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" , AV_CODEC_ID_GIF},
116 {"image/jpeg", AV_CODEC_ID_MJPEG},
117 {"image/jpg", AV_CODEC_ID_MJPEG},
118 {"image/png" , AV_CODEC_ID_PNG},
119 {"image/tiff", AV_CODEC_ID_TIFF},
120 {"image/bmp", AV_CODEC_ID_BMP},
121 {"JPG", AV_CODEC_ID_MJPEG}, /* ID3v2.2 */
122 {"PNG" , AV_CODEC_ID_PNG}, /* ID3v2.2 */
123 {"", AV_CODEC_ID_NONE},
126 int ff_id3v2_match(const uint8_t *buf, const char * magic)
128 return buf[0] == magic[0] &&
129 buf[1] == magic[1] &&
130 buf[2] == magic[2] &&
133 (buf[6] & 0x80) == 0 &&
134 (buf[7] & 0x80) == 0 &&
135 (buf[8] & 0x80) == 0 &&
136 (buf[9] & 0x80) == 0;
139 int ff_id3v2_tag_len(const uint8_t * buf)
141 int len = ((buf[6] & 0x7f) << 21) +
142 ((buf[7] & 0x7f) << 14) +
143 ((buf[8] & 0x7f) << 7) +
147 len += ID3v2_HEADER_SIZE;
151 static unsigned int get_size(AVIOContext *s, int len)
155 v = (v << 7) + (avio_r8(s) & 0x7F);
160 * Free GEOB type extra metadata.
162 static void free_geobtag(void *obj)
164 ID3v2ExtraMetaGEOB *geob = obj;
165 av_free(geob->mime_type);
166 av_free(geob->file_name);
167 av_free(geob->description);
173 * Decode characters to UTF-8 according to encoding type. The decoded buffer is
174 * always null terminated. Stop reading when either *maxread bytes are read from
175 * pb or U+0000 character is found.
177 * @param dst Pointer where the address of the buffer with the decoded bytes is
178 * stored. Buffer must be freed by caller.
179 * @param maxread Pointer to maximum number of characters to read from the
180 * AVIOContext. After execution the value is decremented by the number of bytes
182 * @returns 0 if no error occurred, dst is uninitialized on error
184 static int decode_str(AVFormatContext *s, AVIOContext *pb, int encoding,
185 uint8_t **dst, int *maxread)
191 unsigned int (*get)(AVIOContext*) = avio_rb16;
194 if ((ret = avio_open_dyn_buf(&dynbuf)) < 0) {
195 av_log(s, AV_LOG_ERROR, "Error opening memory stream\n");
201 case ID3v2_ENCODING_ISO8859:
204 PUT_UTF8(ch, tmp, avio_w8(dynbuf, tmp);)
209 case ID3v2_ENCODING_UTF16BOM:
210 if ((left -= 2) < 0) {
211 av_log(s, AV_LOG_ERROR, "Cannot read BOM value, input too short\n");
212 avio_close_dyn_buf(dynbuf, dst);
214 return AVERROR_INVALIDDATA;
216 switch (avio_rb16(pb)) {
222 av_log(s, AV_LOG_ERROR, "Incorrect BOM value\n");
223 avio_close_dyn_buf(dynbuf, dst);
226 return AVERROR_INVALIDDATA;
230 case ID3v2_ENCODING_UTF16BE:
231 while ((left > 1) && ch) {
232 GET_UTF16(ch, ((left -= 2) >= 0 ? get(pb) : 0), break;)
233 PUT_UTF8(ch, tmp, avio_w8(dynbuf, tmp);)
236 left += 2; /* did not read last char from pb */
239 case ID3v2_ENCODING_UTF8:
247 av_log(s, AV_LOG_WARNING, "Unknown encoding\n");
253 avio_close_dyn_buf(dynbuf, dst);
262 static void read_ttag(AVFormatContext *s, AVIOContext *pb, int taglen, const char *key)
265 int encoding, dict_flags = AV_DICT_DONT_OVERWRITE | AV_DICT_DONT_STRDUP_VAL;
271 encoding = avio_r8(pb);
272 taglen--; /* account for encoding type byte */
274 if (decode_str(s, pb, encoding, &dst, &taglen) < 0) {
275 av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", key);
279 if (!(strcmp(key, "TCON") && strcmp(key, "TCO"))
280 && (sscanf(dst, "(%d)", &genre) == 1 || sscanf(dst, "%d", &genre) == 1)
281 && genre <= ID3v1_GENRE_MAX) {
283 dst = av_strdup(ff_id3v1_genre_str[genre]);
284 } else if (!(strcmp(key, "TXXX") && strcmp(key, "TXX"))) {
285 /* dst now contains the key, need to get value */
287 if (decode_str(s, pb, encoding, &dst, &taglen) < 0) {
288 av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", key);
292 dict_flags |= AV_DICT_DONT_STRDUP_KEY;
297 av_dict_set(&s->metadata, key, dst, dict_flags);
301 * Parse GEOB tag into a ID3v2ExtraMetaGEOB struct.
