Tech giant Google has been working on a fusion of artificial intelligence and music called Project Magenta for some time now, and the project has finally yielded a nuanced, useful tool for musicians in the form of the NSynth Super. The NSynth Super is a synthesizer board of sorts that uses machine learning and Magenta’s extensive sound database to allow musicians to essentially craft new soundboards and voices for synthesizers, drum machines, and the like. The kicker is that it can be done on the fly using the NSynth Super’s front touch panel and dials, which means that musicians who are inclined to do so can craft entirely new songs or put a new spin on old favorites in mid-concert.
The NSynth Super is based on a neural synthesizer algorithm, called NSynth, that essentially uses machine learning to not blend two sound profiles, but to “learn” what makes them unique, and combine them to create an in-between sound. NSynth uses acoustic profiling through waveforms, so a wailing electric guitar combined with the sound of a babbling brook is just as much fair game as blending the voices of two different singers. Google Brain and the team behind Magenta crafted NSynth from the open-source Magenta project, and that work was used by Google Creative Labs to materialize it as the NSynth Super.
Any DIY musicians out there who want an NSynth Super of their very own will be pleased to know that Google won’t charge them for it – but they will have to download the open source project files and templates, gather the materials, and build the instrument themselves. Google has produced only one prototype of the NSynth Super, and does not plan to sell it. Like NSynth and the files for the NSynth Super, the entirety of Project Magenta is open-source and readily available on GitHub, which means that enterprising coders and musicians can create and discover brand new things within the project’s framework. Magenta is based on machine learning, but once you have enough AI know-how and a powerful enough machine to work with the project in its original form, the only real limit as far as things to have it do with audio and music is your own imagination.