facebookresearch / denoiser

Real Time Speech Enhancement in the Waveform Domain (Interspeech 2020)We provide a PyTorch implementation of the paper Real Time Speech Enhancement in the Waveform Domain. In which, we present a causal speech enhancement model working on the raw waveform that runs in real-time on a laptop CPU. The proposed model is based on an encoder-decoder architecture with skip-connections. It is optimized on both time and frequency domains, using multiple loss functions. Empirical evidence shows that it is capable of removing various kinds of background noise including stationary and non-stationary noises, as well as room reverb. Additionally, we suggest a set of data augmentation techniques applied directly on the raw waveform which further improve model performance and its generalization abilities.
Other
1.65k stars 302 forks source link

What are the minimum system requirements ? #124

Open the-drunk-coder opened 2 years ago

the-drunk-coder commented 2 years ago

Hi,

I tried to make the live denoiser run on my laptop (2015 Macbook Pro, i7), to no avail. I get constant message that audio processing is too slow, and there's no audio at all (tried various combinations, routing audio through blackhole, soundflower, etc).

denoise.enhance didn't finish, or at least it would have taken a long time with all CPU cores running at maximum.

I finally managed to get the enhance working on a GPU machine, and it's impressive, but on my laptop, no chance.

Is it just that my laptop is too slow ?

Best, N

adefossez commented 2 years ago

For older computers, it might be the RAM that is too slow actually. I know people with DDR3 for instance cannot run it live even though their CPU is quite fast. For non real time usage I would expect the processing time to be around the duration of the audio. Slow processing might also happen if you are trying to denoise very long file.

the-drunk-coder commented 2 years ago

Hmm my laptop has DDR3 RAM in fact ...

The file I was trying to denoise was around 3 minutes ... I thought I waited longer than that but I might try again.

What would be the expected effect if processing is too slow ? Choppy audio, or no audio at all ?