[Summary view] [Print] [Text view]

   1  <?php
   2  
   3  /////////////////////////////////////////////////////////////////
   4  /// getID3() by James Heinrich <info@getid3.org>               //
   5  //  available at https://github.com/JamesHeinrich/getID3       //
   6  //            or https://www.getid3.org                        //
   7  //            or http://getid3.sourceforge.net                 //
   8  //  see readme.txt for more details                            //
   9  /////////////////////////////////////////////////////////////////
  10  //                                                             //
  11  // module.tag.id3v1.php                                        //
  12  // module for analyzing ID3v1 tags                             //
  13  // dependencies: NONE                                          //
  14  //                                                            ///
  15  /////////////////////////////////////////////////////////////////
  16  
  17  if (!defined('GETID3_INCLUDEPATH')) { // prevent path-exposing attacks that access modules directly on public webservers
  18      exit;
  19  }
  20  
  21  class getid3_id3v1 extends getid3_handler
  22  {
  23      /**
  24       * @return bool
  25       */
  26  	public function Analyze() {
  27          $info = &$this->getid3->info;
  28  
  29          if (!getid3_lib::intValueSupported($info['filesize'])) {
  30              $this->warning('Unable to check for ID3v1 because file is larger than '.round(PHP_INT_MAX / 1073741824).'GB');
  31              return false;
  32          } elseif ($info['filesize'] < 128) {
  33              $this->warning('Unable to check for ID3v1 because file is too small');
  34              return false;
  35          }
  36  
  37          if($info['filesize'] < 256) {
  38              $this->fseek(-128, SEEK_END);
  39              $preid3v1 = '';
  40              $id3v1tag = $this->fread(128);
  41          } else {
  42              $this->fseek(-256, SEEK_END);
  43              $preid3v1 = $this->fread(128);
  44              $id3v1tag = $this->fread(128);
  45          }
  46  
  47  
  48          if (substr($id3v1tag, 0, 3) == 'TAG') {
  49  
  50              $info['avdataend'] = $info['filesize'] - 128;
  51  
  52              $ParsedID3v1            = array();
  53              $ParsedID3v1['title']   = $this->cutfield(substr($id3v1tag,   3, 30));
  54              $ParsedID3v1['artist']  = $this->cutfield(substr($id3v1tag,  33, 30));
  55              $ParsedID3v1['album']   = $this->cutfield(substr($id3v1tag,  63, 30));
  56              $ParsedID3v1['year']    = $this->cutfield(substr($id3v1tag,  93,  4));
  57              $ParsedID3v1['comment'] =                 substr($id3v1tag,  97, 30);  // can't remove nulls yet, track detection depends on them
  58              $ParsedID3v1['genreid'] =             ord(substr($id3v1tag, 127,  1));
  59  
  60              // If second-last byte of comment field is null and last byte of comment field is non-null
  61              // then this is ID3v1.1 and the comment field is 28 bytes long and the 30th byte is the track number
  62              if (($id3v1tag[125] === "\x00") && ($id3v1tag[126] !== "\x00")) {
  63                  $ParsedID3v1['track_number'] = ord(substr($ParsedID3v1['comment'], 29,  1));
  64                  $ParsedID3v1['comment']      =     substr($ParsedID3v1['comment'],  0, 28);
  65              }
  66              $ParsedID3v1['comment'] = $this->cutfield($ParsedID3v1['comment']);
  67  
  68              $ParsedID3v1['genre'] = $this->LookupGenreName($ParsedID3v1['genreid']);
  69              if (!empty($ParsedID3v1['genre'])) {
  70                  unset($ParsedID3v1['genreid']);
  71              }
  72              if (empty($ParsedID3v1['genre']) || ($ParsedID3v1['genre'] == 'Unknown')) {
  73                  unset($ParsedID3v1['genre']);
