Add the glyph range to the comment. Cyrillic glyphs were taken from FS#9834 (thanks...
[kugel-rb/myfork.git] / apps / metadata / mp4.c
blobb6aeaf15813b6d6d7ba7ab2c90c47ed7eef12fb1
1 /***************************************************************************
2 * __________ __ ___.
3 * Open \______ \ ____ ____ | | _\_ |__ _______ ___
4 * Source | _// _ \_/ ___\| |/ /| __ \ / _ \ \/ /
5 * Jukebox | | ( <_> ) \___| < | \_\ ( <_> > < <
6 * Firmware |____|_ /\____/ \___ >__|_ \|___ /\____/__/\_ \
7 * \/ \/ \/ \/ \/
8 * $Id$
10 * Copyright (C) 2005 Magnus Holmgren
12 * This program is free software; you can redistribute it and/or
13 * modify it under the terms of the GNU General Public License
14 * as published by the Free Software Foundation; either version 2
15 * of the License, or (at your option) any later version.
17 * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
18 * KIND, either express or implied.
20 ****************************************************************************/
21 #include <stdio.h>
22 #include <string.h>
23 #include <stdlib.h>
24 #include <ctype.h>
25 #include <inttypes.h>
27 #include "system.h"
28 #include "errno.h"
29 #include "metadata.h"
30 #include "metadata_common.h"
31 #include "metadata_parsers.h"
32 #include "logf.h"
33 #include "debug.h"
34 #include "replaygain.h"
36 #define MP4_ID(a, b, c, d) (((a) << 24) | ((b) << 16) | ((c) << 8) | (d))
38 #define MP4_3gp6 MP4_ID('3', 'g', 'p', '6')
39 #define MP4_aART MP4_ID('a', 'A', 'R', 'T')
40 #define MP4_alac MP4_ID('a', 'l', 'a', 'c')
41 #define MP4_calb MP4_ID(0xa9, 'a', 'l', 'b')
42 #define MP4_cART MP4_ID(0xa9, 'A', 'R', 'T')
43 #define MP4_cgrp MP4_ID(0xa9, 'g', 'r', 'p')
44 #define MP4_cgen MP4_ID(0xa9, 'g', 'e', 'n')
45 #define MP4_cnam MP4_ID(0xa9, 'n', 'a', 'm')
46 #define MP4_cwrt MP4_ID(0xa9, 'w', 'r', 't')
47 #define MP4_ccmt MP4_ID(0xa9, 'c', 'm', 't')
48 #define MP4_cday MP4_ID(0xa9, 'd', 'a', 'y')
49 #define MP4_disk MP4_ID('d', 'i', 's', 'k')
50 #define MP4_esds MP4_ID('e', 's', 'd', 's')
51 #define MP4_ftyp MP4_ID('f', 't', 'y', 'p')
52 #define MP4_gnre MP4_ID('g', 'n', 'r', 'e')
53 #define MP4_hdlr MP4_ID('h', 'd', 'l', 'r')
54 #define MP4_ilst MP4_ID('i', 'l', 's', 't')
55 #define MP4_M4A MP4_ID('M', '4', 'A', ' ')
56 #define MP4_M4B MP4_ID('M', '4', 'B', ' ')
57 #define MP4_mdat MP4_ID('m', 'd', 'a', 't')
58 #define MP4_mdia MP4_ID('m', 'd', 'i', 'a')
59 #define MP4_mdir MP4_ID('m', 'd', 'i', 'r')
60 #define MP4_meta MP4_ID('m', 'e', 't', 'a')
61 #define MP4_minf MP4_ID('m', 'i', 'n', 'f')
62 #define MP4_moov MP4_ID('m', 'o', 'o', 'v')
63 #define MP4_mp4a MP4_ID('m', 'p', '4', 'a')
64 #define MP4_mp42 MP4_ID('m', 'p', '4', '2')
65 #define MP4_qt MP4_ID('q', 't', ' ', ' ')
66 #define MP4_soun MP4_ID('s', 'o', 'u', 'n')
67 #define MP4_stbl MP4_ID('s', 't', 'b', 'l')
68 #define MP4_stsd MP4_ID('s', 't', 's', 'd')
69 #define MP4_stts MP4_ID('s', 't', 't', 's')
70 #define MP4_trak MP4_ID('t', 'r', 'a', 'k')
71 #define MP4_trkn MP4_ID('t', 'r', 'k', 'n')
72 #define MP4_udta MP4_ID('u', 'd', 't', 'a')
73 #define MP4_extra MP4_ID('-', '-', '-', '-')
75 /* Read the tag data from an MP4 file, storing up to buffer_size bytes in
76 * buffer.
