suno.ai now offers “stems”.
i.e. voice and playback are separated with - guess what - a different ai algorithm.
unfortunately, the voice is not (re-)rendered solo.
unfortunately, the stems algorithm does not intervene too convincingly.
the vocals sound flat after the separation and vocal artifacts can still be heard on the “instrumental”.
let's hope that #sunoai improves soon and then also offers bass / drums / other instruments as additional tracks.
.
