Devin Arne

Hi everyone, taking a break from X for a bit. I'll be working on a new research project this winter and will most likely return then. Be kind, be safe.

😅😅

If you are a musician, music industry professional or a music lover take a moment to read to the motions filed against Suno and Udio. See links in @ednewtonrex ‘s analysis below

@dylanowenmusic @TheTrueAbstract Amazing! Have a great show, we’re up in MA this week otherwise I’d make the drive out from Philly to come see the show!

They weren’t missing a letter on this sign for last night’s show, they know we don’t take L’s!!! Pittsburgh PA tonight for the final stop of tour

@deeplearnmusic Congratulations! This is an amazing tool, the one stem at a time approach is great for musicians/producers as it gives granular control. I could also see this as helpful for film/TV composers to make variations on themes and cues. Hope to be able to use this in the near future!

🥳Excited to share our latest work: "Diff-A-Riff"! 🥁 A Latent Diffusion Model that generates instrumental accompaniments for any musical input, specifically tailored for music producers! It's faster, lighter, and produces superior audio quality. Control via text/audio references. 48kHz sample rate, (pseudo) stereo, ~3Gb memory, takes 6 seconds to generate 90 seconds of music. Trained on a single GPU. 📜arxiv.org/pdf/2406.08384 🎶sonycslparis.github.io/diffariff-comp… 🎸 "Diff-A-Riff" adapts to any musical input, following the artist's unique style. 🎛️ Optional controls via text prompts, audio references, interpolation slider, pseudo-stereo width and loop intensity. 🎚️ It produces state-of-the-art audio quality indistinguishable from real data by human raters and operates at unprecedented speed. 🧠 "Diff-A-Riff" is smaller and more efficient than previous models thanks to its Consistency Autoencoder, making it accessible and practical for various applications. Big shoutout to my outer space colleagues: Javier Nistal, the Machine in "machine learning" 🚄, Marco Pasini, the neural net whisperer 🤫, Cyran Aouameur, the troubleshootah 🛠️, Maarten Grachten, aka MaartenGPT 🤖. #Teamwork #AI #MusicTech #Innovation @latentspaces @marco_ppasini @cyranaouameur @SonyCSLMusic

@theFinneganegan So cool! Congrats on the release.

The wait is over, Droplets is out and you can buy it now!!! finneganeganegan.gumroad.com/l/Droplets Big thanks to everyone who's shown support throughout this long journey <3 #musicproduction #generativeart #audiovisual #vst #vstplugin

@JLenzyy @navreyort @zlbdad @JLenzyy , looking forward to meeting you as well!

@devinarne @navreyort @zlbdad Will be great to see you there and finally meet!

"How can AI enable the musicians of tomorrow?” I'm thrilled to dive into this topic with @navreyort @zlbdad and Carlos Arana at the "Educating the Next Generation of AI-Enabled Musicians" panel at next week’s AES Symposium on AI and The Musician aimusicpreneur.com/ai-music-news/…

Best Practice

The rollout of AI summaries instead of search results threatens to undermine any collective sense of shared history, thereby unraveling the threads that bind us, & abandoning truth as a guide. On both a literal & philosophical level, what happens when we stop searching?

On World IP Day, a reminder that the internationally agreed Berne Convention forbids copying that “unreasonably prejudice[s] the legitimate interests of the author.” Fair use, and other copyright exceptions, have to meet this standard. Copying people’s work to build a generative AI model that competes with them clearly unreasonably prejudices their interests. It reduces their opportunities and decreases their income. This is already apparent. Most of the gen AI world may have adopted this approach. But that doesn’t make it legal, and it doesn’t mean we should normalize it.

If you want to democratize something, develop tools to make it less expensive and easier to learn. Typing text prompts and receiving ready-made audio further mystifies and obscures the beautiful, joy-filled process of creating music.

AI startups talking about how they will 'democratize' music making (or art in general) is just as cringe as Facebook/Meta's "move fast and break stuff" mantra of their early years. rollingstone.com/music/music-fe…

The first session of my new lecture and discussion series starts today! Very excited to talk about and learn more about our student’s perspectives on generative AI and Music.

new turing test just dropped

After an unreleased performance of the Tomasz Stanko Quintet was found, @AstigmaticRec turned to AudioShake to create stems of the 50-year-old recording. More on our work helping to master this radio broadcast turned vinyl album, “Wooden Music II”: audioshake.ai/post/astigmati…

This blew me away today: a singing cow gate! Geophony (strong wind) meets antrophony (cow gate), with a sprinkle of biophony (passing jackdaws and a late bumblebee!). I was taken aback at how musical this sounded as the wind flowed through the nooks & crannies. Just beautiful

@ethanhein Love this question. I think a rapped verse falls into the same category as a jazz solo. One can transcribe it to learn from it, but it would be gauche to perform or record it as is and pass it off as a genuine artistic statement.

@Komaniecki_R As a grad of McGill, this is sad to see. I would have, and sure many others would too, gladly accepted mandatory French language classes. Punishing out of province and international students is not the move.

To protect and promote the French language, Québec is proposing tuition hikes that will *double* costs for out-of-province students. This would have devastating consequences to McGill’s fantastic school of music in Montreal. montrealgazette.com/news/local-new…

Read this, understand this. Embrace your limitations, your imperfections your humanity. This is it folks, the only way forward.