Demo · libdse
Speech Enhancement using Deep Learning
Upload a noisy speech recording, select a model, and listen to the enhanced result. Spectrograms are rendered in real-time directly in the browser.
Wave U-Net is currently untrained. Selecting it will produce degraded output.
or
0:00
Input · Noisy
Hz ↑