no code implementations • 25 Jun 2021 • Amir Ivry, Israel Cohen, Baruch Berdugo
In this paper, we propose a residual echo suppression method using a UNet neural network that directly maps the outputs of a linear acoustic echo canceler to the desired signal in the spectral domain.
no code implementations • 25 Jun 2021 • Amir Ivry, Baruch Berdugo, Israel Cohen
A deep neural network, which is trained to separate speech from non-speech frames, is obtained by concatenating the decoder to the encoder, resembling the known Diffusion nets architecture.
no code implementations • 25 Jun 2021 • Amir Ivry, Israel Cohen, Baruch Berdugo
To mitigate this mismatch between training data and real data, we simulate an augmented training set that contains nearly five million utterances.
no code implementations • 25 Jun 2021 • Amir Ivry, Israel Cohen, Baruch Berdugo
Second, the network is succeeded by a standard adaptive linear filter that constantly tracks the echo path between the loudspeaker output and the microphone.