date", required_argument
, 0, M4AF_TAG_DATE
},
255 { "composer", required_argument
, 0, M4AF_TAG_COMPOSER
},
256 { "grouping", required_argument
, 0, M4AF_TAG_GROUPING
},
257 { "comment", required_argument
, 0, M4AF_TAG_COMMENT
},
258 { "album-artist", required_argument
, 0, M4AF_TAG_ALBUM_ARTIST
},
259 { "track", required_argument
, 0, M4AF_TAG_TRACK
},
260 { "disk", required_argument
, 0, M4AF_TAG_DISK
},
261 { "tempo", required_argument
, 0, M4AF_TAG_TEMPO
},
262 { "tag", required_argument
, 0, OPT_SHORT_TAG
},
263 { "tag-from-file", required_argument
, 0, OPT_SHORT_TAG_FILE
},
264 { "long-tag", required_argument
, 0, OPT_LONG_TAG
},
265 { "tag-from-json", required_argument
, 0, OPT_TAG_FROM_JSON
},
268 params
->afterburner
= 1;
270 aacenc_getmainargs(&argc
, &argv
);
271 while ((ch
= getopt_long(argc
, argv
, "hp:b:m:w:a:Ls:f:CP:Io:SR",
272 long_options
, 0)) != EOF
) {
277 if (sscanf(optarg
, "%u", &n
) != 1) {
278 fprintf(stderr
, "invalid arg for profile\n");
284 if (sscanf(optarg
, "%u", &n
) != 1) {
285 fprintf(stderr
, "invalid arg for bitrate\n");
291 if (sscanf(optarg
, "%u", &n
) != 1 || n
> 5) {
292 fprintf(stderr
, "invalid arg for bitrate-mode\n");
295 params
->bitrate_mode
= n
;
298 if (sscanf(optarg
, "%u", &n
) != 1) {
299 fprintf(stderr
, "invalid arg for bandwidth\n");
302 params
->bandwidth
= n
;
305 if (sscanf(optarg
, "%u", &n
) != 1 || n
> 1) {
306 fprintf(stderr
, "invalid arg for afterburner\n");
309 params
->afterburner
= n
;
312 params
->lowdelay_sbr
= 1;
315 if (sscanf(optarg
, "%u", &n
) != 1 || n
> 2) {
316 fprintf(stderr
, "invalid arg for sbr-signaling\n");
319 params
->sbr_signaling
= n
;
322 if (sscanf(optarg
, "%u", &n
) != 1) {
323 fprintf(stderr
, "invalid arg for transport-format\n");
326 params
->transport_format
= n
;
329 params
->adts_crc_check
= 1;
332 if (sscanf(optarg
, "%u", &n
) != 1) {
333 fprintf(stderr
, "invalid arg for header-period\n");
336 params
->header_period
= n
;
339 params
->output_filename
= optarg
;
342 params
->ignore_length
= 1;
350 case OPT_RAW_CHANNELS
:
351 if (sscanf(optarg
, "%u", &n
) != 1) {
352 fprintf(stderr
, "invalid arg for raw-channels\n");
355 params
->raw_channels
= n
;
358 if (sscanf(optarg
, "%u", &n
) != 1) {
359 fprintf(stderr
, "invalid arg for raw-rate\n");
362 params
->raw_rate
= n
;
365 params
->raw_format
= optarg
;
368 case M4AF_TAG_ARTIST
:
372 case M4AF_TAG_COMPOSER
:
373 case M4AF_TAG_GROUPING
:
374 case M4AF_TAG_COMMENT
:
375 case M4AF_TAG_ALBUM_ARTIST
:
379 aacenc_param_add_itmf_entry(¶ms
->tags
, ch
, 0, optarg
,
383 case OPT_SHORT_TAG_FILE
:
388 unsigned fcc
= M4AF_FOURCC('-','-','-','-');
390 if ((val
= strchr(optarg
, ':')) == 0) {
391 fprintf(stderr
, "invalid arg for tag\n");
395 if (ch
== OPT_SHORT_TAG
|| ch
== OPT_SHORT_TAG_FILE
) {
397 * take care of U+00A9(COPYRIGHT SIGN).
