PIX AI 2 // Nic Dwa Razy // DiffSinger Re-Cover

I did a new model for DiffSinger PIX yippeee !! 🐱
I really wanted to try the new model type before but I was hesitant since I wanted to record more data ... but I've accepted my fate (no new recordings at all...) and just trained him on whatever I already have ARGHH
Still! The improvement is day and night compared to his previous model, he's way more crispier and more expressive than before c:
- - - - -
Credits:
Nic Dwa Razy - Wisława Szymborska (musical interpretation by sanah)
USTX/Mix: PixPrucer
Model info:
6h of varied acoustic data total, trained with Shallow Diffusion for 100k steps
DL PIX AI 2 here ! pixproducer.wi...

[@dryjakiew]

sanah krulowa

[@frqnchyUTAU]

okay wait i jus listened to the first cover oh god this is such an improvement?? that one was already amazing PIX is an icon in this community

[@PixPrucer]

"Icon" now that's too generous to say that !!!2!!

[@Ac11dTr1p]

sounds so clean - love this cover

[@malcha]

Wonderful voice and nice cover!!!👍

[@heros-p]

Nice job, I love it

[@frqnchyUTAU]

this just makes me wanna redownload open utau for diffsinger to struggle with making him sing polish because im quirky like that

[@jjj5475]

so Amazing cover ;-;