I think it's an really interesting prospect for DJs especially to have more flexibility with the material they are playing... Here a subtle reworking of one of my favourite tracks as realized by the model.

It's also kind of something I like about these things... I don't save the parameters for the outputs, so everything created is unique and will usually disappear after the session is gone. There's something very poetic and fragile about that.

interesting!! 😵‍💫😵‍💫🚀🚀

I would try to add more styles like Disco, drum & bass, metal, folk and jazz to make it more interesting

Some random recent experiments made with my custom AI models trained on a curated selection of around 1000 trance tracks which I really like.

Stable Audio Open was used to train these. SAO is an open source, easily expandable project which itself is trained only on copyright-free music and sounds. The personal test models these were made with were obviously trained on copyrighted stuff but in the future I hope to be able to create copyright free open source models to be shared with the community. I have also made models exclusively on my own music, and in the future I hope to be able to share them with people so they can have a small flavour of the vibe of my music available as inspiration for their own work :)

Stable Audio trance

https://aavepyora.wordpress.com/ for releases.

Here I will probably hold work-in-progress versions, remasters of old tracks and so on. So it will probably pay off to follow what goes on, not all might be uploaded to the main site.

Comments

Sphirala

Dark el Kante

Retroalien

Dixie Krystalz

Aavepyora

Aavepyora