Course-based Project
Authors
Modern browsers expose rich audio processing capabilities through the Web Audio API as it enables dynamic and immersive sound experiences. However, the tight integration of the API with hardware-level rendering introduces subtle platform-specific variations. These differences can be exploited for browser fingerprinting, an emerging privacy concern in which users are uniquely identified without their knowledge or consent. In this paper, we investigate the feasibility of audio-based fingerprinting by generating controlled frequency signals, analyzing time-domain waveforms via the Web Audio API, and applying hash-based comparisons to assess fingerprint consistency and uniqueness. We design and implement a browser-based pipeline to extract distinguishing features from system-level audio behavior. Our results show that even simple waveform configurations can result in stable and unique identifiers, with browser-specific rendering playing a critical role in fingerprinting effectiveness. These findings call for a re-examination of the privacy implications of seemingly benign web technologies.
Web Audio API, Browser Fingerprinting, Audio Fingerprinting, Browser Security, OscillatorNode