Unofficial PyTorch implementation of the paper "HIGH FIDELITY SPEECH ENHANCEMENT WITH BAND-SPLIT RNN" (https://arxiv.org/abs/2212.00406) on VCTK-DEMAND Dataset (https://datashare.ed.ac.uk/handle/10283/2791).

Choosed parameter settings

N (feature dimension) : 64, L (the number of lstm layers) : 5

Audio files are in saved_tracks_best folder.

Used packages are can be installed by:

pip install -r requirements.txt

Download VCTK-DEMAND dataset (https://datashare.ed.ac.uk/handle/10283/2791), change the dataset dir:

-VCTK-DEMAND/
  -train/
    -noisy/
    -clean/
  -test/
    -noisy/
    -clean/

python train.py --data_dir <dir to VCTK-DEMAND dataset>

If you want to adjust parameters (N, L) of the model, change the value of variables in train.py.

self.model = BSRNN(num_channel=64, num_layer=5).cuda()

python evaluation.py --test_dir <dir to VCTK-DEMAND/test> --model_path <path to the best ckpt>

• https://github.com/ruizhecao96/CMGAN (MIT License)