Listen to an AI sing an uncannily human rendition of ‘Jolene’

AI-powered image generators are taking over printers lately. But musical machine learning models have quietly made great strides in recent years. Holly Herndon has been at the forefront of that revolution. For her last album she co-developed (with partner Matt Dryhurst) the vocal neural network Spawn. Proto And Holly+ (in partnership with Never Before Heard Sounds) was released to the public last year, allowing anyone to use a model of Holly’s voice. And now she’s released a new single, the only vocals of which come from her digital twin.

This cover of Dolly Parton’s “Jolene” plays it pretty straight on first listen. Yes, it’s slow and quiet, but Ryan Norris, who handles the instrument, doesn’t take any extreme liberties with the arrangement or the soundscape. It simply replaces frantic despair with plaintive resignation.

Also Read :  A Rubens Masterpiece Could Fetch $35 Million at the Auction of Divorcing Couple Mark Fisch and Rachel Davidson's $177 Million Baroque Art Collection

It’s so arresting that every vocal sound is generated by Holly+, right down to the sharp inhales before the harmonies begin. (That’s right, it “breathes”.) There’s not a human in sight of a vocal booth here. Some phrases are a bit sticky and if you listen closely, there are occasional digital artefacts. But overall, this digital model of the real Holly Herndon’s voice is impressive in its ability to mimic its creator.

Also Read :  Book review of Cheap Land Colorado: Off-Gridders at America’s Edge by Ted Conover

Until now, most mainstream artists’ experiments with AI have focused on creating generative soundscapes or synth melodies. This is the first time (as far as I know) that a machine learning model has taken the mic solo of a pop song.

Herndon previewed the track at the Sonar festival in March, but it largely flew under the radar until a proper release this week. (BTW: Watch the presentation from Sonar for a truly wild real-time demonstration of Holly+ and Never Before Heard Sounds’ technique.) You can try to recreate the performance above by recording your own performance of Jolene and uploading it to Holly+. Don’t expect the same reliability of results via the web app. It’s definitely a fun diversion, and artists serious about using AI to further their craft should explore Spawning, an organization launched by Herdon and Dryhurst earlier this year.

Also Read :  Takeoff killed - latest: Kanye West pays tribute to Migos rapper after new video shows Houston party

All products recommended by Engadget are selected by our editorial team, independent of our parent company. Some of our stories include affiliate links. If you purchase something through one of these links, we may earn an affiliate commission. All prices correct at time of publication.


Leave a Reply

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

Related Articles

Back to top button