78 static unsigned long read_mp4_tag(int fd, unsigned int size_left, char* buffer,
79 unsigned int buffer_left)
81 unsigned int bytes_read = 0;
83 if (buffer_left == 0)
85 lseek(fd, size_left, SEEK_CUR); /* Skip everything */
87 else
89 /* Skip the data tag header - maybe we should parse it properly? */
90 lseek(fd, 16, SEEK_CUR);
91 size_left -= 16;
93 if (size_left > buffer_left)
95 read(fd, buffer, buffer_left);
96 lseek(fd, size_left - buffer_left, SEEK_CUR);
97 bytes_read = buffer_left;
99 else
101 read(fd, buffer, size_left);
102 bytes_read = size_left;
106 return bytes_read;
109 /* Read a string tag from an MP4 file */
110 static unsigned int read_mp4_tag_string(int fd, int size_left, char** buffer,
111 unsigned int* buffer_left, char** dest)
113 unsigned int bytes_read = read_mp4_tag(fd, size_left, *buffer,
114 *buffer_left > 0 ? *buffer_left - 1 : 0);
115 unsigned int length = 0;
117 if (bytes_read)
119 (*buffer)[bytes_read] = 0;
120 *dest = *buffer;
121 length = strlen(*buffer) + 1;
122 *buffer_left -= length;
123 *buffer += length;
125 else
127 *dest = NULL;
130 return length;
133 static unsigned int read_mp4_atom(int fd, unsigned int* size,
134 unsigned int* type, unsigned int size_left)
136 read_uint32be(fd, size);
137 read_uint32be(fd, type);
139 if (*size == 1)
141 /* FAT32 doesn't support files this big, so something seems to
142 * be wrong. (64-bit sizes should only be used when required.)
144 errno = EFBIG;
145 *type = 0;
146 return 0;
149 if (*size > 0)
151 if (*size > size_left)
153 size_left = 0;
155 else
157 size_left -= *size;
160 *size -= 8;
162 else
164 *size = size_left;
165 size_left = 0;
168 return size_left;
171 static unsigned int read_mp4_length(int fd, unsigned int* size)
173 unsigned int length = 0;
174 int bytes = 0;
175 unsigned char c;
179 read(fd, &c, 1);
180 bytes++;
181 (*size)--;
182 length = (length << 7) | (c & 0x7F);
184 while ((c & 0x80) && (bytes < 4) && (*size > 0));
186 return length;
189 static bool read_mp4_esds(int fd, struct mp3entry* id3,
190 unsigned int* size)
192 unsigned char buf[8];
193 bool sbr = false;
195 lseek(fd, 4, SEEK_CUR); /* Version and flags. */
196 read(fd, buf, 1); /* Verify ES_DescrTag. */
197 *size -= 5;
199 if (*buf == 3)
201 /* read length */
202 if (read_mp4_length(fd, size) < 20)
204 return sbr;
207 lseek(fd, 3, SEEK_CUR);
208 *size -= 3;
210 else
212 lseek(fd, 2, SEEK_CUR);
213 *size -= 2;
216 read(fd, buf, 1); /* Verify DecoderConfigDescrTab. */
217 *size -= 1;
219 if (*buf != 4)
221 return sbr;
224 if (read_mp4_length(fd, size) < 13)
226 return sbr;
229 lseek(fd, 13, SEEK_CUR); /* Skip audio type, bit rates, etc. */
230 read(fd, buf, 1);
231 *size -= 14;
233 if (*buf != 5) /* Verify DecSpecificInfoTag. */
235 return sbr;
239 static const int sample_rates[] =
241 96000, 88200, 64000, 48000, 44100, 32000,
242 24000, 22050, 16000, 12000, 11025, 8000
244 unsigned long bits;
245 unsigned int length;
246 unsigned int index;
247 unsigned int type;
249 /* Read the (leading part of the) decoder config. */
250 length = read_mp4_length(fd, size);
251 length = MIN(length, *size);
252 length = MIN(length, sizeof(buf));
253 memset(buf, 0, sizeof(buf));
254 read(fd, buf, length);
255 *size -= length;
257 /* Maybe time to write a simple read_bits function... */
259 /* Decoder config format:
260 * Object type - 5 bits
261 * Frequency index - 4 bits
