An experiment with RNNs applied to audio in the frequency domain, to try to generate new audio.
Note that if you do not have GPUs, you will need to edit the script to install tensorflow
instead of tensorflow-gpu
, and change gpu_count
to 0.
python3 rnn-fft.py
This will start model training and then produce output showing the audio in the frequency domain, and then a plot of some of the generated audio in time domain, and also the frequency domain.
It will output generated audio as out.wav
.
Doesn't really work, but produces some agreeable noise.