src/mp3-util.h

   1 /* libSoX MP3 utilities  Copyright (c) 2007-9 SoX contributors
   2  *
   3  * This library is free software; you can redistribute it and/or modify it
   4  * under the terms of the GNU Lesser General Public License as published by
   5  * the Free Software Foundation; either version 2.1 of the License, or (at
   6  * your option) any later version.
   7  *
   8  * This library is distributed in the hope that it will be useful, but
   9  * WITHOUT ANY WARRANTY; without even the implied warranty of
  10  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU Lesser
  11  * General Public License for more details.
  12  *
  13  * You should have received a copy of the GNU Lesser General Public License
  14  * along with this library; if not, write to the Free Software Foundation,
  15  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
  16  */
  17
  18 #include <sys/stat.h>
  19
  20 #ifdef USING_ID3TAG
  21
  22 static char const * id3tagmap[][2] =
  23 {
  24   {"TIT2", "Title"},
  25   {"TPE1", "Artist"},
  26   {"TALB", "Album"},
  27   {"TRCK", "Tracknumber"},
  28   {"TDRC", "Year"},
  29   {"TCON", "Genre"},
  30   {"COMM", "Comment"},
  31   {"TPOS", "Discnumber"},
  32   {NULL, NULL}
  33 };
  34
  35 #endif /* USING_ID3TAG */
  36
  37 #if defined(HAVE_LAME)
  38
  39 static void write_comments(sox_format_t * ft)
  40 {
  41   priv_t *p = (priv_t *) ft->priv;
  42   const char* comment;
  43
  44   p->id3tag_init(p->gfp);
  45   p->id3tag_set_pad(p->gfp, (size_t)ID3PADDING);
  46
  47   /* Note: id3tag_set_fieldvalue is not present in LAME 3.97, so we're using
  48      the 3.97-compatible methods for all of the tags that 3.97 supported. */
  49   /* FIXME: This is no more necessary, since support for LAME 3.97 has ended. */
  50   if ((comment = sox_find_comment(ft->oob.comments, "Title")))
  51     p->id3tag_set_title(p->gfp, comment);
  52   if ((comment = sox_find_comment(ft->oob.comments, "Artist")))
  53     p->id3tag_set_artist(p->gfp, comment);
  54   if ((comment = sox_find_comment(ft->oob.comments, "Album")))
  55     p->id3tag_set_album(p->gfp, comment);
  56   if ((comment = sox_find_comment(ft->oob.comments, "Tracknumber")))
  57     p->id3tag_set_track(p->gfp, comment);
  58   if ((comment = sox_find_comment(ft->oob.comments, "Year")))
  59     p->id3tag_set_year(p->gfp, comment);
  60   if ((comment = sox_find_comment(ft->oob.comments, "Comment")))
  61     p->id3tag_set_comment(p->gfp, comment);
  62   if ((comment = sox_find_comment(ft->oob.comments, "Genre")))
  63   {
  64     if (p->id3tag_set_genre(p->gfp, comment))
  65       lsx_warn("\"%s\" is not a recognized ID3v1 genre.", comment);
  66   }
  67
  68   if ((comment = sox_find_comment(ft->oob.comments, "Discnumber")))
  69   {
  70     char* id3tag_buf = lsx_malloc(strlen(comment) + 6);
  71     if (id3tag_buf)
  72     {
  73       sprintf(id3tag_buf, "TPOS=%s", comment);
  74       p->id3tag_set_fieldvalue(p->gfp, id3tag_buf);
  75       free(id3tag_buf);
  76     }
  77   }
  78 }
  79
  80 #endif /* HAVE_LAME */
  81
  82 #ifdef USING_ID3TAG
  83
  84 static id3_utf8_t * utf8_id3tag_findframe(
  85     struct id3_tag * tag, const char * const frameid, unsigned index)
  86 {
  87   struct id3_frame const * frame = id3_tag_findframe(tag, frameid, index);
  88   if (frame) {
  89     union id3_field  const * field = id3_frame_field(frame, 1);
  90     unsigned nstrings = id3_field_getnstrings(field);
  91     while (nstrings--){
  92       id3_ucs4_t const * ucs4 = id3_field_getstrings(field, nstrings);
  93       if (ucs4)
  94         return id3_ucs4_utf8duplicate(ucs4); /* Must call free() on this */
  95     }
  96   }
  97   return NULL;
  98 }
  99
 100 struct tag_info_node
 101 {
 102     struct tag_info_node * next;
 103     off_t start;
 104     off_t end;
 105 };
 106
 107 struct tag_info {
 108   sox_format_t * ft;
 109   struct tag_info_node * head;
 110   struct id3_tag * tag;
 111 };
 112
 113 static int add_tag(struct tag_info * info)
 114 {
 115   struct tag_info_node * current;
 116   off_t start, end;
 117   id3_byte_t query[ID3_TAG_QUERYSIZE];
 118   id3_byte_t * buffer;
 119   long size;
 120   int result = 0;
 121
 122   /* Ensure we're at the start of a valid tag and get its size. */
