go-audio / riff Goto Github PK

When calculating the duration with the wavDuration() method, the size used is Parser.Size. In Parser.Parse() this is set to be the RIFF chunkSize, i.e. the file size - 8 bytes. This means that the header chunks are included in the calculation, giving a duration that is too long. The size that should be used is the size from the data chunk.

In most cases the added duration will be small enough that you won't notice, but if you need to know the exact number of samples in the file, or you have a file with large amounts of meta data in the headers (which you could typically get with BWF and derivatives), it becomes a problem.

When fixing the parsing in wav for DecodeListChunk I tripped over the fact that Chunk.Size doesn't represent the head value as described:

Chunk represents the header and containt of a sub block

Instead it also includes the padding byte, which makes parsing more complex as you need to check for a remaining byte and ignore it instead of being able to simply rely on io.EOF.

I would instead suggest that the extra byte is handled by Chunk.Drain.