AI creates never-before heard sounds

Artificial Intelligence takes a new turn as it can now create sounds that have never been heard before, all because of math joined with samples from some music instruments.

Soon enough, people would be able to listen to new sounds from their radio. Through this, the researchers hope to provide musicians with unlimited range of computer-generated instruments.

Called the NSynth, the system has been created by an engineering team Google Magenta which is a little part of Google’s push into AI.

The team explained, “Learning directly from data, NSynth provides artists with intuitive control over timbre and dynamics and the ability to explore new sounds that would be difficult or impossible to produce with a hand-tuned synthesizer.”

The system takes up samples from numerous instruments and mixes them together in a ‘sophisticated’ manner. In order to reproduce the sounds, the program then learns to recognize the audible traits of every instrument.

According to Science Alert, the information is further used for producing a mix of instruments that actually doesn’t sound like a mix of instruments in fact sounds like a single and new instrument due to the adjusted properties.

Because NSynth depends upon deep learning, the system is frequently describes as ‘artificial neural networks’. The system’s demos that are built by the team all work in real time which in turn allows fresh compositions to be created.

Wired reported music critic Marc Weidenbaum saying that the company’s new invention seems promising. “Artistically, it could yield some cool stuff, and because it's Google, people will follow their lead.”

Their research has already been published in arXiv.