Spiritbox - Ultraviolet [Vessel AI Cover]

Created an 11 minute dataset of Vessel's vocals and trained my own model for this. Might add to the dataset and see how much better I can get it without studio quality multi tracks. Current model is at 500 epochs. --
-- The isolated Spiritbox vocals didn't come out super clean in some parts because of how they're layered/mixed. That makes the conversion to Vessel's voice model get glitchy as you'll hear. I did Too Close / Too late by Spiritbox and it came out much better: • spiritbox - too close ...