An end-to-end lightweight multichannel speech enhancement network
A minimum implementation of A Causal U-net based Neural Beamforming Network for Real-Time Multi-Channel Speech Enhancement -- Interspeech 2021
This model acts as a weighted and sum beamformer
1 add unet_miso.py to your model directory
2 import UNet as model and ready to go 👻
- params: 1.1M
- n_fft: number of sample points for STFT
- hop_length: hop size for STFT
- mics: number of channels
Speech and noise dataset from DNS 2022 challenge, RIR is simulated for eight-microphone circular array
sisnr(dB) snr(dB) stoi
1. noisy. 7.588 11.499 0.772
2. enhanced 12.839 14.039 0.837