Processing bytes instead of numpy array (Using PyAudio)

[qalabeabbas49]

Hi, I am working on a project where I have audio data as bytes stream (from PyAudio). I want to use the denoiser to clean the audio in real-time (using live.py).

My question is it possible to process the bytes instead of the NumPy array? If so, what should be my approach, and which parameters do I need to change?

Any leads are much appreciated.

Thanks

[adiyoss]

At the moment the model gets as input a numpy array representing the raw-waveform. If you want to use bytes instead, the easiest way imo would be to convert it to numpy.