This is a PoCoNet like model for noise suppression to make speech cleaner. The model is based on PoCoNet architecure and trained on DNS-Challenge dataset paper. The network works with mono audio sampled on 16kHz. The audio processed iterative by patches with 2048 size. On each iteration it takes 2048 (128ms) samples as input and returns 2048 (128ms) samples as output with 640 (40ms) samples delay. In addition the network required 50 state tensors to make processing seamless. On the first iteration these state tensors have to be filled with 0. On the consequences iterations theses tensors have to be taken from corresponding outputs of previous iteration. You can try Noise Suppression Python* Demo to see how it works.
to process 2048 samples that is 128ms for 16kHz
| Metric | Value |
|---|---|
| GOps | 1.2 |
| MParams | 7.22 |
| Source framework | PyTorch* |
The SISDR quality metric was calculated on the 100 dev test synthetic speech clips from DNS-Challenge 2021 dataset.
| Metric | Value |
|---|---|
| SISDR for input noisy signal | 11.73 dB |
| SISDR for output cleaned signal | 20.54 dB |
| SISDR increase | +8.81 dB |
Sequence patch, name: input, shape: 1, 2048, format: B, T, where:
B- batch sizeT- number of samples in patch
input states, names: inp_state_*, should be filled by corresponding out_state_* from previous step
Sequence patch, name: output, shape: 1, 2048, format: B, T, where:
B- batch sizeT- number of samples in patch Note: The output patch is "shifted" by 640 (40ms) samples in time. So output[0,i] sample is synced with input[0,i-640] sample
output states, names: out_state_*, should be used to fill corresponding inp_state_* on next step
[*] Other names and brands may be claimed as the property of others.