  74              }
  75  
  76              foreach ($ParsedID3v1 as $key => $value) {
  77                  $ParsedID3v1['comments'][$key][0] = $value;
  78              }
  79              $ID3v1encoding = $this->getid3->encoding_id3v1;
  80              if ($this->getid3->encoding_id3v1_autodetect) {
  81                  // ID3v1 encoding detection hack START
  82                  // ID3v1 is defined as always using ISO-8859-1 encoding, but it is not uncommon to find files tagged with ID3v1 using Windows-1251 or other character sets
  83                  // Since ID3v1 has no concept of character sets there is no certain way to know we have the correct non-ISO-8859-1 character set, but we can guess
  84                  foreach ($ParsedID3v1['comments'] as $tag_key => $valuearray) {
  85                      foreach ($valuearray as $key => $value) {
  86                          if (preg_match('#^[\\x00-\\x40\\x80-\\xFF]+$#', $value) && !ctype_digit((string) $value)) { // check for strings with only characters above chr(128) and punctuation/numbers, but not just numeric strings (e.g. track numbers or years)
  87                              foreach (array('Windows-1251', 'KOI8-R') as $id3v1_bad_encoding) {
  88                                  if (function_exists('mb_convert_encoding') && @mb_convert_encoding($value, $id3v1_bad_encoding, $id3v1_bad_encoding) === $value) {
  89                                      $ID3v1encoding = $id3v1_bad_encoding;
  90                                      $this->warning('ID3v1 detected as '.$id3v1_bad_encoding.' text encoding in '.$tag_key);
  91                                      break 3;
  92                                  } elseif (function_exists('iconv') && @iconv($id3v1_bad_encoding, $id3v1_bad_encoding, $value) === $value) {
  93                                      $ID3v1encoding = $id3v1_bad_encoding;
  94                                      $this->warning('ID3v1 detected as '.$id3v1_bad_encoding.' text encoding in '.$tag_key);
  95                                      break 3;
  96                                  }
  97                              }
  98                          }
  99                      }
 100                  }
 101                  // ID3v1 encoding detection hack END
 102              }
 103  
 104              // ID3v1 data is supposed to be padded with NULL characters, but some taggers pad with spaces
 105              $GoodFormatID3v1tag = $this->GenerateID3v1Tag(
 106                                              $ParsedID3v1['title'],
 107                                              $ParsedID3v1['artist'],
 108                                              $ParsedID3v1['album'],
 109                                              $ParsedID3v1['year'],
 110                                              (isset($ParsedID3v1['genre']) ? $this->LookupGenreID($ParsedID3v1['genre']) : false),
 111                                              $ParsedID3v1['comment'],
 112                                              (!empty($ParsedID3v1['track_number']) ? $ParsedID3v1['track_number'] : ''));
 113              $ParsedID3v1['padding_valid'] = true;
 114              if ($id3v1tag !== $GoodFormatID3v1tag) {
 115                  $ParsedID3v1['padding_valid'] = false;
 116                  $this->warning('Some ID3v1 fields do not use NULL characters for padding');
 117              }
 118  
 119              $ParsedID3v1['tag_offset_end']   = $info['filesize'];
 120              $ParsedID3v1['tag_offset_start'] = $ParsedID3v1['tag_offset_end'] - 128;
 121  
 122              $info['id3v1'] = $ParsedID3v1;
 123              $info['id3v1']['encoding'] = $ID3v1encoding;
 124          }
 125  
 126          if (substr($preid3v1, 0, 3) == 'TAG') {
 127              // The way iTunes handles tags is, well, brain-damaged.
 128              // It completely ignores v1 if ID3v2 is present.