398 * 1) if length of fcc is 3, we prepend '\xa9'.
399 * 2) U+00A9 becomes "\xc2\xa9" in UTF-8. Therefore
400 * we remove first '\xc2'.
402 if (optarg
[0] == '\xc2')
404 if ((klen
= strlen(optarg
))== 3)
406 else if (klen
!= 4) {
407 fprintf(stderr
, "invalid arg for tag\n");
410 for (; *optarg
; ++optarg
)
411 fcc
= ((fcc
<< 8) | (*optarg
& 0xff));
413 aacenc_param_add_itmf_entry(¶ms
->tags
, fcc
, optarg
,
415 ch
== OPT_SHORT_TAG_FILE
);
418 case OPT_TAG_FROM_JSON
:
419 params
->json_filename
= optarg
;
428 if (!params
->bitrate
&& !params
->bitrate_mode
) {
429 fprintf(stderr
, "bitrate or bitrate-mode is mandatory\n");
432 if (params
->output_filename
&& !strcmp(params
->output_filename
, "-") &&
433 !params
->transport_format
) {
434 fprintf(stderr
, "stdout streaming is not available on M4A output\n");
437 if (params
->bitrate
&& params
->bitrate
< 10000)
438 params
->bitrate
*= 1000;
440 if (params
->is_raw
) {
441 if (!params
->raw_channels
)
442 params
->raw_channels
= 2;
443 if (!params
->raw_rate
)
444 params
->raw_rate
= 44100;
445 if (!params
->raw_format
)
446 params
->raw_format
= "S16L";
448 params
->input_filename
= argv
[optind
];
453 int write_sample(FILE *ofp
, m4af_ctx_t
*m4af
,
454 const void *data
, uint32_t size
, uint32_t duration
)
457 fwrite(data
, 1, size
, ofp
);
459 fprintf(stderr
, "ERROR: fwrite(): %s\n", strerror(errno
));
462 } else if (m4af_write_sample(m4af
, 0, data
, size
, duration
) < 0) {
463 fprintf(stderr
, "ERROR: failed to write m4a sample\n");
470 int encode(wav_reader_t
*wavf
, HANDLE_AACENCODER encoder
,
471 uint32_t frame_length
, FILE *ofp
, m4af_ctx_t
*m4af
,
476 uint32_t pcmsize
= 0;
483 int frames_written
= 0;
484 aacenc_progress_t progress
= { 0 };
485 const pcm_sample_description_t
*format
= wav_get_format(wavf
);
487 ibuf
= malloc(frame_length
* format
->bytes_per_frame
);
488 aacenc_progress_init(&progress
, wav_get_length(wavf
), format
->sample_rate
);
493 if ((nread
= wav_read_frames(wavf
, ibuf
, frame_length
)) < 0) {
494 fprintf(stderr
, "ERROR: read failed\n");
496 } else if (nread
> 0) {
497 if (pcm_convert_to_native_sint16(format
, ibuf
, nread
,
498 &pcmbuf
, &pcmsize
) < 0) {
499 fprintf(stderr
, "ERROR: unsupported sample format\n");
504 aacenc_progress_update(&progress
, wav_get_position(wavf
),
505 format
->sample_rate
* 2);
507 if ((consumed
= aac_encode_frame(encoder
, format
, pcmbuf
, nread
,
508 &obuf
, &olen
, &osize
)) < 0)
511 if (write_sample(ofp
, m4af
, obuf
, olen
, frame_length
) < 0)
515 } while (nread
> 0 || olen
> 0);
518 aacenc_progress_finish(&progress
, wav_get_position(wavf
));
521 if (ibuf
) free(ibuf
);
522 if (pcmbuf
) free(pcmbuf
);
523 if (obuf
) free(obuf
);
528 void put_tool_tag(m4af_ctx_t
