Jump to content
IGNORED

Riffusion - Music Made From AI Stable Diffusion

Rate this topic


Recommended Posts

https://www.riffusion.com/about

  Quote

You've heard of Stable Diffusion, the open-source AI model that generates images from text?

photograph of an astronaut riding a horse

image.png
Well, we fine-tuned the model to generate images of spectrograms, like this:

funk bassline with a jazzy saxophone solo

image.png
The magic is that this spectrogram can then be converted to an audio clip:

????

Really? Yup.

This is the v1.5 stable diffusion model with no modifications, just fine-tuned on images of spectrograms paired with text. Audio processing happens downstream of the model.

It can generate infinite variations of a prompt by varying the seed. All the same web UIs and techniques like img2img, inpainting, negative prompts, and interpolation work out of the box.

Expand  

 

WATMM-Records-Signature-Banner-500x80.jpg

 

Follow WATMM on Twitter: @WATMMOfficial

dootdoot.gif.ddabfa450296178304c8972fae97f00d.gif

 

 

 

Edited by luke viia

GHOST: have you killed Claudius yet
HAMLET: no
GHOST: why
HAMLET: fuck you is why
im going to the cemetery to touch skulls

[planet of dinosaurs - the album [bc] [archive]]

The samples on that 'about' page are amazing, see the 'Looping and Interpolation' section, where it interpolates between 'typing' and 'jazz'

https://www.riffusion.com/about/typing_to_jazz.mp3

Another interpolation between 'church bells' and 'electronic beats'

https://www.riffusion.com/about/church_bells_to_electronic_beats.mp3

latent_space_interpolation.1adf6eb9.png&w=3840&q=75

Edited by zazen

                                                     

 This is freaking RIDICULOUS. Some highlights of Experimental electronic noir waltz, into avantgarde atonal string quartet, into rhythmic bucket clattering. The thing just does an effortless DJ set for you. Very inspiring! (my post formatting is super messy) 

Edited by chim
  On 12/17/2022 at 10:55 PM, logakght said:

Can someone try black metal IDM?

I tried all evening, with various followup prompts to try and nudge it, but all I got was BOCCY variations on some grunge guitar riffs and rock drums. The dataset is fairly limited to a low-mid tempo, and not a lot of material from black metal or blast beasts. I'm pretty sure they put various constraints to keep the tempo and keys similar across variations. 

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×