metadata
license: gpl-3.0
A dereverb model specifically for mono vocal room reverb.
Model type: bs_roformer
Channels: mono
Reverb in training data: only convolutional reverbs, generated with pyroomacoustics
Example:
- input.flac
- noreverb.flac
- reverb.flac
for refercence dereverb_mel_band_roformer_mono got SDR: 7.6685 on same valid set.