Data & Demos

Data Sets

Wearable Microphone Impulse Responses

An open-access dataset of acoustic impulse responses measured at 160 positions on the body and various wearable accessories from 24 directions of arrival. (Blog post)

Massive Distributed Microphone Array Data Set

An open-access dataset of speech clips from 10 loudspeakers recorded by 160 microphones on wearable and smart-speaker arrays spread throughout a large, reverberant room. (Blog post)

Video Demonstrations

Turning the Music Down with Wireless Assistive Listening Systems

[ASA Meeting 2022] We used a wireless assistive listening system, usually used to enhance the broadcast sound, to cancel it instead. The system removes the nuisance sound being played over the speakers.

Immersive Remote Microphones with Tympan

[ASA Meeting 2021] We implemented our immersive multiple-talker remote microphone system on the Tympan embedded hearing aid platform. The adaptive filter removes noise from the fan while preserving the spatial cues of each talker, even when they move. Best experienced with headphones.

Adaptive Remote Microphone

[WASPAA 2021] A binaural adaptive filtering system preserves the spatial cues of a moving talker while reducing background noise. Best experienced with headphones.

Illinois RapidAlarm

The Illinois RapidAlarm is a sensor and alarm system for use with emergency ventilators such as the Illinois RapidVent. It tracks breaths using a pressure sensor and alerts clinicians when it detects a problem.

Face Mask Acoustics

Face masks muffle speech, making it harder to communicate, but not all face masks have the same effect. This video compares the acoustic attenuation of different face masks and demonstrates that lapel microphones can improve clarity.

Mic-Array-Na

[WASPAA 2019] A motion-robust beamforming algorithm can separate speech even when the microphones in the array are moving relative to each other. Best experienced with headphones.

Cooperative Listening

[CAMSAP 2019] Demonstration of cooperative processing using wearable and smart-speaker devices. A distributed array of 160 microphones is used to separate sources and learn room acoustics parameters, then a wearable array of 16 microphones is used to perform real-time binaural beamforming. Best experienced with headphones.

Audio Demonstrations

Cooperative Listening Enhancement

Demonstration of cooperative source separation and listening enhancement using a massive-scale distributed array and wearable devices. The method is described in our CAMSAP 2019 paper.

Deformable Microphone Array

Demonstration of source separation using a wearable microphone array on a moving human subject from our WASPAA 2019 paper.

Asynchronous Source Separation

Demonstration of asynchronous source separation using hat-mounted microphone arrays on moving listeners. The algorithm is from our entry into SiSEC 2018 ASY and an IWAENC 2018 paper.