Page 1 of 1
Current "known issues" in getID3()
Posted: Mon Apr 01, 2002 1:43 am
- MP4 support incomplete (audio-only works some of the time, video works none of the time). Current implementation may report a large number of unknown quicktime atoms.
- MPEG audio+video streams may have inaccurate bitrate & playtime (usually within 0.1% but may be off by as much as 2%) - need better documentation
- Cannot determine bitrate for MPEG video with VBR video data (need documentation)
- AAC-RAW mode files cannot be identified
- WavPack-RAW mode files cannot be identified
- Encrypted ASF/WMA/WMV files warn about "unhandled GUID ASF_Content_Encryption_Object"
- Bitrate split between audio and video cannot be calculated for NSV, only the total bitrate. (need documentation)
- All Ogg formats (Vorbis, OggFLAC, Speex) are affected by the problem of large VorbisComments spanning multiple Ogg pages, but but only OggVorbis files can be processed with vorbiscomment.
- Matroska files with AAC audio are not correctly handled
- Type-1 DV AVI files do not return resolution information (need documentation)
Known Bugs/Issues in getID3() that cannot be fixed
Posted: Wed Sep 21, 2011 3:53 pm
- 32-bit PHP installations only:
- Files larger than 2GB cannot always be parsed fully by getID3() due to limitations in the 32-bit PHP filesystem functions.
NOTE: Since v1.7.8b3 there is partial support for larger-than-2GB files, most of which will parse OK, as long as no critical data is located beyond the 2GB offset.
- all file formats on 64-bit PHP
- ZIP (format doesn't support files >2GB)
- FLAC (current encoders don't support files >2GB)
- ID3v1 tags (always located at end-of-file)
- Lyrics3 tags (always located at end-of-file)
- APE tags (always located at end-of-file)
- Quicktime (will work if needed metadata is before 2GB offset, that is if the file has been hinted/optimized for streaming)
- RIFF.WAV (should work fine, but gives warnings about not being able to parse all chunks)
- RIFF.AVI (playtime will probably be wrong, is only based on "movi" chunk that fits in the first 2GB, should issue error to show that playtime is incorrect. Other data should be mostly correct, assuming that data is constant throughout the file)
Known Bugs/Issues in other programs
Posted: Wed Sep 21, 2011 3:58 pm
- Windows Media Player (up to v11) and iTunes (up to v10+) do not correctly handle ID3v2.3 tags with UTF-16BE+BOM encoding (they assume the data is UTF-16LE+BOM and either crash (WMP) or output Asian character set (iTunes)
- Winamp (up to v2.80 at least) does not support ID3v2.4 tags, only ID3v2.3 see: http://forums.winamp.com/showthread.php?postid=387524
- Some versions of Helium2 do not write ID3v2.4-compliant Frame Sizes, even though the tag is marked as ID3v2.4) (detected by getID3())
- MP3ext V3.3.17 places a non-compliant padding string at the end of the ID3v2 header. This is supposedly fixed in v3.4b21 but only if you manually add a registry key. This fix is not yet confirmed. (detected by getID3())
- CDex v1.40 (fixed by v1.50b7) writes non-compliant Ogg comment strings, supposed to be in the format "NAME=value" but actually written just "value" (detected by getID3())
- Oggenc 0.9-rc3 flags the encoded file as ABR whether it's actually ABR or VBR.
- iTunes (versions "X v2.0.3", "v3.0.1" are known-guilty, probably other versions are too) writes ID3v2.3 comment tags using a frame name 'COM ' which is not valid for ID3v2.3+ (it's an ID3v2.2-style frame name) (detected by getID3())
- MP2enc does not encode mono CBR MP2 files properly (half speed sound and double playtime)
- MP2enc does not encode mono VBR MP2 files properly (actually encoded as stereo)
- tooLAME does not encode mono VBR MP2 files properly (actually encoded as stereo)
- AACenc encodes files in VBR mode (actually ABR) even if CBR is specified
- AAC/ADIF - bitrate_mode = cbr for vbr files
- LAME 3.90-3.92 prepends one frame of null data (space for the LAME/VBR header, but it never gets written) when encoding in CBR mode with the DLL
- Ahead Nero encodes TwinVQF with a DSIZ value (which is supposed to be the filesize in bytes) of "0" for TwinVQF v1.0 and "1" for TwinVQF v2.0 (detected by getID3())
- Ahead Nero encodes TwinVQF files 1 second shorter than they should be
- AAC-ADTS files are always actually encoded VBR, even if CBR mode is specified (the CBR-mode switches on the encoder enable ABR mode, not CBR as such, but it's not possible to tell the difference between such ABR files and true VBR)
- STREAMINFO.audio_signature in OggFLAC is always null. "The reason it's like that is because there is no seeking support in libOggFLAC yet, so it has no way to go back and write the computed sum after encoding. Seeking support in Ogg FLAC is the #1 item for the next release." - Josh Coalson (FLAC developer) NOTE: getID3() will calculate md5_data in a method similar to other file formats, but that value cannot be compared to the md5_data value from FLAC data in a FLAC file format.
- STREAMINFO.audio_signature is not calculated in FLAC v0.3.0 & v0.4.0 - getID3() will calculate md5_data in a method similar to other file formats, but that value cannot be compared to the md5_data value from FLAC v0.5.0+
- RioPort (various versions including 2.0 and 3.11) tags ID3v2 with a WCOM frame that has no data portion
- Earlier versions of Coolplayer adds illegal ID3 tags to Ogg Vorbis files, thus making them corrupt.
- Meracl ID3 Tag Writer v1.3.4 (and older) incorrectly truncates the last byte of data from an MP3 file when appending a new ID3v1 tag. (detected by getID3())
- Lossless-Audio files encoded with and without the -noseek switch do actually differ internally and therefore cannot match md5_data
- iTunes has been known to append a new ID3v1 tag on the end of an existing ID3v1 tag when ID3v2 tag is also present (detected by getID3())
- MediaMonkey may write a blank RGAD ID3v2 frame but put actual replay gain adjustments in a series of user-defined TXXX frames (detected and handled by getID3() since v1.9.2)