 123   if (ID3_TAG_QUERYSIZE != lsx_readbuf(info->ft, query, ID3_TAG_QUERYSIZE) ||
 124       !(size = id3_tag_query(query, ID3_TAG_QUERYSIZE))) {
 125     return 0;
 126   }
 127   if (size < 0) {
 128     if (0 != lsx_seeki(info->ft, size, SEEK_CUR) ||
 129         ID3_TAG_QUERYSIZE != lsx_readbuf(info->ft, query, ID3_TAG_QUERYSIZE) ||
 130         (size = id3_tag_query(query, ID3_TAG_QUERYSIZE)) <= 0) {
 131       return 0;
 132     }
 133   }
 134
 135   /* Don't read a tag more than once. */
 136   start = lsx_tell(info->ft);
 137   end = start + size;
 138   for (current = info->head; current; current = current->next) {
 139     if (start == current->start && end == current->end) {
 140       return 1;
 141     } else if (start < current->end && current->start < end) {
 142       return 0;
 143     }
 144   }
 145
 146   buffer = lsx_malloc((size_t)size);
 147   if (!buffer) {
 148     return 0;
 149   }
 150   memcpy(buffer, query, ID3_TAG_QUERYSIZE);
 151   if ((unsigned long)size - ID3_TAG_QUERYSIZE ==
 152       lsx_readbuf(info->ft, buffer + ID3_TAG_QUERYSIZE, (size_t)size - ID3_TAG_QUERYSIZE)) {
 153     struct id3_tag * tag = id3_tag_parse(buffer, (size_t)size);
 154     if (tag) {
 155       current = lsx_malloc(sizeof(struct tag_info_node));
 156       if (current) {
 157         current->next = info->head;
 158         current->start = start;
 159         current->end = end;
 160         info->head = current;
 161         if (info->tag && (info->tag->extendedflags & ID3_TAG_EXTENDEDFLAG_TAGISANUPDATE)) {
 162           struct id3_frame * frame;
 163           unsigned i;
 164           for (i = 0; (frame = id3_tag_findframe(tag, NULL, i)); i++) {
 165             id3_tag_attachframe(info->tag, frame);
 166           }
 167           id3_tag_delete(tag);
 168         } else {
 169           if (info->tag) {
 170             id3_tag_delete(info->tag);
 171           }
 172           info->tag = tag;
 173         }
 174       }
 175     }
 176   }
 177   free(buffer);
 178   return result;
 179 }
 180
 181 static void read_comments(sox_format_t * ft)
 182 {
 183   struct tag_info   info;
 184   id3_utf8_t        * utf8;
 185   int               i;
 186   int               has_id3v1 = 0;
 187
 188   info.ft = ft;
 189   info.head = NULL;
 190   info.tag = NULL;
 191
 192   /*
 193   We look for:
 194   ID3v1 at end (EOF - 128).
 195   ID3v2 at start.
 196   ID3v2 at end (but before ID3v1 from end if there was one).
 197   */
 198
 199   if (0 == lsx_seeki(ft, -128, SEEK_END)) {
 200     has_id3v1 =
 201       add_tag(&info) &&
 202       1 == ID3_TAG_VERSION_MAJOR(id3_tag_version(info.tag));
 203   }
 204   if (0 == lsx_seeki(ft, 0, SEEK_SET)) {
 205     add_tag(&info);
 206   }
 207   if (0 == lsx_seeki(ft, has_id3v1 ? -138 : -10, SEEK_END)) {
 208     add_tag(&info);
 209   }
 210   if (info.tag && info.tag->frames) {
 211     for (i = 0; id3tagmap[i][0]; ++i) {
 212       if ((utf8 = utf8_id3tag_findframe(info.tag, id3tagmap[i][0], 0))) {
 213         char * comment = lsx_malloc(strlen(id3tagmap[i][1]) + 1 + strlen((char *)utf8) + 1);
 214         sprintf(comment, "%s=%s", id3tagmap[i][1], utf8);
 215         sox_append_comment(&ft->oob.comments, comment);
 216         free(comment);
 217         free(utf8);
 218       }
 219     }
 220     if ((utf8 = utf8_id3tag_findframe(info.tag, "TLEN", 0))) {
 221       unsigned long tlen = strtoul((char *)utf8, NULL, 10);
 222       if (tlen > 0 && tlen < ULONG_MAX) {
 223         ft->signal.length= tlen; /* In ms; convert to samples later */
 224         lsx_debug("got exact duration from ID3 TLEN");
 225       }
 226       free(utf8);
 227     }
 228   }
 229   while (info.head) {
 230     struct tag_info_node * head = info.head;
 231     info.head = head->next;
 232     free(head);
 233   }
 234   if (info.tag) {
 235     id3_tag_delete(info.tag);
 236   }
 237 }
 238
 239 #endif /* USING_ID3TAG */
 240
 241 #ifdef HAVE_MAD_H
 242
 243 static unsigned long xing_frames(priv_t * p, struct mad_bitptr ptr, unsigned bitlen)
 244 {
 245   #define XING_MAGIC ( ('X' << 24) | ('i' << 16) | ('n' << 8) | 'g' )
 246   if (bitlen >= 96 && p->mad_bit_read(&ptr, 32) == XING_MAGIC &&
 247       (p->mad_bit_read(&ptr, 32) & 1 )) /* XING_FRAMES */
