Lesson 06 · Walking chunks and skipping the unknown

Walking chunks and skipping the unknown

A real WAV file often carries chunks you do not care about between fmt and data. Today you walk the whole chunk list in order, stepping over anything unfamiliar to find the two chunks you need.

The goal

Iterate every chunk after the header, collecting ids in order and locating fmt and data.

Start here - the target

TO DO

Scenario: The walker visits every chunk and finds fmt and data

Givena WAVE body with chunks in order: fmt (16 bytes), then JUNK (4 bytes), then data (8 bytes)

Whenthe chunks are walked from just after the 12-byte header

Thenthe visited ids in order are ["fmt ", "JUNK", "data"]

Andthe fmt chunk is found and the data chunk payload is the 8 data bytes, with JUNK skipped

Background

The fmt and data chunks are the only two a minimal WAV reader must understand, but files in the wild are full of others - LIST (metadata), fact, cue , JUNK (padding placeholders), and vendor-specific tags. A robust reader walks the chunk list from the first chunk after the header, and for each chunk it either handles it or steps over it using the size field. That is the whole point of a self-describing container: you never have to know a chunk to skip it.

The walk is a loop: read a chunk, record it, advance the offset by the bytes that chunk consumed (its 8-byte header plus its payload), repeat until you run out. To find data, walk until you see that id and keep its payload. This loop is the backbone of the reader - next lesson hardens it against the one alignment quirk RIFF has, the padding byte.

Make it work

// start just past the 12-byte header; readChunk returns consumed bytes
func walk(b []byte) []Chunk {
  var out []Chunk
  off := 12
  for off+8 <= len(b) {
    c, used, _ := readChunk(b[off:])
    out = append(out, c)
    off += used   // advance to the next chunk
  }
  return out
}

CheckpointDONE

You can walk every chunk and skip ones you do not recognise. Commit and stop here.