 129              // This goes as far as adding a new v1 tag *even if there already is one*
 130  
 131              // A suspected double-ID3v1 tag has been detected, but it could be that
 132              // the "TAG" identifier is a legitimate part of an APE or Lyrics3 tag
 133              if (substr($preid3v1, 96, 8) == 'APETAGEX') {
 134                  // an APE tag footer was found before the last ID3v1, assume false "TAG" synch
 135              } elseif (substr($preid3v1, 119, 6) == 'LYRICS') {
 136                  // a Lyrics3 tag footer was found before the last ID3v1, assume false "TAG" synch
 137              } else {
 138                  // APE and Lyrics3 footers not found - assume double ID3v1
 139                  $this->warning('Duplicate ID3v1 tag detected - this has been known to happen with iTunes');
 140                  $info['avdataend'] -= 128;
 141              }
 142          }
 143  
 144          return true;
 145      }
 146  
 147      /**
 148       * @param string $str
 149       *
 150       * @return string
 151       */
 152  	public static function cutfield($str) {
 153          return trim(substr($str, 0, strcspn($str, "\x00")));
 154      }
 155  
 156      /**
 157       * @param bool $allowSCMPXextended
 158       *
 159       * @return string[]
 160       */
 161  	public static function ArrayOfGenres($allowSCMPXextended=false) {
 162          static $GenreLookup = array(
 163              0    => 'Blues',
 164              1    => 'Classic Rock',
 165              2    => 'Country',
 166              3    => 'Dance',
 167              4    => 'Disco',
 168              5    => 'Funk',
 169              6    => 'Grunge',
 170              7    => 'Hip-Hop',
 171              8    => 'Jazz',
 172              9    => 'Metal',
 173              10   => 'New Age',
 174              11   => 'Oldies',
 175              12   => 'Other',
 176              13   => 'Pop',
 177              14   => 'R&B',
 178              15   => 'Rap',
 179              16   => 'Reggae',
 180              17   => 'Rock',
 181              18   => 'Techno',
 182              19   => 'Industrial',
 183              20   => 'Alternative',
 184              21   => 'Ska',
 185              22   => 'Death Metal',
 186              23   => 'Pranks',
 187              24   => 'Soundtrack',
 188              25   => 'Euro-Techno',
 189              26   => 'Ambient',
 190              27   => 'Trip-Hop',
 191              28   => 'Vocal',
 192              29   => 'Jazz+Funk',
 193              30   => 'Fusion',
 194              31   => 'Trance',
 195              32   => 'Classical',
 196              33   => 'Instrumental',
 197              34   => 'Acid',
 198              35   => 'House',
 199              36   => 'Game',
 200              37   => 'Sound Clip',
 201              38   => 'Gospel',
 202              39   => 'Noise',
 203              40   => 'Alt. Rock',
 204              41   => 'Bass',
 205              42   => 'Soul',
 206              43   => 'Punk',
 207              44   => 'Space',
 208              45   => 'Meditative',
 209              46   => 'Instrumental Pop',
 210              47   => 'Instrumental Rock',
 211              48   => 'Ethnic',
 212              49   => 'Gothic',
 213              50   => 'Darkwave',
 214              51   => 'Techno-Industrial',
 215              52   => 'Electronic',
 216              53   => 'Pop-Folk',
 217              54   => 'Eurodance',
 218              55   => 'Dream',
 219              56   => 'Southern Rock',
 220              57   => 'Comedy',
 221              58   => 'Cult',
 222              59   => 'Gangsta Rap',
 223              60   => 'Top 40',
 224              61   => 'Christian Rap',
 225              62   => 'Pop/Funk',
 226              63   => 'Jungle',
 227              64   => 'Native American',
 228              65   => 'Cabaret',
 229              66   => 'New Wave',
 230              67   => 'Psychedelic',
 231              68   => 'Rave',
 232              69   => 'Showtunes',
 233              70   => 'Trailer',
 234              71   => 'Lo-Fi',
 235              72   => 'Tribal',
 236              73   => 'Acid Punk',
 237              74   => 'Acid Jazz',
 238              75   => 'Polka',
 239              76   => 'Retro',
 240              77   => 'Musical',
 241              78   => 'Rock & Roll',
 242              79   => 'Hard Rock',
 243              80   => 'Folk',
 244              81   => 'Folk/Rock',
 245              82   => 'National Folk',
 246              83   => 'Swing',
 247              84   => 'Fast-Fusion',
 248              85   => 'Bebob',
 249              86   => 'Latin',
 250              87   => 'Revival',
