Audio QA & Prompt Engineer @ https://nitter.cf/t.co/tU4QTC4QUn
West Midlands, England
Joined May 2017
- Tweets93
- Following91
- Followers160
- Likes153
Hot take: if you're making any form of electronic dance music, and the piano roll in your DAW is constantly looking like a piano recital at all points in the song, you probably sound Disney.
I'm noticing Image models are getting to the same state of diminishing returns in fidelity as video game graphics, where sure things are improving, but it's getting increasingly harder to notice. The more important innovations will come with controllability and interface.
I made the track in the promo while doing QA for the model. If it isn't obvious from the promo vid the instrumental utilized only sounds coming from the SAO Small model.
I've uploaded the full track to SoundCloud if you want to check it out.
Soundcloud
soundcloud.com/atoddswiththe…
Today we’re open-sourcing Stable Audio Open Small, a 341M-parameter text-to-audio model optimized to run entirely on @Arm CPUs.
This means 99% of smartphones can now generate music-production samples in seconds, right on-device with no internet required.
Built for fast, on-the-go creation, it turns your next quick idea into up to 11 seconds of audio. Generate drum loops, foley, riffs, and textures right where you are. No cords 🔌 just chords 🎹
You can learn more here: bit.ly/3GUomcC
@ODDs retweeted
I wonder what @PurzBeats and Techno Viking were talking about! :D
@LumaLabsAI Image to vid with Ray2 is too fun! :D
@ODDs retweeted
Had a great time on @AIForHumansShow Episode 52 chatting with @gavinpurcell and @Attack
Full Episode: youtube.com/watch?v=vPQemOTh…
Hey, ODDs here prompt engineer and Model QA for Stable Audio 2.0.
Very excited that you're all able to get your hands on the new 3 min model so I've put together a prompt tutorial where I discuss some of the more philosophical parallels between prompting and the more involved artistic process.
I hope that you can get some good tips on how to approach audio from someone who has spent more time with the model than anyone else and hopefully inspire you to start seeing the potential of AI in music in a much broader context
@stableaudio
One of my favorite things about using generative AI in music workflows is that my problem with using sample packs in a limited transformative manner makes me feel guilty or paranoid I will hear the same sound down that line. Prompting for my samples removes this paranoia.
soundcloud.com/at-odds-69/ru……
I hate how I can't edit tweets to change a link...
Thought I'd share one of the tracks I've been working on. Plenty of stable audio-based sound design in this one. Lovely Jubbly. #stableaudio
