Each version is pored over. There is cover after cover. People train AI models to improve the quality of shoddy vocals.The desire to hear the song in high quality is so immense it is often described as painful.

↩︎ The Guardian

View Post →