 251              88   => 'Celtic',
 252              89   => 'Bluegrass',
 253              90   => 'Avantgarde',
 254              91   => 'Gothic Rock',
 255              92   => 'Progressive Rock',
 256              93   => 'Psychedelic Rock',
 257              94   => 'Symphonic Rock',
 258              95   => 'Slow Rock',
 259              96   => 'Big Band',
 260              97   => 'Chorus',
 261              98   => 'Easy Listening',
 262              99   => 'Acoustic',
 263              100  => 'Humour',
 264              101  => 'Speech',
 265              102  => 'Chanson',
 266              103  => 'Opera',
 267              104  => 'Chamber Music',
 268              105  => 'Sonata',
 269              106  => 'Symphony',
 270              107  => 'Booty Bass',
 271              108  => 'Primus',
 272              109  => 'Porn Groove',
 273              110  => 'Satire',
 274              111  => 'Slow Jam',
 275              112  => 'Club',
 276              113  => 'Tango',
 277              114  => 'Samba',
 278              115  => 'Folklore',
 279              116  => 'Ballad',
 280              117  => 'Power Ballad',
 281              118  => 'Rhythmic Soul',
 282              119  => 'Freestyle',
 283              120  => 'Duet',
 284              121  => 'Punk Rock',
 285              122  => 'Drum Solo',
 286              123  => 'A Cappella',
 287              124  => 'Euro-House',
 288              125  => 'Dance Hall',
 289              126  => 'Goa',
 290              127  => 'Drum & Bass',
 291              128  => 'Club-House',
 292              129  => 'Hardcore',
 293              130  => 'Terror',
 294              131  => 'Indie',
 295              132  => 'BritPop',
 296              133  => 'Negerpunk',
 297              134  => 'Polsk Punk',
 298              135  => 'Beat',
 299              136  => 'Christian Gangsta Rap',
 300              137  => 'Heavy Metal',
 301              138  => 'Black Metal',
 302              139  => 'Crossover',
 303              140  => 'Contemporary Christian',
 304              141  => 'Christian Rock',
 305              142  => 'Merengue',
 306              143  => 'Salsa',
 307              144  => 'Thrash Metal',
 308              145  => 'Anime',
 309              146  => 'JPop',
 310              147  => 'Synthpop',
 311              148 => 'Abstract',
 312              149 => 'Art Rock',
 313              150 => 'Baroque',
 314              151 => 'Bhangra',
 315              152 => 'Big Beat',
 316              153 => 'Breakbeat',
 317              154 => 'Chillout',
 318              155 => 'Downtempo',
 319              156 => 'Dub',
 320              157 => 'EBM',
 321              158 => 'Eclectic',
 322              159 => 'Electro',
 323              160 => 'Electroclash',
 324              161 => 'Emo',
 325              162 => 'Experimental',
 326              163 => 'Garage',
 327              164 => 'Global',
 328              165 => 'IDM',
 329              166 => 'Illbient',
 330              167 => 'Industro-Goth',
 331              168 => 'Jam Band',
 332              169 => 'Krautrock',
 333              170 => 'Leftfield',
 334              171 => 'Lounge',
 335              172 => 'Math Rock',
 336              173 => 'New Romantic',
 337              174 => 'Nu-Breakz',
 338              175 => 'Post-Punk',
 339              176 => 'Post-Rock',
 340              177 => 'Psytrance',
 341              178 => 'Shoegaze',
 342              179 => 'Space Rock',
 343              180 => 'Trop Rock',
 344              181 => 'World Music',
 345              182 => 'Neoclassical',
 346              183 => 'Audiobook',
 347              184 => 'Audio Theatre',
 348              185 => 'Neue Deutsche Welle',
 349              186 => 'Podcast',
 350              187 => 'Indie-Rock',
 351              188 => 'G-Funk',
 352              189 => 'Dubstep',
 353              190 => 'Garage Rock',
 354              191 => 'Psybient',
 355  
 356              255  => 'Unknown',
 357  
 358              'CR' => 'Cover',
 359              'RX' => 'Remix'
 360          );
 361  
 362          static $GenreLookupSCMPX = array();
 363          if ($allowSCMPXextended && empty($GenreLookupSCMPX)) {
 364              $GenreLookupSCMPX = $GenreLookup;
 365              // http://www.geocities.co.jp/SiliconValley-Oakland/3664/alittle.html#GenreExtended
 366              // Extended ID3v1 genres invented by SCMPX
 367              // Note that 255 "Japanese Anime" conflicts with standard "Unknown"
 368              $GenreLookupSCMPX[240] = 'Sacred';
 369              $GenreLookupSCMPX[241] = 'Northern Europe';
