Under the hood, it does just cut together the first 50ms (after pitch shifting) of 32 sound effects, and puts them on 50ms alignments.
So, it is slightly similar, but not really.

Converted a song into a datapack using playsound/stopsound, suprisingly the lyrics are quite understandable.
(volume warning?)
www.mediafire.com/file/4qqku9d...