Transcribing polyphonic audio is a hard challenge.

Currently we can detect two musically independent voices (which is one per hand for a piano song).

But in the editor you can create up to two voices per staff (this means four voices for piano songs).