Welcome to my blog. It will be focused on some AI experiments related to Audio and Music.
I want to start with my first software to use AI called “AI Audio player”. Couldn’t think of a better name, let’s leave that for later.
Anyway, this Windows OS app plays audio files and applies a CNN model on almost realtime. It uses a 56 million parameter Unet with skip connections.
The results are pretty decent for something I consider lightweight on the CPU usage (although memory usage can be bigger).