Google introduced a tool for finding new sounds with the help of AI
Research project Google is called Magenta, which is using AI helps people to show their creativity, presented experimental musical instrument open source NSynth Super. It uses a neural network NSynth able to generate different sounds.
NSynth Super — device that explores the potential of neural networks. Says research scientist at Google Brain, Douglas Eck (Douglas Eck), NSynth of creating notes, and directly the sounds of the instrument. The algorithm examines the key quality that what comprises a single sound, and can combine these sounds, thus creating something completely new.
The device that Google offers to experiment with the algorithm that has area control, each sector which can be bound to a specific tool. The sounds of these instruments can be combined, finger on the pad. But NSynth not just Super impose on each other existing sounds system synthesizes a completely new, based on the acoustic qualities of individual instruments.
NSynth Super can work with any source of MIDI sound, whether it’s a digital audio workstation, sequencer or a keyboard synthesizer. The device allows you to configure some audio parameters — for example, attack, decay, support and damping.
Google does not allow to buy NSynth Super, but offers instructions for creating their own devices based on the Raspberry Pi. There is even information on how to make the printed circuit Board and the housing.