Audio fingerprinting
A fingerprinting technique that processes a silent audio signal through the Web Audio API and hashes the output, which differs slightly across devices.
Audio fingerprinting uses the Web Audio API to identify devices without playing any sound. A script builds an audio processing graph, typically an oscillator fed through a compressor in an OfflineAudioContext, renders it, and hashes the resulting samples. Floating-point audio processing differs subtly across hardware, operating systems, and browser builds, so the hash is stable on one machine and different on another. The whole measurement runs silently in the background.
Like canvas fingerprinting, audio fingerprinting is attractive to trackers because it needs no permissions and survives cookie clearing. It also serves as a consistency check: the audio hash should plausibly correspond to the platform the browser claims to be. For multi-profile work, the requirement is the same as for other noise-based surfaces. Each profile should produce its own stable audio output, the way distinct physical machines would, rather than blocking the API or randomizing on every load.
Oculr applies seeded, deterministic per-profile noise to AudioContext output at the engine level. The noise comes from the profile's fingerprint seed, so a given profile returns the same audio hash on every launch while different profiles diverge, matching the per-profile, deterministic behavior of its canvas and WebGL surfaces.
