What is this exactly? o.o
I don't know if any of these characters are capable of speaking normally. I wonder if characters from anime can be sampled? If you could get them to speak you could replace their voices in anime.
... separating vocal from BGM requires that the two are separate tracks
we have the technology to remove the vocals from a recorded song (these devices are available at the consumer level, if ya know where to look). It would be a simple matter to reprogram that same device/software to remove BGM and SFX from a recorded track, thus isolating the vocals.