262 * Channel configuration - 4 bits
264 bits = get_long_be(buf);
265 type = bits >> 27; /* Object type - 5 bits */
266 index = (bits >> 23) & 0xf; /* Frequency index - 4 bits */
268 if (index < (sizeof(sample_rates) / sizeof(*sample_rates)))
270 id3->frequency = sample_rates[index];
273 if (type == 5)
275 DEBUGF("MP4: SBR\n");
276 unsigned int old_index = index;
278 sbr = true;
279 index = (bits >> 15) & 0xf; /* Frequency index - 4 bits */
281 if (index == 15)
283 /* 17 bits read so far... */
284 bits = get_long_be(&buf[2]);
285 id3->frequency = (bits >> 7) & 0x00ffffff;
287 else if (index < (sizeof(sample_rates) / sizeof(*sample_rates)))
289 id3->frequency = sample_rates[index];
292 if (old_index == index)
294 /* Downsampled SBR */
295 id3->frequency *= 2;
298 /* Skip 13 bits from above, plus 3 bits, then read 11 bits */
299 else if ((length >= 4) && (((bits >> 5) & 0x7ff) == 0x2b7))
301 /* extensionAudioObjectType */
302 DEBUGF("MP4: extensionAudioType\n");
303 type = bits & 0x1f; /* Object type - 5 bits*/
304 bits = get_long_be(&buf[4]);
306 if (type == 5)
308 sbr = bits >> 31;
310 if (sbr)
312 unsigned int old_index = index;
314 /* 1 bit read so far */
315 index = (bits >> 27) & 0xf; /* Frequency index - 4 bits */
317 if (index == 15)
319 /* 5 bits read so far */
320 id3->frequency = (bits >> 3) & 0x00ffffff;
322 else if (index < (sizeof(sample_rates) / sizeof(*sample_rates)))
324 id3->frequency = sample_rates[index];
327 if (old_index == index)
329 /* Downsampled SBR */
330 id3->frequency *= 2;
336 if (!sbr && (id3->frequency <= 24000) && (length <= 2))
338 /* Double the frequency for low-frequency files without a "long"
339 * DecSpecificConfig header. The file may or may not contain SBR,
340 * but here we guess it does if the header is short. This can
341 * fail on some files, but it's the best we can do, short of
342 * decoding (parts of) the file.
344 id3->frequency *= 2;
348 return sbr;
351 static bool read_mp4_tags(int fd, struct mp3entry* id3,
352 unsigned int size_left)
354 unsigned int size;
355 unsigned int type;
356 unsigned int buffer_left = sizeof(id3->id3v2buf) + sizeof(id3->id3v1buf);
357 char* buffer = id3->id3v2buf;
358 bool cwrt = false;
362 size_left = read_mp4_atom(fd, &size, &type, size_left);
364 /* DEBUGF("Tag atom: '%c%c%c%c' (%d bytes left)\n", type >> 24 & 0xff,
365 type >> 16 & 0xff, type >> 8 & 0xff, type & 0xff, size); */
367 switch (type)
369 case MP4_cnam:
370 read_mp4_tag_string(fd, size, &buffer, &buffer_left,
371 &id3->title);
372 break;
374 case MP4_cART:
375 read_mp4_tag_string(fd, size, &buffer, &buffer_left,
376 &id3->artist);
377 break;
379 case MP4_aART:
380 read_mp4_tag_string(fd, size, &buffer, &buffer_left,
381 &id3->albumartist);
382 break;
384 case MP4_cgrp:
385 read_mp4_tag_string(fd, size, &buffer, &buffer_left,
386 &id3->grouping);
387 break;
389 case MP4_calb:
390 read_mp4_tag_string(fd, size, &buffer, &buffer_left,
391 &id3->album);
392 break;
394 case MP4_cwrt:
395 read_mp4_tag_string(fd, size, &buffer, &buffer_left,
396 &id3->composer);
397 cwrt = false;
398 break;
400 case MP4_ccmt:
401 read_mp4_tag_string(fd, size, &buffer, &buffer_left,
402 &id3->comment);
403 break;
405 case MP4_cday:
406 read_mp4_tag_string(fd, size, &buffer, &buffer_left,
407 &id3->year_string);
409 /* Try to parse it as a year, for the benefit of the database.