303 static void read_geobtag(AVFormatContext *s, AVIOContext *pb, int taglen, char *tag, ID3v2ExtraMeta **extra_meta)
305 ID3v2ExtraMetaGEOB *geob_data = NULL;
306 ID3v2ExtraMeta *new_extra = NULL;
313 geob_data = av_mallocz(sizeof(ID3v2ExtraMetaGEOB));
315 av_log(s, AV_LOG_ERROR, "Failed to alloc %zu bytes\n", sizeof(ID3v2ExtraMetaGEOB));
319 new_extra = av_mallocz(sizeof(ID3v2ExtraMeta));
321 av_log(s, AV_LOG_ERROR, "Failed to alloc %zu bytes\n", sizeof(ID3v2ExtraMeta));
325 /* read encoding type byte */
326 encoding = avio_r8(pb);
329 /* read MIME type (always ISO-8859) */
330 if (decode_str(s, pb, ID3v2_ENCODING_ISO8859, &geob_data->mime_type, &taglen) < 0
335 if (decode_str(s, pb, encoding, &geob_data->file_name, &taglen) < 0
339 /* read content description */
340 if (decode_str(s, pb, encoding, &geob_data->description, &taglen) < 0
345 /* save encapsulated binary data */
346 geob_data->data = av_malloc(taglen);
347 if (!geob_data->data) {
348 av_log(s, AV_LOG_ERROR, "Failed to alloc %d bytes\n", taglen);
351 if ((len = avio_read(pb, geob_data->data, taglen)) < taglen)
352 av_log(s, AV_LOG_WARNING, "Error reading GEOB frame, data truncated.\n");
353 geob_data->datasize = len;
355 geob_data->data = NULL;
356 geob_data->datasize = 0;
359 /* add data to the list */
360 new_extra->tag = "GEOB";
361 new_extra->data = geob_data;
362 new_extra->next = *extra_meta;
363 *extra_meta = new_extra;
368 av_log(s, AV_LOG_ERROR, "Error reading frame %s, skipped\n", tag);
369 free_geobtag(geob_data);
374 static int is_number(const char *str)
376 while (*str >= '0' && *str <= '9') str++;
380 static AVDictionaryEntry* get_date_tag(AVDictionary *m, const char *tag)
382 AVDictionaryEntry *t;
383 if ((t = av_dict_get(m, tag, NULL, AV_DICT_MATCH_CASE)) &&
384 strlen(t->value) == 4 && is_number(t->value))
389 static void merge_date(AVDictionary **m)
391 AVDictionaryEntry *t;
392 char date[17] = {0}; // YYYY-MM-DD hh:mm
394 if (!(t = get_date_tag(*m, "TYER")) &&
395 !(t = get_date_tag(*m, "TYE")))
397 av_strlcpy(date, t->value, 5);
398 av_dict_set(m, "TYER", NULL, 0);
399 av_dict_set(m, "TYE", NULL, 0);
401 if (!(t = get_date_tag(*m, "TDAT")) &&
402 !(t = get_date_tag(*m, "TDA")))
404 snprintf(date + 4, sizeof(date) - 4, "-%.2s-%.2s", t->value + 2, t->value);
405 av_dict_set(m, "TDAT", NULL, 0);
406 av_dict_set(m, "TDA", NULL, 0);
408 if (!(t = get_date_tag(*m, "TIME")) &&
409 !(t = get_date_tag(*m, "TIM")))
411 snprintf(date + 10, sizeof(date) - 10, " %.2s:%.2s", t->value, t->value + 2);
412 av_dict_set(m, "TIME", NULL, 0);
413 av_dict_set(m, "TIM", NULL, 0);
417 av_dict_set(m, "date", date, 0);
420 static void free_apic(void *obj)
422 ID3v2ExtraMetaAPIC *apic = obj;
423 av_buffer_unref(&apic->buf);
424 av_freep(&apic->description);
428 static void read_apic(AVFormatContext *s, AVIOContext *pb, int taglen, char *tag, ID3v2ExtraMeta **extra_meta)
432 const CodecMime *mime = ff_id3v2_mime_tags;
433 enum AVCodecID id = AV_CODEC_ID_NONE;
434 ID3v2ExtraMetaAPIC *apic = NULL;
435 ID3v2ExtraMeta *new_extra = NULL;
436 int64_t end = avio_tell(pb) + taglen;
441 new_extra = av_mallocz(sizeof(*new_extra));
442 apic = av_mallocz(sizeof(*apic));
443 if (!new_extra || !apic)