*m4af
, const aacenc_param_ex_t
*params
,
529 HANDLE_AACENCODER encoder
)
533 LIB_INFO
*lib_info
= 0;
535 p
+= sprintf(p
, PROGNAME
" %s, ", fdkaac_version
);
537 lib_info
= calloc(FDK_MODULE_LAST
, sizeof(LIB_INFO
));
538 if (aacEncGetLibInfo(lib_info
) == AACENC_OK
) {
540 for (i
= 0; i
< FDK_MODULE_LAST
; ++i
)
541 if (lib_info
[i
].module_id
== FDK_AACENC
)
543 p
+= sprintf(p
, "libfdk-aac %s, ", lib_info
[i
].versionStr
);
546 if (params
->bitrate_mode
)
547 sprintf(p
, "VBR mode %d", params
->bitrate_mode
);
549 sprintf(p
, "CBR %dkbps",
550 aacEncoder_GetParam(encoder
, AACENC_BITRATE
) / 1000);
552 m4af_add_itmf_string_tag(m4af
, M4AF_TAG_TOOL
, tool_info
);
556 int finalize_m4a(m4af_ctx_t
*m4af
, const aacenc_param_ex_t
*params
,
557 HANDLE_AACENCODER encoder
)
560 aacenc_tag_entry_t
*tag
= params
->tags
.tag_table
;
562 if (params
->json_filename
)
563 aacenc_put_tags_from_json(m4af
, params
->json_filename
);
565 for (i
= 0; i
< params
->tags
.tag_count
; ++i
, ++tag
)
566 aacenc_put_tag_entry(m4af
, tag
);
568 put_tool_tag(m4af
, params
, encoder
);
570 if (m4af_finalize(m4af
) < 0) {
571 fprintf(stderr
, "ERROR: failed to finalize m4a\n");
578 char *generate_output_filename(const char *filename
, const char *ext
)
581 size_t ext_len
= strlen(ext
);
583 if (strcmp(filename
, "-") == 0) {
584 p
= malloc(ext_len
+ 6);
585 sprintf(p
, "stdin%s", ext
);
587 const char *base
= aacenc_basename(filename
);
588 size_t ilen
= strlen(base
);
589 const char *ext_org
= strrchr(base
, '.');
590 if (ext_org
) ilen
= ext_org
- base
;
591 p
= malloc(ilen
+ ext_len
+ 1);
592 sprintf(p
, "%.*s%s", ilen
, base
, ext
);
598 int parse_raw_spec(const char *spec
, pcm_sample_description_t
*desc
)
601 unsigned char c_type
, c_endian
= 'L';
604 if (sscanf(spec
, "%c%u%c", &c_type
, &bits
, &c_endian
) < 2)
606 c_type
= toupper(c_type
);
607 c_endian
= toupper(c_endian
);
611 else if (c_type
== 'U')
613 else if (c_type
== 'F')
620 else if (c_endian
!= 'L')
623 if (c_type
== 'F' && bits
!= 32 && bits
!= 64)
625 if (c_type
!= 'F' && (bits
< 8 || bits
> 32))
628 desc
->sample_type
= type
;
629 desc
->bits_per_channel
= bits
;
633 int main(int argc
, char **argv
)
635 wav_io_context_t wav_io
= { read_callback
, seek_callback
, tell_callback
};
637 m4af_io
= { 0, write_callback
, seek_callback
, tell_callback
};
638 aacenc_param_ex_t params
= { 0 };
643 char *output_filename
= 0;
644 wav_reader_t
*wavf
= 0;
645 HANDLE_AACENCODER encoder
= 0;
646 AACENC_InfoStruct aacinfo
= { 0 };
647 m4af_ctx_t
*m4af
= 0;
648 const pcm_sample_description_t
*sample_format
;
649 int downsampled_timescale
= 0;
651 struct stat stb
= { 0 };
653 setlocale(LC_CTYPE
, "");
656 if (parse_options(argc
, argv
, ¶ms
) < 0)
