Reading id3v2 tags for MP3 files isn't too bad, but I can already tell that writing will be a bit of a challenge. id3v2 tags are divided into frames, with each frame having a sort of name/value pair. In id3v1, all of the values are strings. In id3v2 values can be any data at all.

One of the goals of the id3v2 component is to separate it from AudioMan so it may be used elsewhere. So I didn't want dependencies on AudioMan-specific things. As well the id3v2 spec has changed quite a bit (4 revisions so far) over the years which can lead to problems using some of its own information.

Something that changed a lot in the id3v2 spec were the frame id strings, which started as three character and then went to four. These inconsistencies between versions makes using the id strings in the API difficult, so I've decided not to do that.

Instead I've used a simple Map of name/value pairs. The name is a string for the field, like "TRACK_NUMBER" and the value is a string. So when the API is used, the data must be converted from AudioData to this format to id3v2, or the reverse. This creates a very loose dependency from AudioData to id3v2 and a loose dependency between revisions of id3v2.

Another issue is when I write data to the id3v2 tag. There will be a lot of frames I won't have to change, so I'll just copy them from the original instead of remaking them from scratch. This will prevent me from deleting data I don't understand and don't have any business trying to rewrite.