Noise-canceling Electrostatic Headphones for fMRI

Home - Updates


Electrostatic Transducers

Noise Cancellation



About Me

Active Noise Reduction

What is Active Noise Reduction?

Noise reduction is typically a desired effect in any audio environment. The most common approach is a passive approach--using physical sound-absorbing or sound-blocking materials

Active Noise Reduction is then used to augument or replace passive systems where they are not good enough, or impractical for the application. The general concept of an active system is emitting a anti-noise signal to null out ("cancel") the unwanted noise at the desired point. The correct anti-noise signal can either be determined parametrically, or by sampling and processing existing noise. A well-implemented noise canceling system can easily outperform most practical passive systems for similar applications.

ANR Implementations

Active noise canceling may be implemented many ways, generally classified as open-loop, closed-loop, or adaptive.

In an open-loop implementation, a microphone picks up only the unwanted noise (i.e. outside of a headphone earcup), which is processed (inverted, adjusted for gain and delay) and sent to the transducer(s). Open-loop systems are the easiest to implement, but are limited in their accuracy and effectiveness.

In a closed-loop implementation, a microphone picks up both the wanted audio and the unwanted noise (i.e. inside the earcup). The original audio signal is then subtracted from the microphone signal, which is then processed, and fed back to the transducer. While more complex than open-loop, closed-loop implementations offer more accuracy, bandwidth and flexibility.

Adaptive noise canceling uses DSP algorithms to determine how best to cancel noise picked up by one or more microphones. Comprehensive algorithms paired with complex networks of microphones and transducers can be very successful at eliminating noise and providing high quality audio--however, these solutions are neceessarily expensive and laborious to implement.

The Project

Both closed-loop and adaptive approaches to noise canceling are appropriate for this project; for the sake of simplicity, I plan to begin with a simpler implementation of closed-loop ANR. I will post more details later, when I reach this part of the project.