AudioTech Refiner: Eliminating Noise and Distortion in Low-Bandwidth Audio

Introduction

For many of us, the only way to stay connected during the pandemic is through online calls. However, the quality of these calling experiences is often less than ideal. Issues such as audio latency, choppy voice, unpleasant noise, and even dropped calls are common and never resolved.

For this project, we will explore possible ways to use deep learning to improve the quality of online phone calls, by 1) reducing the bandwidth required to transfer audio data, and 2) using image-to-image translation to refine the details of audio.

We are looking for one or two developers with experience in research, audio, NLP, and/or computer vision to join us. Email richardxuyixuan@gmail.com or DM “Richard Xu” in the UTMIST Discord if you’re interested!

Proposal