3 * Copyright (c) 2003 Fabrice Bellard
5 * This file is part of Libav.
7 * Libav 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 * Libav 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 Libav; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
24 #include "libavutil/avstring.h"
25 #include "libavutil/intreadwrite.h"
26 #include "libavutil/dict.h"
27 #include "avio_internal.h"
30 const AVMetadataConv ff_id3v2_34_metadata_conv[] = {
32 { "TCOM", "composer"},
34 { "TCOP", "copyright"},
35 { "TENC", "encoded_by"},
37 { "TLAN", "language"},
39 { "TPE2", "album_artist"},
40 { "TPE3", "performer"},
42 { "TPUB", "publisher"},
48 const AVMetadataConv ff_id3v2_4_metadata_conv[] = {
51 { "TDEN", "creation_time"},
52 { "TSOA", "album-sort"},
53 { "TSOP", "artist-sort"},
54 { "TSOT", "title-sort"},
58 static const AVMetadataConv id3v2_2_metadata_conv[] = {
62 { "TEN", "encoded_by"},
64 { "TP2", "album_artist"},
65 { "TP3", "performer"},
71 const char ff_id3v2_tags[][4] = {
72 "TALB", "TBPM", "TCOM", "TCON", "TCOP", "TDLY", "TENC", "TEXT",
73 "TFLT", "TIT1", "TIT2", "TIT3", "TKEY", "TLAN", "TLEN", "TMED",
74 "TOAL", "TOFN", "TOLY", "TOPE", "TOWN", "TPE1", "TPE2", "TPE3",
75 "TPE4", "TPOS", "TPUB", "TRCK", "TRSN", "TRSO", "TSRC", "TSSE",
79 const char ff_id3v2_4_tags[][4] = {
80 "TDEN", "TDOR", "TDRC", "TDRL", "TDTG", "TIPL", "TMCL", "TMOO",
81 "TPRO", "TSOA", "TSOP", "TSOT", "TSST",
85 const char ff_id3v2_3_tags[][4] = {
86 "TDAT", "TIME", "TORY", "TRDA", "TSIZ", "TYER",
90 const char *ff_id3v2_picture_types[21] = {
92 "32x32 pixels 'file icon'",
97 "Media (e.g. label side of CD)",
98 "Lead artist/lead performer/soloist",
103 "Lyricist/text writer",
104 "Recording Location",
106 "During performance",
107 "Movie/video screen capture",
108 "A bright coloured fish",
110 "Band/artist logotype",
111 "Publisher/Studio logotype",
114 const CodecMime ff_id3v2_mime_tags[] = {
115 {"image/gif" , CODEC_ID_GIF},
116 {"image/jpeg", CODEC_ID_MJPEG},
117 {"image/jpg", CODEC_ID_MJPEG},
118 {"image/png" , CODEC_ID_PNG},
119 {"image/tiff", CODEC_ID_TIFF},
123 int ff_id3v2_match(const uint8_t *buf, const char * magic)
125 return buf[0] == magic[0] &&
126 buf[1] == magic[1] &&
127 buf[2] == magic[2] &&
130 (buf[6] & 0x80) == 0 &&
131 (buf[7] & 0x80) == 0 &&
132 (buf[8] & 0x80) == 0 &&
133 (buf[9] & 0x80) == 0;
136 int ff_id3v2_tag_len(const uint8_t * buf)
138 int len = ((buf[6] & 0x7f) << 21) +
139 ((buf[7] & 0x7f) << 14) +
140 ((buf[8] & 0x7f) << 7) +
144 len += ID3v2_HEADER_SIZE;
148 static unsigned int get_size(AVIOContext *s, int len)
152 v = (v << 7) + (avio_r8(s) & 0x7F);
157 * Free GEOB type extra metadata.
159 static void free_geobtag(void *obj)
161 ID3v2ExtraMetaGEOB *geob = obj;
162 av_free(geob->mime_type);
163 av_free(geob->file_name);
164 av_free(geob->description);
170 * Decode characters to UTF-8 according to encoding type. The decoded buffer is
171 * always null terminated. Stop reading when either *maxread bytes are read from
172 * pb or U+0000 character is found.