450 taglen -= avio_get_str(pb, taglen, mimetype, sizeof(mimetype));
451 while (mime->id != AV_CODEC_ID_NONE) {
452 if (!av_strncasecmp(mime->str, mimetype, sizeof(mimetype))) {
458 if (id == AV_CODEC_ID_NONE) {
459 av_log(s, AV_LOG_WARNING, "Unknown attached picture mimetype: %s, skipping.\n", mimetype);
465 pic_type = avio_r8(pb);
467 if (pic_type < 0 || pic_type >= FF_ARRAY_ELEMS(ff_id3v2_picture_types)) {
468 av_log(s, AV_LOG_WARNING, "Unknown attached picture type %d.\n", pic_type);
471 apic->type = ff_id3v2_picture_types[pic_type];
473 /* description and picture data */
474 if (decode_str(s, pb, enc, &apic->description, &taglen) < 0) {
475 av_log(s, AV_LOG_ERROR, "Error decoding attached picture description.\n");
479 apic->buf = av_buffer_alloc(taglen + FF_INPUT_BUFFER_PADDING_SIZE);
480 if (!apic->buf || avio_read(pb, apic->buf->data, taglen) != taglen)
482 memset(apic->buf->data + taglen, 0, FF_INPUT_BUFFER_PADDING_SIZE);
484 new_extra->tag = "APIC";
485 new_extra->data = apic;
486 new_extra->next = *extra_meta;
487 *extra_meta = new_extra;
494 av_freep(&new_extra);
495 avio_seek(pb, end, SEEK_SET);
498 typedef struct ID3v2EMFunc {
501 void (*read)(AVFormatContext*, AVIOContext*, int, char*, ID3v2ExtraMeta **);
502 void (*free)(void *obj);
505 static const ID3v2EMFunc id3v2_extra_meta_funcs[] = {
506 { "GEO", "GEOB", read_geobtag, free_geobtag },
507 { "PIC", "APIC", read_apic, free_apic },
512 * Get the corresponding ID3v2EMFunc struct for a tag.
513 * @param isv34 Determines if v2.2 or v2.3/4 strings are used
514 * @return A pointer to the ID3v2EMFunc struct if found, NULL otherwise.
516 static const ID3v2EMFunc *get_extra_meta_func(const char *tag, int isv34)
519 while (id3v2_extra_meta_funcs[i].tag3) {
521 (isv34 ? id3v2_extra_meta_funcs[i].tag4 :
522 id3v2_extra_meta_funcs[i].tag3),
524 return &id3v2_extra_meta_funcs[i];
530 static void id3v2_parse(AVFormatContext *s, int len, uint8_t version,
531 uint8_t flags, ID3v2ExtraMeta **extra_meta)
533 int isv34, tlen, unsync;
535 int64_t next, end = avio_tell(s->pb) + len;
537 const char *reason = NULL;
540 unsigned char *buffer = NULL;
542 const ID3v2EMFunc *extra_func;
547 reason = "compression";
565 unsync = flags & 0x80;
567 if (isv34 && flags & 0x40) { /* Extended header present, just skip over it */
568 int extlen = get_size(s->pb, 4);
570 extlen -= 4; // in v2.4 the length includes the length field we just read
573 reason = "invalid extended header length";
576 avio_skip(s->pb, extlen);
579 while (len >= taghdrlen) {
580 unsigned int tflags = 0;
584 avio_read(s->pb, tag, 4);
587 tlen = avio_rb32(s->pb);
589 tlen = get_size(s->pb, 4);
590 tflags = avio_rb16(s->pb);
591 tunsync = tflags & ID3v2_FLAG_UNSYNCH;
593 avio_read(s->pb, tag, 3);
595 tlen = avio_rb24(s->pb);
597 if (tlen < 0 || tlen > len - taghdrlen) {
598 av_log(s, AV_LOG_WARNING, "Invalid size in frame %s, skipping the rest of tag.\n", tag);
601 len -= taghdrlen + tlen;
602 next = avio_tell(s->pb) + tlen;
606 av_log(s, AV_LOG_DEBUG, "Invalid empty frame %s, skipping.\n", tag);
610 if (tflags & ID3v2_FLAG_DATALEN) {
615 if (tflags & (ID3v2_FLAG_ENCRYPTION | ID3v2_FLAG_COMPRESSION)) {
