6 min read
The WAV file format explained: chunks, headers and why it never dies
A format from 1991 that still runs every film set. Here is what is actually inside the file you hand to post.
WAV is the plainest professional audio container in use. No compression, no licensing, no surprises. That simplicity is exactly why every recorder, DAW and delivery spec still asks for it.
A WAV is a list of chunks
- RIFF header: four bytes saying RIFF, the file size, then WAVE.
- fmt chunk: sample rate, channel count, bit depth and encoding.
- data chunk: the raw interleaved PCM samples, nothing else.
- Optional chunks: bext, iXML, cue, LIST. Production metadata lives here.
Because chunks are self describing, a player that does not understand iXML simply skips it. That is why a BWF file from a Sound Devices recorder opens in a consumer player without complaint.
The 4 GB wall and RF64
Classic WAV stores sizes in 32 bit fields, so a file stops at roughly 4 GB. At 48 kHz, 24 bit, eight channels, that is about two hours. Long interviews and concert recordings hit the wall, which is why recorders switch to RF64 or WAV64, formats that widen those fields while keeping the same chunk structure.
Interleaving is the part people forget
Inside the data chunk, samples alternate channel by channel: L, R, L, R. In an eight track poly WAV the pattern repeats every eight samples. Splitting a poly file is simply reading that pattern and writing each channel out on its own, which is why a correct split is bit identical to the source.
Why WAV survives
- Lossless by definition, so generation loss never accumulates.
- Sample accurate editing with no decoder latency or priming samples.
- Metadata chunks carry scene, take, timecode and track names.
- Every NLE, DAW and archive system reads it without a plugin.
Inspect a WAV in the browser to see its header fields, channel layout and metadata before it reaches the edit. Nothing is uploaded, the parse happens on your machine.