If, like Music Ally, your singing voice is akin to a circus sealion with a car horn and even your humming sounds like a fat bee trying to bench press a fridge, then technology could still make that unfulfilled pop career a reality.

Adobe has created Project Kazoo – currently in a prototype stage and based on machine learning – that uses voice conversion to turn honking and atonal sounds into something resembling music. The software can apparently turn the human voice into the sound of one of several in-tune instruments and even reconfigure the sound of one instrument into that of another.

“It listens to a voice and takes the essence of your voice, how the pitch variates and how it changes over time, and then it converts that kind of essence into music using machine learning,” Adobe research scientist Zeyu Jin tells Mashable. As it is still at the experimental stage, it is not clear if it will ever get a full public release, but don’t discount [name of genuinely terrible singer] buying it up to continue their carpet bombing of the charts.

EarPods and phone

Tools: platforms to help you reach new audiences

Tools: Kaiber

In the year or so since its launch, AI startup Kaiber has been making waves,…

Read all Tools >>

Leave a comment

Your email address will not be published. Required fields are marked *