Uh, looks like I've found a bug. I have a lot of mp3s with ID3 tags written by EAC. The SYLT plugin seems to be incorrectly interpreting them when trying to get data from lrcdb.
For example, a couple of the tags might look like this:
Code:
TIT2(00)(00)(00)(09)(00)(00)(00)La CercaTPE1(00)(00)(00)(07)(00)(00)(00)SpartaTALB(00)(00)(00)(10)(00)(00)(00)Porcelain
where the two-digit numbers inside parentheses describe unprintable bytes in decimal.
For some reason, the SYLT plugin is asking lrcdb for "(00)La Cerca" instead of just "La Cerca". At first, I thought that the ID3 tags were wrong, that the null terminator existed at the beginning of the text instead of the end, but the ID3v2.3 spec doesn't require null termination on strings and the leading one is actually stating that the text is in ISO-8859-1 and not Unicode.
I think that the SYLT plugin is somehow forgetting to interpret the encoding byte and including it in the text string itself.
Reference:
ID3v2.3 spec §4.2To deconstruct the first frame there:
TIT2 : frame type
(00)(00)(00)(09) : frame data size
(00)(00) : flags
<-- End of frame header -->
(00) : text encoding
La Cerca : actual textual data
Previous Topic
Index
