Closed Pocwiardo closed 2 weeks ago
The tool itself can be used as a standalone app.
It processes audio in real time, and we get to choose the buffer size and if we want to use mono or stereo.
At the input, we get a Tensor (its shape depending on the mentioned parameters) and the values of the parameters the user chose.
We can define 4 custom parameters, and only knob parameters are available for now.
All the magic must happen within a forward
function, which has to return a modified Tensor.
The library provides an easy-to-use function to export the model into a usable .nm
file, which can be loaded into the Neutone application.
Run Neutone in python, configure the environment and find out how it works.