Remove a speaker's voice from a recording using ultrasound

What if you could effectively prevent someone from recording your voice? This is the subject of a study by Guo et al. (2022) at Michigan State University, in which they use a dynamically computed audio signal that effectively cancels voice in a recording device. This relies on an interesting aspect of some Micro-Electro-Mechanical System (MEMS) microphones, which are commonly used in smartphones and other recording devices.

Pressure sensitivity of a MEMS microphone. (credit: Brian R. Elbing)

A specially crafted ultrasonic signal sent to the same microphone recording his voice can cause the voice audio signal to disappear in the final recording. The approach taken by the authors is to use a neural network trained on voice samples of the person (“Bob”) whose voice is to be canceled. After recording Bob's voice during a conversation, the creatively named Neural Enhanced Cancellation (NEC) system determines which ultrasonic signal to send to the target recording device. During this time, the person holding the recording device ("Alice") will still hear Bob's voice normally.

Because ultrasound is highly directional, the system can only block a specific microphone and will not affect hidden microphones in a room. As the authors noted, it is possible to perform general microphone jamming using other systems, but this poses a legal problem, which should not be an issue with their NEC system. p>

Thanks to [JohnU] for the tip!