174 * @param dst Pointer where the address of the buffer with the decoded bytes is
175 * stored. Buffer must be freed by caller.
176 * @param maxread Pointer to maximum number of characters to read from the
177 * AVIOContext. After execution the value is decremented by the number of bytes
179 * @returns 0 if no error occurred, dst is uninitialized on error
181 static int decode_str(AVFormatContext *s, AVIOContext *pb, int encoding,
182 uint8_t **dst, int *maxread)
188 unsigned int (*get)(AVIOContext*) = avio_rb16;
191 if ((ret = avio_open_dyn_buf(&dynbuf)) < 0) {
192 av_log(s, AV_LOG_ERROR, "Error opening memory stream\n");
198 case ID3v2_ENCODING_ISO8859:
201 PUT_UTF8(ch, tmp, avio_w8(dynbuf, tmp);)
206 case ID3v2_ENCODING_UTF16BOM:
207 if ((left -= 2) < 0) {
208 av_log(s, AV_LOG_ERROR, "Cannot read BOM value, input too short\n");
209 avio_close_dyn_buf(dynbuf, dst);
211 return AVERROR_INVALIDDATA;
213 switch (avio_rb16(pb)) {
219 av_log(s, AV_LOG_ERROR, "Incorrect BOM value\n");
220 avio_close_dyn_buf(dynbuf, dst);
223 return AVERROR_INVALIDDATA;
227 case ID3v2_ENCODING_UTF16BE:
228 while ((left > 1) && ch) {
229 GET_UTF16(ch, ((left -= 2) >= 0 ? get(pb) : 0), break;)
230 PUT_UTF8(ch, tmp, avio_w8(dynbuf, tmp);)
233 left += 2; /* did not read last char from pb */
236 case ID3v2_ENCODING_UTF8:
244 av_log(s, AV_LOG_WARNING, "Unknown encoding\n");
250 avio_close_dyn_buf(dynbuf, dst);
259 static void read_ttag(AVFormatContext *s, AVIOContext *pb, int taglen, const char *key)
262 int encoding, dict_flags = AV_DICT_DONT_OVERWRITE;
268 encoding = avio_r8(pb);
269 taglen--; /* account for encoding type byte */
271 if (decode_str(s, pb, encoding, &dst, &taglen) < 0) {
272 av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", key);
276 if (!(strcmp(key, "TCON") && strcmp(key, "TCO"))
277 && (sscanf(dst, "(%d)", &genre) == 1 || sscanf(dst, "%d", &genre) == 1)
278 && genre <= ID3v1_GENRE_MAX) {
280 dst = ff_id3v1_genre_str[genre];
281 } else if (!(strcmp(key, "TXXX") && strcmp(key, "TXX"))) {
282 /* dst now contains the key, need to get value */
284 if (decode_str(s, pb, encoding, &dst, &taglen) < 0) {
285 av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", key);
289 dict_flags |= AV_DICT_DONT_STRDUP_VAL | AV_DICT_DONT_STRDUP_KEY;
292 dict_flags |= AV_DICT_DONT_STRDUP_VAL;
295 av_dict_set(&s->metadata, key, dst, dict_flags);
299 * Parse GEOB tag into a ID3v2ExtraMetaGEOB struct.
