struct RiffChunkHeader

The RIFF file specification (and the child specification for WAV content)

defines the layout and contents of WAV audio files.

RIFF files consist of so-called _chunks_ (self-describing sections of the

file). These files begin with a RIFF header chunk that describes the primary

format of the file contents, followed by the data itself (in a chunk of its

own). Additional chunks may also be present, containing metadata and/or other

information to support optional features. Because all chunks include a length

field, any unknown chunks can be safely skipped by file readers.

The WAV file format specifies an initial 'RIFF' chunk of type 'WAVE' (length

24), followed by two required Subchunks: 'fmt ' (length 24) and 'data'

(length 8 + the size of the subsequent audio data). Audio data should

immediately follow these first 8 bytes of the 'data' subchunk. Once the

entirety of audio data has been written into the file, the 'length' field for

the 'data' subchunk should be updated with the number of bytes of audio.

Likewise, the overall length for the parent 'RIFF' chunk (which conceptually

contains the two 'fmt ' and 'data' subchunks) must be updated at this point,

to describe its total size (including subchunk headers and the audio data).

Thus, although all audio data follows the file headers, we must update the

headers once all audio has been written.

** Note, lest our RiffChunkHeader struct definition mislead the uninformed **

These struct definitions actually conceptually relocate the final 32-bit

value of the initial RIFF chunk (the RIFF format-type) into the subsequent

'fmt ' subchunk instead. Because the sequence of fields is maintained, this

does not create a problem. We do this so that we can reuse our RIFF struct

definition for the 'data' subchunk as well.