Description

made by me! houston has 40+ minutes of training data on 100 epochs with mangio-crepe. its mostly his assault lines so he constantly sounds like he's yelling. probably not too good for softer songs but feel free to do so if you really wanna. < hoxton has 29 minutes of training data on 300 epochs with crepe. hoxton also shouts alot. i think he sounds pretty okay on softer songs though. i encourage you to fiddle with the conversion strength slider for him. i think you can get away with significantly lowering the conversion strength if the singer already has a british accent, helps reduce the crustiness on higher levels... < they both sound pretty good on higher vocals but after a certain point they kinda break, so -5 to -10 semitones for female vocals is best.