301 static void read_geobtag(AVFormatContext *s, AVIOContext *pb, int taglen, char *tag, ID3v2ExtraMeta **extra_meta)
303 ID3v2ExtraMetaGEOB *geob_data = NULL;
304 ID3v2ExtraMeta *new_extra = NULL;
311 geob_data = av_mallocz(sizeof(ID3v2ExtraMetaGEOB));
313 av_log(s, AV_LOG_ERROR, "Failed to alloc %zu bytes\n", sizeof(ID3v2ExtraMetaGEOB));
317 new_extra = av_mallocz(sizeof(ID3v2ExtraMeta));
319 av_log(s, AV_LOG_ERROR, "Failed to alloc %zu bytes\n", sizeof(ID3v2ExtraMeta));
323 /* read encoding type byte */
324 encoding = avio_r8(pb);
327 /* read MIME type (always ISO-8859) */
328 if (decode_str(s, pb, ID3v2_ENCODING_ISO8859, &geob_data->mime_type, &taglen) < 0
333 if (decode_str(s, pb, encoding, &geob_data->file_name, &taglen) < 0
337 /* read content description */
338 if (decode_str(s, pb, encoding, &geob_data->description, &taglen) < 0
343 /* save encapsulated binary data */
344 geob_data->data = av_malloc(taglen);
345 if (!geob_data->data) {
346 av_log(s, AV_LOG_ERROR, "Failed to alloc %d bytes\n", taglen);
349 if ((len = avio_read(pb, geob_data->data, taglen)) < taglen)
350 av_log(s, AV_LOG_WARNING, "Error reading GEOB frame, data truncated.\n");
351 geob_data->datasize = len;
353 geob_data->data = NULL;
354 geob_data->datasize = 0;
357 /* add data to the list */
358 new_extra->tag = "GEOB";
359 new_extra->data = geob_data;
360 new_extra->next = *extra_meta;
361 *extra_meta = new_extra;
366 av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", tag);
367 free_geobtag(geob_data);
372 static int is_number(const char *str)
374 while (*str >= '0' && *str <= '9') str++;
378 static AVDictionaryEntry* get_date_tag(AVDictionary *m, const char *tag)
380 AVDictionaryEntry *t;
381 if ((t = av_dict_get(m, tag, NULL, AV_DICT_MATCH_CASE)) &&
382 strlen(t->value) == 4 && is_number(t->value))
387 static void merge_date(AVDictionary **m)
389 AVDictionaryEntry *t;
390 char date[17] = {0}; // YYYY-MM-DD hh:mm
392 if (!(t = get_date_tag(*m, "TYER")) &&
393 !(t = get_date_tag(*m, "TYE")))
395 av_strlcpy(date, t->value, 5);
396 av_dict_set(m, "TYER", NULL, 0);
397 av_dict_set(m, "TYE", NULL, 0);
399 if (!(t = get_date_tag(*m, "TDAT")) &&
400 !(t = get_date_tag(*m, "TDA")))
402 snprintf(date + 4, sizeof(date) - 4, "-%.2s-%.2s", t->value + 2, t->value);
403 av_dict_set(m, "TDAT", NULL, 0);
404 av_dict_set(m, "TDA", NULL, 0);
406 if (!(t = get_date_tag(*m, "TIME")) &&
407 !(t = get_date_tag(*m, "TIM")))
409 snprintf(date + 10, sizeof(date) - 10, " %.2s:%.2s", t->value, t->value + 2);
410 av_dict_set(m, "TIME", NULL, 0);
411 av_dict_set(m, "TIM", NULL, 0);
415 av_dict_set(m, "date", date, 0);
418 static void free_apic(void *obj)
420 ID3v2ExtraMetaAPIC *apic = obj;
421 av_freep(&apic->data);
422 av_freep(&apic->description);
426 static void read_apic(AVFormatContext *s, AVIOContext *pb, int taglen, char *tag, ID3v2ExtraMeta **extra_meta)
430 const CodecMime *mime = ff_id3v2_mime_tags;
431 enum CodecID id = CODEC_ID_NONE;
432 ID3v2ExtraMetaAPIC *apic = NULL;
433 ID3v2ExtraMeta *new_extra = NULL;
434 int64_t end = avio_tell(pb) + taglen;
439 new_extra = av_mallocz(sizeof(*new_extra));