659 if ((ifp
= aacenc_fopen(params
.input_filename
, "rb")) == 0) {
660 aacenc_fprintf(stderr
, "ERROR: %s: %s\n", params
.input_filename
,
664 if (fstat(fileno(ifp
), &stb
) == 0 && (stb
.st_mode
& S_IFMT
) != S_IFREG
) {
668 if (!params
.is_raw
) {
669 if ((wavf
= wav_open(&wav_io
, ifp
, params
.ignore_length
)) == 0) {
670 fprintf(stderr
, "ERROR: broken / unsupported input file\n");
674 int bytes_per_channel
;
675 pcm_sample_description_t desc
= { 0 };
676 if (parse_raw_spec(params
.raw_format
, &desc
) < 0) {
677 fprintf(stderr
, "ERROR: invalid raw-format spec\n");
680 desc
.sample_rate
= params
.raw_rate
;
681 desc
.channels_per_frame
= params
.raw_channels
;
682 bytes_per_channel
= (desc
.bits_per_channel
+ 7) / 8;
683 desc
.bytes_per_frame
= params
.raw_channels
* bytes_per_channel
;
684 if ((wavf
= raw_open(&wav_io
, ifp
, &desc
)) == 0) {
685 fprintf(stderr
, "ERROR: failed to open raw input\n");
689 sample_format
= wav_get_format(wavf
);
691 if (aacenc_init(&encoder
, (aacenc_param_t
*)¶ms
, sample_format
,
695 if (!params
.output_filename
) {
696 const char *ext
= params
.transport_format
? ".aac" : ".m4a";
697 output_filename
= generate_output_filename(params
.input_filename
, ext
);
698 params
.output_filename
= output_filename
;
701 if ((ofp
= aacenc_fopen(params
.output_filename
, "wb")) == 0) {
702 aacenc_fprintf(stderr
, "ERROR: %s: %s\n", params
.output_filename
,
707 if (!params
.transport_format
) {
709 unsigned framelen
= aacinfo
.frameLength
;
710 int sbr_mode
= aacenc_is_sbr_active((aacenc_param_t
*)¶ms
);
711 int sig_mode
= aacEncoder_GetParam(encoder
, AACENC_SIGNALING_MODE
);
712 if (sbr_mode
&& !sig_mode
)
713 downsampled_timescale
= 1;
714 scale
= sample_format
->sample_rate
>> downsampled_timescale
;
715 if ((m4af
= m4af_create(M4AF_CODEC_MP4A
, scale
, &m4af_io
, ofp
)) < 0)
717 m4af_set_decoder_specific_info(m4af
, 0, aacinfo
.confBuf
,
719 m4af_set_fixed_frame_duration(m4af
, 0,
720 framelen
>> downsampled_timescale
);
721 m4af_begin_write(m4af
);
723 frame_count
= encode(wavf
, encoder
, aacinfo
.frameLength
, ofp
, m4af
,
728 uint32_t delay
= aacinfo
.encoderDelay
;
729 int64_t frames_read
= wav_get_position(wavf
);
730 uint32_t padding
= frame_count
* aacinfo
.frameLength
731 - frames_read
- aacinfo
.encoderDelay
;
732 m4af_set_priming(m4af
, 0, delay
>> downsampled_timescale
,
733 padding
>> downsampled_timescale
);
734 if (finalize_m4a(m4af
, ¶ms
, encoder
) < 0)
739 if (wavf
) wav_teardown(&wavf
);
740 if (ifp
) fclose(ifp
);
741 if (m4af
) m4af_teardown(&m4af
);
742 if (ofp
) fclose(ofp
);
743 if (encoder
) aacEncClose(&encoder
);
744 if (output_filename
) free(output_filename
);
745 if (params
.tags
.tag_table
) free(params
.tags
.tag_table
);