Google brings sounds that have never been heard before with AI

June 5, 2017, 2:32 p.m. By: Vishakha Jha

AI Sound

Artificial Intelligence explores the connection to engineering and technology. It is a set of tools that are driving forward key parts of the futurist agenda, at a rapid pace. Researchers have been providing us with all sorts of wonderful things with AI, from directing adverts, virtual personal assistants, video games, smart home devices, and so much more. Recently AI is generating brand new sounds that have never been heard before.

The new system, called NSynth( Neural Synthesiser), developed by Magenta, a part of Google Brain where a team of researchers works on neural networks and other forms of machine learning. Now the Magenta is using neural networks as a way of teaching machines to introduce new kinds of music and other art. Neural networks are complex mathematical systems that can learn tasks by analyzing large amounts of data.

NSynth relies on deep learning and begins with a massive database of sounds. The team has collected a wide range of notes from different instruments and then fed them into a neural network. By analyzing the notes, the neural net assessed the audible characteristics of each instrument and created a mathematical vector for each one. Using these vectors, a machine can produce the sound of each instrument and it can also blend the sounds of the two. The properties of audio are adjusted to create something that sounds like a single, new instrument rather than a mash of multiple sounds.

Research paper has been released describing the NSynth algorithms and anyone can download and use their database of sounds new compositions to be created. The team has opened up the world to a plethora of new possibilities through AI for any artist, not just musicians.

Image Source: themerkle