440 apic = av_mallocz(sizeof(*apic));
441 if (!new_extra || !apic)
448 taglen -= avio_get_str(pb, taglen, mimetype, sizeof(mimetype));
449 while (mime->id != CODEC_ID_NONE) {
450 if (!strncmp(mime->str, mimetype, sizeof(mimetype))) {
456 if (id == CODEC_ID_NONE) {
457 av_log(s, AV_LOG_WARNING, "Unknown attached picture mimetype: %s, skipping.\n", mimetype);
463 pic_type = avio_r8(pb);
465 if (pic_type < 0 || pic_type >= FF_ARRAY_ELEMS(ff_id3v2_picture_types)) {
466 av_log(s, AV_LOG_WARNING, "Unknown attached picture type %d.\n", pic_type);
469 apic->type = ff_id3v2_picture_types[pic_type];
471 /* description and picture data */
472 if (decode_str(s, pb, enc, &apic->description, &taglen) < 0) {
473 av_log(s, AV_LOG_ERROR, "Error decoding attached picture description.\n");
478 apic->data = av_malloc(taglen);
479 if (!apic->data || avio_read(pb, apic->data, taglen) != taglen)
482 new_extra->tag = "APIC";
483 new_extra->data = apic;
484 new_extra->next = *extra_meta;
485 *extra_meta = new_extra;
492 av_freep(&new_extra);
493 avio_seek(pb, end, SEEK_SET);
496 typedef struct ID3v2EMFunc {
499 void (*read)(AVFormatContext*, AVIOContext*, int, char*, ID3v2ExtraMeta **);
500 void (*free)(void *obj);
503 static const ID3v2EMFunc id3v2_extra_meta_funcs[] = {
504 { "GEO", "GEOB", read_geobtag, free_geobtag },
505 { "PIC", "APIC", read_apic, free_apic },
510 * Get the corresponding ID3v2EMFunc struct for a tag.
511 * @param isv34 Determines if v2.2 or v2.3/4 strings are used
512 * @return A pointer to the ID3v2EMFunc struct if found, NULL otherwise.
514 static const ID3v2EMFunc *get_extra_meta_func(const char *tag, int isv34)
517 while (id3v2_extra_meta_funcs[i].tag3) {
519 (isv34 ? id3v2_extra_meta_funcs[i].tag4 :
520 id3v2_extra_meta_funcs[i].tag3),
522 return &id3v2_extra_meta_funcs[i];
528 static void ff_id3v2_parse(AVFormatContext *s, int len, uint8_t version, uint8_t flags, ID3v2ExtraMeta **extra_meta)
530 int isv34, tlen, unsync;
532 int64_t next, end = avio_tell(s->pb) + len;
534 const char *reason = NULL;
537 unsigned char *buffer = NULL;
539 const ID3v2EMFunc *extra_func;
544 reason = "compression";
562 unsync = flags & 0x80;
564 if (isv34 && flags & 0x40) { /* Extended header present, just skip over it */
565 int extlen = get_size(s->pb, 4);
567 extlen -= 4; // in v2.4 the length includes the length field we just read
570 reason = "invalid extended header length";
573 avio_skip(s->pb, extlen);
576 while (len >= taghdrlen) {
577 unsigned int tflags = 0;
581 avio_read(s->pb, tag, 4);
584 tlen = avio_rb32(s->pb);
586 tlen = get_size(s->pb, 4);
587 tflags = avio_rb16(s->pb);
588 tunsync = tflags & ID3v2_FLAG_UNSYNCH;
590 avio_read(s->pb, tag, 3);
592 tlen = avio_rb24(s->pb);
594 if (tlen < 0 || tlen > len - taghdrlen) {
595 av_log(s, AV_LOG_WARNING, "Invalid size in frame %s, skipping the rest of tag.\n", tag);
598 len -= taghdrlen + tlen;
599 next = avio_tell(s->pb) + tlen;
603 av_log(s, AV_LOG_DEBUG, "Invalid empty frame %s, skipping.\n", tag);
607 if (tflags & ID3v2_FLAG_DATALEN) {
612 if (tflags & (ID3v2_FLAG_ENCRYPTION | ID3v2_FLAG_COMPRESSION)) {
