OptimiseLab uses AI to predict ‘music popularity scores’


Another day, another AI/music startup – this one hailing from Toronto and called OptimiseLab. It’s using the slogan ‘538 for music’ – a reference (we think) to well-known statistical-analysis blog FiveThirtyEight, run by data-journalist Nate Silver.

“We are building patent pending AI technology to analyse, improve, and compose music,” explains the company’s profile on investment site AngelList. “Our first product Timbre is for music producers. Timbre accurately predicts music popularity scores and provides customised advice on song improvement instantly.”

The company’s website provides more information: it claims to have trained its AI system on “tens of millions of songs” across all genres to build a “powerful, gigantic music knowledge base” which will then be used to benchmark the quality of new songs.

Oh, and then: “Better yet, Timbre generates customised tuning advice: delay the fade-in for another 2 seconds, repeat the 17th bar one more time, lower the pitch in segment 5 by 20%, and you’ve got yourself a more appealing track”.

The obligatory question here is whether rating new songs purely on what’s gone before risks missing out on tracks that break new ground. We’ll find out once Timbre is out in the wild being used.

Update: OptimiseLab is now known as Resonoo.

Stuart Dredge

Read More: News Startups
Leave a Reply

(All fields required)