usage.md

Usage:

This demo takes a speech sample and a noise sample and mixes them at the provided signal to noise ratio (SNR). You can either record a speech sample or alternatively provide one via upload. Furthermore, you may upload a noise sample which will be mixed with the speech sample. If no samples are provided, a default will be used. Long audio samples will be trimmed to 10s.

DeepFilterNet (link) is used to denoise the noisy mixture.

Note:

Before you edit the inputs (i.e. record a new sample), you need to clear all inputs and outputs. Otherwise the output audio will not be updated.