Voice-to-MIDI synthesis requires instant processing to be usable. Tampered code frequently disrupts ASIO or CoreAudio driver paths, introducing massive delay between your voice and the MIDI output.