 248     return p->mad_bit_read(&ptr, 32);
 249   return 0;
 250 }
 251
 252 static size_t mp3_duration(sox_format_t * ft)
 253 {
 254   priv_t              * p = (priv_t *) ft->priv;
 255   struct mad_stream   mad_stream;
 256   struct mad_header   mad_header;
 257   struct mad_frame    mad_frame;
 258   size_t              initial_bitrate = 0; /* Initialised to prevent warning */
 259   size_t              tagsize = 0, consumed = 0, frames = 0;
 260   sox_bool            vbr = sox_false, depadded = sox_false;
 261   size_t              num_samples = 0;
 262
 263   p->mad_stream_init(&mad_stream);
 264   p->mad_header_init(&mad_header);
 265   p->mad_frame_init(&mad_frame);
 266
 267   do {  /* Read data from the MP3 file */
 268     int read, padding = 0;
 269     size_t leftover = mad_stream.bufend - mad_stream.next_frame;
 270
 271     memmove(p->mp3_buffer, mad_stream.this_frame, leftover);
 272     read = lsx_readbuf(ft, p->mp3_buffer + leftover, p->mp3_buffer_size - leftover);
 273     if (read <= 0) {
 274       lsx_debug("got exact duration by scan to EOF (frames=%" PRIuPTR " leftover=%" PRIuPTR ")", frames, leftover);
 275       break;
 276     }
 277     for (; !depadded && padding < read && !p->mp3_buffer[padding]; ++padding);
 278     depadded = sox_true;
 279     p->mad_stream_buffer(&mad_stream, p->mp3_buffer + padding, leftover + read - padding);
 280
 281     while (sox_true) {  /* Decode frame headers */
 282       mad_stream.error = MAD_ERROR_NONE;
 283       if (p->mad_header_decode(&mad_header, &mad_stream) == -1) {
 284         if (mad_stream.error == MAD_ERROR_BUFLEN)
 285           break;  /* Normal behaviour; get some more data from the file */
 286         if (!MAD_RECOVERABLE(mad_stream.error)) {
 287           lsx_warn("unrecoverable MAD error");
 288           break;
 289         }
 290         if (mad_stream.error == MAD_ERROR_LOSTSYNC) {
 291           unsigned available = (mad_stream.bufend - mad_stream.this_frame);
 292           tagsize = tagtype(mad_stream.this_frame, (size_t) available);
 293           if (tagsize) {   /* It's some ID3 tags, so just skip */
 294             if (tagsize >= available) {
 295               lsx_seeki(ft, (off_t)(tagsize - available), SEEK_CUR);
 296               depadded = sox_false;
 297             }
 298             p->mad_stream_skip(&mad_stream, min(tagsize, available));
 299           }
 300           else lsx_warn("MAD lost sync");
 301         }
 302         else lsx_warn("recoverable MAD error");
 303         continue; /* Not an audio frame */
 304       }
 305
 306       num_samples += MAD_NSBSAMPLES(&mad_header) * 32;
 307       consumed += mad_stream.next_frame - mad_stream.this_frame;
 308
 309       lsx_debug_more("bitrate=%lu", mad_header.bitrate);
 310       if (!frames) {
 311         initial_bitrate = mad_header.bitrate;
 312
 313         /* Get the precise frame count from the XING header if present */
 314         mad_frame.header = mad_header;
 315         if (p->mad_frame_decode(&mad_frame, &mad_stream) == -1)
 316           if (!MAD_RECOVERABLE(mad_stream.error)) {
 317             lsx_warn("unrecoverable MAD error");
 318             break;
 319           }
 320         if ((frames = xing_frames(p, mad_stream.anc_ptr, mad_stream.anc_bitlen))) {
 321           num_samples *= frames;
 322           lsx_debug("got exact duration from XING frame count (%" PRIuPTR ")", frames);
 323           break;
 324         }
 325       }
 326       else vbr |= mad_header.bitrate != initial_bitrate;
 327
 328       /* If not VBR, we can time just a few frames then extrapolate */
 329       if (++frames == 25 && !vbr) {
 330         double frame_size = (double) consumed / frames;
 331         size_t num_frames = (lsx_filelength(ft) - tagsize) / frame_size;
 332         num_samples = num_samples / frames * num_frames;
 333         lsx_debug("got approx. duration by CBR extrapolation");
 334         break;
 335       }
 336     }
 337   } while (mad_stream.error == MAD_ERROR_BUFLEN);
 338
 339   p->mad_frame_finish(&mad_frame);
 340   mad_header_finish(&mad_header);
 341   p->mad_stream_finish(&mad_stream);
 342   lsx_rewind(ft);
 343
 344   return num_samples;
 345 }
 346
 347 #endif /* HAVE_MAD_H */