616 av_log(s, AV_LOG_WARNING, "Skipping encrypted/compressed ID3v2 frame %s.\n", tag);
617 avio_skip(s->pb, tlen);
618 /* check for text tag or supported special meta tag */
619 } else if (tag[0] == 'T' || (extra_meta && (extra_func = get_extra_meta_func(tag, isv34)))) {
620 if (unsync || tunsync) {
621 int64_t end = avio_tell(s->pb) + tlen;
623 av_fast_malloc(&buffer, &buffer_size, tlen);
625 av_log(s, AV_LOG_ERROR, "Failed to alloc %d bytes\n", tlen);
629 while (avio_tell(s->pb) < end) {
630 *b++ = avio_r8(s->pb);
631 if (*(b - 1) == 0xff && avio_tell(s->pb) < end - 1) {
632 uint8_t val = avio_r8(s->pb);
633 *b++ = val ? val : avio_r8(s->pb);
636 ffio_init_context(&pb, buffer, b - buffer, 0, NULL, NULL, NULL, NULL);
638 pbx = &pb; // read from sync buffer
640 pbx = s->pb; // read straight from input
644 read_ttag(s, pbx, tlen, tag);
646 /* parse special meta tag */
647 extra_func->read(s, pbx, tlen, tag, extra_meta);
651 av_log(s, AV_LOG_WARNING, "invalid frame id, assuming padding");
652 avio_skip(s->pb, tlen);
655 /* Skip to end of tag */
657 avio_seek(s->pb, next, SEEK_SET);
660 if (version == 4 && flags & 0x10) /* Footer preset, always 10 bytes, skip over it */
665 av_log(s, AV_LOG_INFO, "ID3v2.%d tag skipped, cannot handle %s\n", version, reason);
666 avio_seek(s->pb, end, SEEK_SET);
671 void ff_id3v2_read(AVFormatContext *s, const char *magic, ID3v2ExtraMeta **extra_meta)
674 uint8_t buf[ID3v2_HEADER_SIZE];
679 /* save the current offset in case there's nothing to read/skip */
680 off = avio_tell(s->pb);
681 ret = avio_read(s->pb, buf, ID3v2_HEADER_SIZE);
682 if (ret != ID3v2_HEADER_SIZE)
684 found_header = ff_id3v2_match(buf, magic);
686 /* parse ID3v2 header */
687 len = ((buf[6] & 0x7f) << 21) |
688 ((buf[7] & 0x7f) << 14) |
689 ((buf[8] & 0x7f) << 7) |
691 id3v2_parse(s, len, buf[3], buf[5], extra_meta);
693 avio_seek(s->pb, off, SEEK_SET);
695 } while (found_header);
696 ff_metadata_conv(&s->metadata, NULL, ff_id3v2_34_metadata_conv);
697 ff_metadata_conv(&s->metadata, NULL, id3v2_2_metadata_conv);
698 ff_metadata_conv(&s->metadata, NULL, ff_id3v2_4_metadata_conv);
699 merge_date(&s->metadata);
702 void ff_id3v2_free_extra_meta(ID3v2ExtraMeta **extra_meta)
704 ID3v2ExtraMeta *current = *extra_meta, *next;
705 const ID3v2EMFunc *extra_func;
708 if ((extra_func = get_extra_meta_func(current->tag, 1)))
709 extra_func->free(current->data);
710 next = current->next;
716 int ff_id3v2_parse_apic(AVFormatContext *s, ID3v2ExtraMeta **extra_meta)
720 for (cur = *extra_meta; cur; cur = cur->next) {
721 ID3v2ExtraMetaAPIC *apic;
724 if (strcmp(cur->tag, "APIC"))
728 if (!(st = avformat_new_stream(s, NULL)))
729 return AVERROR(ENOMEM);
731 st->disposition |= AV_DISPOSITION_ATTACHED_PIC;
732 st->codec->codec_type = AVMEDIA_TYPE_VIDEO;
733 st->codec->codec_id = apic->id;
734 av_dict_set(&st->metadata, "title", apic->description, 0);
735 av_dict_set(&st->metadata, "comment", apic->type, 0);
737 av_init_packet(&st->attached_pic);
738 st->attached_pic.buf = apic->buf;
739 st->attached_pic.data = apic->buf->data;
740 st->attached_pic.size = apic->buf->size - FF_INPUT_BUFFER_PADDING_SIZE;
741 st->attached_pic.stream_index = st->index;
742 st->attached_pic.flags |= AV_PKT_FLAG_KEY;