 370              $GenreLookupSCMPX[242] = 'Irish & Scottish';
 371              $GenreLookupSCMPX[243] = 'Scotland';
 372              $GenreLookupSCMPX[244] = 'Ethnic Europe';
 373              $GenreLookupSCMPX[245] = 'Enka';
 374              $GenreLookupSCMPX[246] = 'Children\'s Song';
 375              $GenreLookupSCMPX[247] = 'Japanese Sky';
 376              $GenreLookupSCMPX[248] = 'Japanese Heavy Rock';
 377              $GenreLookupSCMPX[249] = 'Japanese Doom Rock';
 378              $GenreLookupSCMPX[250] = 'Japanese J-POP';
 379              $GenreLookupSCMPX[251] = 'Japanese Seiyu';
 380              $GenreLookupSCMPX[252] = 'Japanese Ambient Techno';
 381              $GenreLookupSCMPX[253] = 'Japanese Moemoe';
 382              $GenreLookupSCMPX[254] = 'Japanese Tokusatsu';
 383              //$GenreLookupSCMPX[255] = 'Japanese Anime';
 384          }
 385  
 386          return ($allowSCMPXextended ? $GenreLookupSCMPX : $GenreLookup);
 387      }
 388  
 389      /**
 390       * @param string $genreid
 391       * @param bool   $allowSCMPXextended
 392       *
 393       * @return string|false
 394       */
 395  	public static function LookupGenreName($genreid, $allowSCMPXextended=true) {
 396          switch ($genreid) {
 397              case 'RX':
 398              case 'CR':
 399                  break;
 400              default:
 401                  if (!is_numeric($genreid)) {
 402                      return false;
 403                  }
 404                  $genreid = intval($genreid); // to handle 3 or '3' or '03'
 405                  break;
 406          }
 407          $GenreLookup = self::ArrayOfGenres($allowSCMPXextended);
 408          return (isset($GenreLookup[$genreid]) ? $GenreLookup[$genreid] : false);
 409      }
 410  
 411      /**
 412       * @param string $genre
 413       * @param bool   $allowSCMPXextended
 414       *
 415       * @return string|false
 416       */
 417  	public static function LookupGenreID($genre, $allowSCMPXextended=false) {
 418          $GenreLookup = self::ArrayOfGenres($allowSCMPXextended);
 419          $LowerCaseNoSpaceSearchTerm = strtolower(str_replace(' ', '', $genre));
 420          foreach ($GenreLookup as $key => $value) {
 421              if (strtolower(str_replace(' ', '', $value)) == $LowerCaseNoSpaceSearchTerm) {
 422                  return $key;
 423              }
 424          }
 425          return false;
 426      }
 427  
 428      /**
 429       * @param string $OriginalGenre
 430       *
 431       * @return string|false
 432       */
 433  	public static function StandardiseID3v1GenreName($OriginalGenre) {
 434          if (($GenreID = self::LookupGenreID($OriginalGenre)) !== false) {
 435              return self::LookupGenreName($GenreID);
 436          }
 437          return $OriginalGenre;
 438      }
 439  
 440      /**
 441       * @param string     $title
 442       * @param string     $artist
 443       * @param string     $album
 444       * @param string     $year
 445       * @param int        $genreid
 446       * @param string     $comment
 447       * @param int|string $track
 448       *
 449       * @return string
 450       */
 451  	public static function GenerateID3v1Tag($title, $artist, $album, $year, $genreid, $comment, $track='') {
 452          $ID3v1Tag  = 'TAG';
 453          $ID3v1Tag .= str_pad(trim(substr($title,  0, 30)), 30, "\x00", STR_PAD_RIGHT);
 454          $ID3v1Tag .= str_pad(trim(substr($artist, 0, 30)), 30, "\x00", STR_PAD_RIGHT);
 455          $ID3v1Tag .= str_pad(trim(substr($album,  0, 30)), 30, "\x00", STR_PAD_RIGHT);
 456          $ID3v1Tag .= str_pad(trim(substr($year,   0,  4)),  4, "\x00", STR_PAD_LEFT);
 457          if (!empty($track) && ($track > 0) && ($track <= 255)) {
 458              $ID3v1Tag .= str_pad(trim(substr($comment, 0, 28)), 28, "\x00", STR_PAD_RIGHT);
 459              $ID3v1Tag .= "\x00";
 460              if (gettype($track) == 'string') {
 461                  $track = (int) $track;
 462              }
 463              $ID3v1Tag .= chr($track);
 464          } else {
 465              $ID3v1Tag .= str_pad(trim(substr($comment, 0, 30)), 30, "\x00", STR_PAD_RIGHT);
 466          }
 467          if (($genreid < 0) || ($genreid > 147)) {
 468              $genreid = 255; // 'unknown' genre
 469          }
 470          switch (gettype($genreid)) {
 471              case 'string':
 472              case 'integer':
 473                  $ID3v1Tag .= chr(intval($genreid));
 474                  break;
 475              default:
 476                  $ID3v1Tag .= chr(255); // 'unknown' genre
 477                  break;
 478          }
 479  
 480          return $ID3v1Tag;
 481      }
 482  
 483  }