Can AI Reinvent Song Creation Without Cutting Out Humans?

As the use of AI in the music industry continues to spark debate, one of the hottest sound sample libraries in the business is launching a machine learning tool that opens new avenues of creativity for subscribers while keeping human-created loops and beats at the forefront.

Splice, which has a market valuation of $500 million thanks in part to a $55 million investment from Goldman Sachs and investment firm MUSIC in 2021, is embracing AI both to redefine the creative process for professional musicians and ease song composition for a growing group of hobbyist creators.

The company is recontextualizing its library of 2.5 million samples to focus on creative flow rather than disparate sample searching. By using machine learning to comb its library and deliver “Stacks” of up to eight layers of sounds based on a selected genre, Splice AI can deliver unique musical flows to its more than 4 million subscribers who range from hitmakers to dabblers. The process goes beyond metadata such as keys or chords, and continually adapts and matches sounds in real time.

Users can continually refine each Stack by adding or taking away samples. As with all Splice services, each sample is an original loop created by a music producer or sound designer who is offering their work via the platform for purchase on a royalty-free agreement.

“It has to be about the human at the center,” says Splice CEO Kakul Srivastava. “This is a ‘Yes, and…’ It’s not like AI versus creative, but AI plus creative—and then we can do really amazing things. There are no dead-ends.”

In this regard, Srivastava views AI as the ultimate boon for music creators—a gateway that enables them to bust out of writer’s block, stretch their musical horizons and collaborate with their peers.

“Our core customers are professionals, the people who are using the digital audio workstations and all of these complicated pieces of software, and it’s really important for them to feel inspired and not stuck. They can bring it right into their professional workflow,” she says.

The new AI is the next iteration of two existing Splice offerings: Similar Sounds (SiSo), launched in 2019, and Complementary Sounds (CoSo), which debuted in 2021.

“When I joined Splice, we asked ourselves as creators, ‘What are the points of friction in digital music making today?’” says Ale Koretzky, Splice’s head of AI and audio science innovation.

“It seemed that even with a vast catalog at your fingertips, finding the ‘right’ sound wasn’t always simple. That led us to the first user-facing AI-powered feature, Similar Sounds. The second point of friction is figuring out how and where to use a sound. That derived into an even more fundamental question: What makes two sounds work well together? So we redefined our strategy as discovery-driven creation, which paved the way for CoSo.”

Noting there are about 60 million such music professionals who regularly use digital audio workstations, Srivastava—who previously held executive positions at Adobe
ADBE
, Flickr, Yahoo! and WeWork—has an even loftier goal in mind: Easing music creation for what she says is a market of nearly 400 million “future producers.”

“There’s another, bigger market of people who want to create music. They’ve tried it, they’ve opened a doc but it’s so overwhelming to look at all those buttons and keys. A lot of that software was built 30 years ago and hasn’t changed much. So we’re also thinking about what kind of creative experience those people need.”

Having worked in the creator space for decades, Srivastava believes “music creation has been chronically underserved.” While she notes the process is in part hamstrung by having to navigate licensing, she believes there’s a larger inhibitor at work.

“I think there’s this mindset that when you come into the music industry there’s so much focus on the fame and fortune, and people forget that there’s another world here as well. It’s a tale of two cities. There’s one city that’s very shiny, with all this fame and fortune. And underneath it is a less glittering city, but one that is just as important—and that’s the city of the creative person,” she says.

“I don’t know why we think that if someone wants to create music they must want to become famous. I think that expectation worsens the mental health issues in thus industry because you come into it and all of a sudden those expectations are there. The pressure is immense. It’s OK to create just for creation itself.”

With the launch of the new Splice AI, Srivastava says the company will continue to observe patterns of use. “Our core business is around our sounds catalog and our sounds subscription service, which is still really important and relevant to the core market,” she notes.

“But I do I think there is more demand for music creation than people expected.”

Source: https://www.forbes.com/sites/cathyolson/2023/06/27/can-ai-reinvent-song-creation-without-cutting-out-humans-music-sample-platform-splice-says-yes/