411 if(id3->year_string)
413 id3->year = atoi(id3->year_string);
414 if (id3->year < 1900)
416 id3->year = 0;
419 else
420 id3->year = 0;
422 break;
424 case MP4_gnre:
426 unsigned short genre;
428 read_mp4_tag(fd, size, (char*) &genre, sizeof(genre));
429 id3->genre_string = id3_get_num_genre(betoh16(genre) - 1);
431 break;
433 case MP4_cgen:
434 read_mp4_tag_string(fd, size, &buffer, &buffer_left,
435 &id3->genre_string);
436 break;
438 case MP4_disk:
440 unsigned short n[2];
442 read_mp4_tag(fd, size, (char*) &n, sizeof(n));
443 id3->discnum = betoh16(n[1]);
445 break;
447 case MP4_trkn:
449 unsigned short n[2];
451 read_mp4_tag(fd, size, (char*) &n, sizeof(n));
452 id3->tracknum = betoh16(n[1]);
454 break;
456 case MP4_extra:
458 char tag_name[TAG_NAME_LENGTH];
459 unsigned int sub_size;
461 /* "mean" atom */
462 read_uint32be(fd, &sub_size);
463 size -= sub_size;
464 lseek(fd, sub_size - 4, SEEK_CUR);
465 /* "name" atom */
466 read_uint32be(fd, &sub_size);
467 size -= sub_size;
468 lseek(fd, 8, SEEK_CUR);
469 sub_size -= 12;
471 if (sub_size > sizeof(tag_name) - 1)
473 read(fd, tag_name, sizeof(tag_name) - 1);
474 lseek(fd, sub_size - (sizeof(tag_name) - 1), SEEK_CUR);
475 tag_name[sizeof(tag_name) - 1] = 0;
477 else
479 read(fd, tag_name, sub_size);
480 tag_name[sub_size] = 0;
483 if ((strcasecmp(tag_name, "composer") == 0) && !cwrt)
485 read_mp4_tag_string(fd, size, &buffer, &buffer_left,
486 &id3->composer);
488 else if (strcasecmp(tag_name, "iTunSMPB") == 0)
490 char value[TAG_VALUE_LENGTH];
491 char* value_p = value;
492 char* any;
493 unsigned int length = sizeof(value);
495 read_mp4_tag_string(fd, size, &value_p, &length, &any);
496 id3->lead_trim = get_itunes_int32(value, 1);
497 id3->tail_trim = get_itunes_int32(value, 2);
498 DEBUGF("AAC: lead_trim %d, tail_trim %d\n",
499 id3->lead_trim, id3->tail_trim);
501 else if (strcasecmp(tag_name, "musicbrainz track id") == 0)
503 read_mp4_tag_string(fd, size, &buffer, &buffer_left,
504 &id3->mb_track_id);
506 else if ((strcasecmp(tag_name, "album artist") == 0))
508 read_mp4_tag_string(fd, size, &buffer, &buffer_left,
509 &id3->albumartist);
511 else
513 char* any;
514 unsigned int length = read_mp4_tag_string(fd, size,
515 &buffer, &buffer_left, &any);
517 if (length > 0)
519 /* Re-use the read buffer as the dest buffer... */
520 buffer -= length;
521 buffer_left += length;
523 if (parse_replaygain(tag_name, buffer, id3,
524 buffer, buffer_left) > 0)
526 /* Data used, keep it. */
527 buffer += length;
528 buffer_left -= length;
533 break;
535 default:
536 lseek(fd, size, SEEK_CUR);
537 break;
540 while ((size_left > 0) && (errno == 0));
542 return true;
545 static bool read_mp4_container(int fd, struct mp3entry* id3,
546 unsigned int size_left)
548 unsigned int size;
549 unsigned int type;
550 unsigned int handler = 0;
551 bool rc = true;
555 size_left = read_mp4_atom(fd, &size, &type, size_left);
557 /* DEBUGF("Atom: '%c%c%c%c' (0x%08x, %d bytes left)\n",
558 (type >> 24) & 0xff, (type >> 16) & 0xff, (type >> 8) & 0xff,
559 type & 0xff, type, size); */
561 switch (type)
563 case MP4_ftyp:
565 unsigned int id;
567 read_uint32be(fd, &id);
568 size -= 4;
570 if ((id != MP4_M4A) && (id != MP4_M4B) && (id != MP4_mp42)
