mizapf said that a future update to the implementation will use synthesized waveforms and filters, rather than stitching together samples, which I'm looking forward to seeing/hearing.

LN