Every voice is synthesised from oscillators and filtered noise at the moment you press a step, the way the original machines made them with analogue circuits. Nothing is sampled and no audio is downloaded. Click a voice name to audition it. Use the arrow keys to move around the grid and space to toggle a step.