Magenta, a Google research project to ‘make music and art using machine learning’, has collaborated with Google Creative Lab to create a Raspberry Pi musical instrument that uses AI to create entirely new sounds.

The project is a hardware implementation of the previously announced NSynth algorithm.

This ‘neural audio synthesis’ algorithm doesn’t blend different input sounds – instead it analyses the sonic characteristics of the sources to ‘birth’ an entirely new sound.

NSynth Super is a Raspberry Pi AI sound machine

The NSynth Super takes the NSynth algorithm and implements it in an easy-to-use hardware form. Each of the four corner dials allows you to select one of four instruments, and then you can set the level of combination by dragging your finger across the touchscreen ‘sound map’.

The custom PCB and bill of materials are open-source, as is the code, but assembly requires advanced soldering skills – see the NSynth Super GitHub page at and Andrew Back’s build guide for RS-Online. You can see the NSynth Super in action at Google.