Introduction
This repository is trying to create a unique audio effect for guitar using a trained autoencoder. The goal is to train the autoencoder with a sample played from the guitar. Using that trained network, incoming audio will processed by the network and hopefully a useful musical effect will be applied.
Getting started
Use the record.py script to capture sound from your computer's default audio input source.
Use the train.py script to train on the captured audio files.
Use the test.py script to generate audio based on an input audio file.
Sound Experiments
Decreasing the kernal size and reducing epochs gave the sound a more distorted/fuzz effect.