571 && (id != MP4_qt) && (id != MP4_3gp6))
573 DEBUGF("Unknown MP4 file type: '%c%c%c%c'\n",
574 id >> 24 & 0xff, id >> 16 & 0xff, id >> 8 & 0xff,
575 id & 0xff);
576 return false;
579 break;
581 case MP4_meta:
582 lseek(fd, 4, SEEK_CUR); /* Skip version */
583 size -= 4;
584 /* Fall through */
586 case MP4_moov:
587 case MP4_udta:
588 case MP4_mdia:
589 case MP4_stbl:
590 case MP4_trak:
591 rc = read_mp4_container(fd, id3, size);
592 size = 0;
593 break;
595 case MP4_ilst:
596 if (handler == MP4_mdir)
598 rc = read_mp4_tags(fd, id3, size);
599 size = 0;
601 break;
603 case MP4_minf:
604 if (handler == MP4_soun)
606 rc = read_mp4_container(fd, id3, size);
607 size = 0;
609 break;
611 case MP4_stsd:
612 lseek(fd, 8, SEEK_CUR);
613 size -= 8;
614 rc = read_mp4_container(fd, id3, size);
615 size = 0;
616 break;
618 case MP4_hdlr:
619 lseek(fd, 8, SEEK_CUR);
620 read_uint32be(fd, &handler);
621 size -= 12;
622 /* DEBUGF(" Handler '%c%c%c%c'\n", handler >> 24 & 0xff,
623 handler >> 16 & 0xff, handler >> 8 & 0xff,handler & 0xff); */
624 break;
626 case MP4_stts:
628 unsigned int entries;
629 unsigned int i;
631 lseek(fd, 4, SEEK_CUR);
632 read_uint32be(fd, &entries);
633 id3->samples = 0;
635 for (i = 0; i < entries; i++)
637 unsigned int n;
638 unsigned int l;
640 read_uint32be(fd, &n);
641 read_uint32be(fd, &l);
642 id3->samples += n * l;
645 size = 0;
647 break;
649 case MP4_mp4a:
650 case MP4_alac:
652 unsigned int frequency;
654 id3->codectype = (type == MP4_mp4a) ? AFMT_AAC : AFMT_ALAC;
655 lseek(fd, 22, SEEK_CUR);
656 read_uint32be(fd, &frequency);
657 size -= 26;
658 id3->frequency = frequency;
660 if (type == MP4_mp4a)
662 unsigned int subsize;
663 unsigned int subtype;
665 /* Get frequency from the decoder info tag, if possible. */
666 lseek(fd, 2, SEEK_CUR);
667 /* The esds atom is a part of the mp4a atom, so ignore
668 * the returned size (it's already accounted for).
670 read_mp4_atom(fd, &subsize, &subtype, size);
671 size -= 10;
673 if (subtype == MP4_esds)
675 read_mp4_esds(fd, id3, &size);
679 break;
681 case MP4_mdat:
682 id3->filesize = size;
683 break;
685 default:
686 break;
689 lseek(fd, size, SEEK_CUR);
691 while (rc && (size_left > 0) && (errno == 0) && (id3->filesize == 0));
692 /* Break on non-zero filesize, since Rockbox currently doesn't support
693 * metadata after the mdat atom (which sets the filesize field).
696 return rc;
699 bool get_mp4_metadata(int fd, struct mp3entry* id3)
701 id3->codectype = AFMT_UNKNOWN;
702 id3->filesize = 0;
703 errno = 0;
705 if (read_mp4_container(fd, id3, filesize(fd)) && (errno == 0)
706 && (id3->samples > 0) && (id3->frequency > 0)
707 && (id3->filesize > 0))
709 if (id3->codectype == AFMT_UNKNOWN)
711 logf("Not an ALAC or AAC file");
712 return false;
715 id3->length = ((int64_t) id3->samples * 1000) / id3->frequency;
717 if (id3->length <= 0)
719 logf("mp4 length invalid!");
720 return false;
723 id3->bitrate = ((int64_t) id3->filesize * 8) / id3->length;
724 DEBUGF("MP4 bitrate %d, frequency %ld Hz, length %ld ms\n",
725 id3->bitrate, id3->frequency, id3->length);
727 else
729 logf("MP4 metadata error");
730 DEBUGF("MP4 metadata error. errno %d, frequency %ld, filesize %ld\n",
731 errno, id3->frequency, id3->filesize);
732 return false;
735 return true;