Synthesizer with a neural network that sequentially studies everyday noise and builds a unique wavetable for each user.

The slider controlling the neural network’s sensitivity to new sounds appears as eyes — the more eyes are open, the more sensitive it is; if they are closed, it doesn’t learn. The slider for selecting neuron save dates becomes a planet with satellites: the crescent shadow shows the day of the year, the moon’s position hints at the day of the month, and the satellite’s position indicates the time of day. The neural network “swells” while it learns.

The neuron that stores information about a sound similar to a new one “swells” more than the others and updates its own sound. Nearby neurons also update, but to a lesser extent, and thus “swell” less. An orange dot highlights the selected neuron, and the curve under it shows the sound fragment learned by that neuron — a wavetable used by the app to synthesize notes. For now, a logarithmic scale of frequencies and amplitudes takes the place of a keyboard.