Reproduce “EBEN: Extreme bandwidth extension network applied to speech signals captured with noise-resilient body-conduction microphones” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Julien Hauret, Thomas Joubaud, Véronique Zimpfer, Éric Bavu

EBEN restores speech detail from low-pass body-conduction microphone audio by splitting signals into frequency subbands and using a compact neural generator to infer missing content. In simulated French LibriSpeech tests, listeners’ median understanding rating was 73, matching Seanet, while the median audio-quality rating was 76 versus Seanet’s 78. These results apply to simulated degradation, not real-microphone recordings.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.