The history of music is full of technological changes that displace and innovate simultaneously. Electric amplification allowed small combos to play at a volume that previously required big bands while inventing entirely new forms of music, most notably rock & roll. Many early synthesizers were developed in the hopes of replicating the sonic qualities of traditional instruments, but more often than not ended up being used to create sounds that could never be duplicated with horns and strings. At the same time, drum machines were used to popularize rhythms that few mere mortals could play in real-time.
In each case, what was presumed to be disastrous to some opened up creative avenues enjoyed by the majority.
Today, many see generative AI music as threatening to devour all of the music made by humans. A threat that’s not not real. But while we watch labels and DSPs struggle with countering the impending AI music surge, there are music and tech entrepreneurs looking to build systems that can accomplish what no live musicians can achieve.
One of these is Infinite Album, a generative AI music tool targeted at Twitch gamers looking for music that responds in real-time depending on inputs from the game while giving players and their fans a more diverse soundscape than usual game soundtracks.
“People mute game soundtracks. And it's not that the soundtrack is bad. They're actually really good. It's just soundtrack fatigue,” says CEO/Co-Founder Karen Allen. A respected industry veteran, Allen literally wrote the book about music in the Twitch ecosystem, Twitch For Musicians. In many ways, Infinite Album is the answer to many of the challenges she discovered while doing research for the book.
“We're targeting the Twitch streamer because there is an immediate need for copyright-safe music for gamers on Twitch. And gamer creators on Twitch also need more ways to engage their viewers and monetize them.”
Obviously, AI-generated music is copyright safe by definition. And as someone who began her career working for the RIAA (at the height of the Napster era, no less) Allen and her colleagues understand the importance of ethical training data and compensation.
On the latter point, Infinite Album engages fans and followers of Twitch streamers by allowing them to manipulate the music in tandem with the game inputs. A creator can allow viewers to adjust the music in real time using Twitch’s robust micropayment system.
“The viewer has this little app on the channel page of the streamer that lets them change the genre, change the emotion, change the instrumentation and add sound effects,” Allen explains. “And the streamer can decide how many bits each costs and for how long each reaction should last.”
The money is then split automatically between the streamer, Twitch, and Infinite Album, who can divide it up further to pay royalties to rightsholders who contributed to the AI creation. IA just launched its first sound pack with French musician Thomas Barrandon, who will receive payments on music trained on his inputs. The potential is there to eventually launch a marketplace where music creators can offer their own algo to IA users.
But even if this doesn’t come to pass (tech roadmap are usually more like overgrown footpaths), Infinite Album has already identified a unique use case that could never be rendered by live or recorded music.
Yet, somewhat ironically, it’s the older instruments, the ones that have been largely displaced by previous tech, that have the most natural resistance to this innovation.
“Rock is hard. Guitar sounds are really hard to do when they're infinitely generated,” Allen admits in an example of modesty that is all too rare in startup land. “So we'd love to work with someone who can really help us with that.”
Watch our full interview with Karen Allen to learn more about how Twitch and AI Music come together:
TAKEAWAYS
Salient statements from this week’s music news.
1. Apple Might Soon Pay Artists More for Offering Their Music in Atmos
Dolby Atmos music mixes are already everywhere, but Apple seems determined to make spatial audio ubiquitous — even as Spotify still doesn’t offer it.
Takeaway: It’s possible that Apple senses an opening amid all the upheaval, and this attempt to make Atmos more enticing to artists could help the company’s standing with artists and publishers.
2. Latest IFPI Report Examines Music Consumption Habits in 2023
Global music body the IFPI has published its latest ‘Engaging With Music’ report, based on a survey of more than 43,000 people across 18 countries.
Takeaway: The average time spent listening to music each week has grown from 20.1 hours per fan in 2022 to 20.7 hours now. 32% of that listening is on audio streaming services and 31% on video streaming services.
3. Masterchannel Has Launched the “World’s First” AI Clone of an Audio Engineer
The company has emulated the expertise of two-time Grammy nominee, Wez Clarke, encapsulating his own unique flair when it comes to the engineering and mastering process.
Takeaway: Clarke licensed his expertise to Masterchannel for the collaboration, and will receive 100 percent of the extra revenue generated via use of the AI tool.