I'm trying to figure out how binary data is embedded in Ogg Vorbis comment tags. Specifically, COVERART pictures, such as the way Media Jukebox v8 does. According to the Ogg Vorbis comment specification the contents of the comment field is supposed to be "8 bit clean UTF-8 encoded field contents to the end of the field", but that doesn't make . The sample file I'm working with has 91400 bytes of data that looks to be uuencoded or base64 encoded, but scattered throughout those 91400 bytes are 783 bytes outside the acceptable 64 character set (A-Z, a-z, 0-9, /, +), usually just one or two of each unacceptable value, but with 280 counts of 0x00 and 340 of 0xFF.
The best luck I've had so far is not UTF8-decoding the comment contents, then uudecoding (using the function I found here - PHP's built-in base64_decode() seems to die when it encounters an invalid character), but those invalid charcters are still throwing it off...
I assume I'm doing something wrong in the decoding - is there some documentation somewhere for COVERART comments, both embedded and linked? Are there any other programs that read embedded images?
Original:
Decoded (sort of):
Author
Topic: Encoding of embedded COVERART pics in Ogg? (Read 694 times)