613 av_log(s, AV_LOG_WARNING, "Skipping encrypted/compressed ID3v2 frame %s.\n", tag);
614 avio_skip(s->pb, tlen);
615 /* check for text tag or supported special meta tag */
616 } else if (tag[0] == 'T' || (extra_meta && (extra_func = get_extra_meta_func(tag, isv34)))) {
617 if (unsync || tunsync) {
619 av_fast_malloc(&buffer, &buffer_size, tlen);
621 av_log(s, AV_LOG_ERROR, "Failed to alloc %d bytes\n", tlen);
624 for (i = 0, j = 0; i < tlen; i++, j++) {
625 buffer[j] = avio_r8(s->pb);
626 if (j > 0 && !buffer[j] && buffer[j - 1] == 0xff) {
627 /* Unsynchronised byte, skip it */
631 ffio_init_context(&pb, buffer, j, 0, NULL, NULL, NULL, NULL);
633 pbx = &pb; // read from sync buffer
635 pbx = s->pb; // read straight from input
639 read_ttag(s, pbx, tlen, tag);
641 /* parse special meta tag */
642 extra_func->read(s, pbx, tlen, tag, extra_meta);
646 av_log(s, AV_LOG_WARNING, "invalid frame id, assuming padding");
647 avio_skip(s->pb, tlen);
650 /* Skip to end of tag */
652 avio_seek(s->pb, next, SEEK_SET);
655 if (version == 4 && flags & 0x10) /* Footer preset, always 10 bytes, skip over it */
660 av_log(s, AV_LOG_INFO, "ID3v2.%d tag skipped, cannot handle %s\n", version, reason);
661 avio_seek(s->pb, end, SEEK_SET);
666 void ff_id3v2_read(AVFormatContext *s, const char *magic, ID3v2ExtraMeta **extra_meta)
669 uint8_t buf[ID3v2_HEADER_SIZE];
674 /* save the current offset in case there's nothing to read/skip */
675 off = avio_tell(s->pb);
676 ret = avio_read(s->pb, buf, ID3v2_HEADER_SIZE);
677 if (ret != ID3v2_HEADER_SIZE)
679 found_header = ff_id3v2_match(buf, magic);
681 /* parse ID3v2 header */
682 len = ((buf[6] & 0x7f) << 21) |
683 ((buf[7] & 0x7f) << 14) |
684 ((buf[8] & 0x7f) << 7) |
686 ff_id3v2_parse(s, len, buf[3], buf[5], extra_meta);
688 avio_seek(s->pb, off, SEEK_SET);
690 } while (found_header);
691 ff_metadata_conv(&s->metadata, NULL, ff_id3v2_34_metadata_conv);
692 ff_metadata_conv(&s->metadata, NULL, id3v2_2_metadata_conv);
693 ff_metadata_conv(&s->metadata, NULL, ff_id3v2_4_metadata_conv);
694 merge_date(&s->metadata);
697 void ff_id3v2_free_extra_meta(ID3v2ExtraMeta **extra_meta)
699 ID3v2ExtraMeta *current = *extra_meta, *next;
700 const ID3v2EMFunc *extra_func;
703 if ((extra_func = get_extra_meta_func(current->tag, 1)))
704 extra_func->free(current->data);
705 next = current->next;
711 int ff_id3v2_parse_apic(AVFormatContext *s, ID3v2ExtraMeta **extra_meta)
715 for (cur = *extra_meta; cur; cur = cur->next) {
716 ID3v2ExtraMetaAPIC *apic;
719 if (strcmp(cur->tag, "APIC"))
723 if (!(st = avformat_new_stream(s, NULL)))
724 return AVERROR(ENOMEM);
726 st->disposition |= AV_DISPOSITION_ATTACHED_PIC;
727 st->codec->codec_type = AVMEDIA_TYPE_VIDEO;
728 st->codec->codec_id = apic->id;
729 av_dict_set(&st->metadata, "title", apic->description, 0);
730 av_dict_set(&st->metadata, "comment", apic->type, 0);
732 av_init_packet(&st->attached_pic);
733 st->attached_pic.data = apic->data;
734 st->attached_pic.size = apic->len;
735 st->attached_pic.destruct = av_destruct_packet;
736 st->attached_pic.stream_index = st->index;
737 st->attached_pic.flags |= AV_PKT_FLAG_KEY;