How to build an audio-audio neural network?

Question

Neural networks

What methods to use and where to read about creating a neural network, the input and output of which is audio, and of different lengths and with other differences.

As an example (purely an example), at the input of audio with noise - at the output without noise, and implement it this way: give the neuron one hundred audio with noise and one hundred of the same audio, but without noise, and then purely predict.

Or at the entrance with a high voice - at the exit with a low one, in the same way. And so on and so forth.

How to do it at all? Which side to take?

Answer 1

Neuron is not needed. It is enough to determine what you need, and use the appropriate methods from digital signal processing.
I can't help you with more specific code. But I think in Python everything you need is already there.