In collaboration with CJ Carr from Dadabots (machine learning expert with raw audio), he fed 1 hour of my vocals into a neural network that generated 10 hours of new audio over 2 1/2 days. I just finished making an album with all this audio, that’s turned into a neural vocal duet. I was curious to find out if AI is smart enough to create new melodies with emotional delivery and lyrics that made sense. The end result was a lot of curation on my end, chopping up a ton of these audio samples into compositions that made sense. Lyrically and melodically it ended up being a game of ‚fill in the blank‘, making it a collaborative writing journey. Some tracks off the album are a duet of me singing back and forth with my second self. Others are pure sampling of the AI generated audio files. The track submitted here is the single off the album currently being mixed and mastered (this is rough mix). The vocals here are all the AI generated clips.