1Real time streaming audio denoising using a [Dual-Signal Transformation LSTM Network for Real-Time Noise Suppression](https://arxiv.org/abs/2005.07551).
 2
 3Trivial to build as it uses the native rust Candle crate for inference. Easy to integrate into any Rodio pipeline.
 4
 5```rust
 6    # use rodio::{nz, source::UniformSourceIterator, wav_to_file};
 7    let file = std::fs::File::open("clips_airconditioning.wav")?;
 8    let decoder = rodio::Decoder::try_from(file)?;
 9    let resampled = UniformSourceIterator::new(decoder, nz!(1), nz!(16_000));
10
11    let mut denoised = denoise::Denoiser::try_new(resampled)?;
12    wav_to_file(&mut denoised, "denoised.wav")?;
13    Result::Ok<(), Box<dyn std::error::Error>>
14```
15
16## Acknowledgements & License
17
18The trained models in this repo are optimized versions of the models in the [breizhn/DTLN](https://github.com/breizhn/DTLN?tab=readme-ov-file#model-conversion-and-real-time-processing-with-onnx). These are licensed under MIT.
19
20The FFT code was adapted from Datadog's [dtln-rs Repo](https://github.com/DataDog/dtln-rs/tree/main) also licensed under MIT.