Audio you can put in a receipt.
The same samples, everywhere.
wai.audio.int_codec is a learned audio codec whose whole decode is integer:
a stack of integer 1-D transposed convolutions and integer IGDN, run over a latent recovered by
an integer rANS decoder. This page runs the reference implementation's decoder, compiled to
WebAssembly, with no ML runtime and no floating point. So the 16-bit PCM is byte-identical
on every conforming sink, and the page re-checks its checksum against the native reference
before playing it. That is sample-equivalence. It is the audio counterpart of the
learned image codec wai.neural.int_hyper.
Why the whole decode is integer
Before the waveform played, the page hashed the decoded PCM and checked it against the checksum the native reference recorded; the badge shows the result. The transposed convolutions are integer accumulation with an arithmetic-shift requantize, IGDN uses an integer square root, and the latent comes off an integer rANS decoder, so nothing in the decode depends on the machine.
The model
A small codec trained on a generated signal (tones, chirps, noise bursts and speech-like formant sweeps, written by its export tool from a seed), then post-training-quantized to integers; the page shows its size once fetched (— KiB). The clip it decodes here is another signal of the same kind from the same seed. It demonstrates the integer decode.
Where the code lives
The capability and its integer decoder are in the standard; the reference engine is
Apache-2.0 in the open-standards repository. Its test vectors live with the reference implementation
(wai-rs/tests/vectors/int_audio). The clip on this page is generated, not recorded, and its export tool takes no audio input. The PCM hash shown is 64-bit FNV-1a.