My Music
The prompt for this week's Disquiet Junto project was to explore what we think the room in the image of the cover art sounds like. I normally do not compose this kind of track, but I found myself immersed once I started working on it. I layered sound effects I found on Pixabay along with a heavily muffled version of one of my own tracks to create my submission. If you would like to take a listen to what I came up with, you can do so here.
Music and Artificial Intelligence (AI)
Spotify announced an AI playlist feature. An AI system that creates a playlist based on a prompt (which also seems to require mentioning a genre, or some word closely tied to one) seems much less interesting than the “Daylist” feature that actually reflects a user’s listening habits throughout the day, or the AI DJ that presumably transitions smoothly between two tracks with complimentary features.
This announced partnership between Warner Music Group and Medimusic will be an impressive application of AI technology for playlist generation. I can imagine wanting to relax by listening to a certain song, but if the sensor in my wearable indicates I am not relaxed, I would want help finding a new song to give me the desired effect. I also wonder if this will create a market for medical musicians - those who create tracks specifically to treat certain illnesses.
Another announcement about an AI system that can create an entire song from a prompt. I participate in prompt-based music projects, namely the Naviar Haiku Music Challenge and the Disquiet Junto, where I create a song in response to a Haiku or a short, one or two-sentence instruction, respectively. Never would anyone providing those prompts claim to be the “creator” of my song. Why do the purveyors of these systems say you can “create” a song based on a text prompt?
Rotor Videos has been acquired. I used Rotor after receiving 3 credits with my SoundCloud subscription. I thought that would equate to 3 videos, but it turned out each video costs 2 credits, leaving me with 1 credit after creating a video. I need to buy additional credits if I want another. How clever. The one I created is just a simple animation incorporating the album artwork. I’m still looking for a #AI platform that allows easy video creation since it is a medium fans discover #music through.
I wondered who is paying $30K for one track, but the fine print in the linked article clarifies this often happens through a “group buy.” I can see someone throwing a small amount of money at a bet that a snippet that sounds authentic really represents a pre-release track. Also interesting is that the majority of takedown notices are not #copyright claims since no one would own the copyright in #AI-generated output. Rather the claims are to enforce trademark rights or the right to publicity.
This new AI music venture involving Deadmau5 sounds like interesting technology, and I like the idea of musicians being able to monetize their contributions, even if those contributions are just audio recording stems. But when I read something like, “The company is still figuring out its business model…” I lose interest and realize there is still a long way to go until a service like this is a reality. Until there is a business strategy, this is just playing around with other people’s money - and a lot of it.
Lost in the details in this story about chatbots used to tackle the problem of loneliness with seniors is that the #AI chatbot plays #music for the user. I read this as it “plays music” in the same way it tells a joke, for the purpose of making the end user happier, similar to how a person might notice a friend is feeling bad and then might tell a joke or put on a song to try to improve that friend’s mood. It would be interesting if an AI system could predict what I want to hear and also determine a personal and friendly way to present it to me.
This is a very touching video about disabled people using AI-enabled technology to perform music. There are so many examples of how #AI technology will enable non-musicians to make #music. Many of them are uninspired attempts to commodify the music-making process. I see this technology as something else entirely, where musicians who have disabilities will be able to faithfully recreate performances that they otherwise have been prevented from playing. It could also give musicians without disabilities new avenues to creatively explore.
To every music artist signed to a record label, I would ask: Does your contract contain the language quoted by the record label in this article about posthumous DMX songs, including “comprehensive rights to [your] distinctive vocals [or other attributes], allowing [your record label] to explore a variety of creative avenues, including remixes, and alternative versions”? If so, do you believe those rights cover the use of your works with #AI technology after your death? Maybe so, but I’d still like to see how this is laid out in any contract drafted before 2022.
Haiku
Cold gray cloudy sky
Frigid wind collects itself
Avoiding conflict
Mossy overgrowth
Garland on the banister
Nowhere to hold on
Early Sunset
Frost glistens on the grass
Continuing grief
Snow capped tree
Exalts the blue sky
Memories return
Purebred white whippet
Black whippet from a rescue
I will miss my dogs
Snow drift undulates
Shadows dance on the surface
Sapling emerges
Pressed against the wind
Yellow wild flower faces
She loved that old barn
I prop myself up
With every stick I can find
Unshakable love
Comments