Stability AI music platform welcomes three-minute “song” creation

Stability AI has launched the latest generation of its music creation platform, capable of producing three-minute long “song” tracks.

Stable Audio 2.0 is a new version of the software that allows users to convert text prompts into extended audio. Stable Audio takes the form of a maximum of three minutes, which is the standard length of a regular song, and will create the intro, main track, and outro of the music.

It’s an impressive feat, but how useful is the content, and how much interest is there in this on-the-fly form of artificial intelligence (AI)-generated music that defies creativity and artist expression?

Introducing Stable Audio 2.0 – a new model capable of generating high-quality, complete tracks with a coherent musical structure in 44.1 kHz stereo for up to three minutes from a single cue.

Explore models and start creating for free: https://t.co/E9ZIGagmPf

read… pic.twitter.com/rFGb0KpdeX

— Stability Artificial Intelligence (@StabilityAI) April 3, 2024

How does Stable Audio 2.0 work?

This review from End Gadget is a critical review that casts doubt on Stability AI’s release prospects.

With a fill time of up to three minutes, Stable Audio 2.0 provides an excellent tool for experimentation and fun. Depending on the quality and final product, it can serve as a companion piece to social media content, but there is a limit to how much content can be created.

Stable Audio 2.0 works via text prompts but has an option to upload an audio clip, which must be copyright-free.

The program will evaluate the uploaded content and then generate similar content based on the prompts. Even if you can’t rely on existing music, you should adjust your expectations based on what Audio 2.0 has to offer.

It’s essentially AI-generated music, there’s no getting around that. The instrumental aspects of the music are inspiring, but the vocals are a reminder of the limitations of the software.

Stability AI’s product is a fast product capable of building a “song” in 60 seconds, with the content length extended from 90 seconds to potentially three minutes thanks to the 2.0 update.

There will be uses for this kind of audio, but it doesn’t appear to be a major breakthrough for original, AI-generated music content. Not yet anyway.

Go to stableaudio.com to try it out for free

image credit, Stable Artificial Intelligence