This is part of our Inno on the Road Series, where we explore and showcase a neighboring ecosystem. Inno on the Road: Hampton Roads, which will feature a couple new stories a month and an event in May, is presented by BDO.
What do you get when you combine artificial intelligence with years of neuroscience research and four passionate musicians who just happen to also be scientists?
For nearly five years, Secret Chord Labs has been developing a way to test for the likelihood of enjoyment of popular music using an AI system informed by years of neuroscience research.
CEO Scott Miles was working on his dissertation in Neuroscience at Georgetown University and, with the help of now-COO David Rosen, published “A Statistical Analysis of the Relationship between Harmonic Surprise and Preference in Popular Music.” Rosen and Brian Owens, who is now chief communications officer, joined forces with Miles in 2018 to complete the behavioral study, resulting in the official formation of the Secret Chord Labs team.
The Virginia Beach-based tech company picked up steam in June 2018, when Rosen and Owens went full time to participate in the Google Project Music Accelerator in Nashville. During this 10-month program, the team got the chance to sit down with executives from major labels and music tech companies to discuss the real obstacles that they encountered every day and how their platform could evolve and be used to solve those problems.
All members of the four-person team identify first and foremost as musicians. If that wasn't cool enough, Rosen also more recently released a study investigating the brain activity of jazz musicians while improvising on the spot.
"There's something that you feel on the musician side when things are going really well, when you're having an amazing improvisation moment or you wrote a new part in a song that really just clicked,” Rosen said. “I think the question we are looking at is what does that moment look like, and is there a way that we can understand our own process of [creating music] by understanding that state that we achieve?”
"We are studying the most human element of music"
The insight gained from their research on the neuroscience of music enjoyment eventually evolved into Secret Chord Labs’ flagship product, dopr, an AI platform that models the brains of the target audience to test for the likelihood of enjoyment.
The dopr platform provides metrics for two major uses: A&R, which is the identification of musical talent for commercial success (such as choosing a hit single from a list of 50 recorded songs) and marketing (such as identifying target demographics or identifying songs for playlists to increase listener retention). But Owens wants to be sure that people see that they are leaning into the human element of music, rather than removing its humanity through the application of excessive science.
“We're looking at neuroscience, the thing that makes us the most human,” Owens said. “We are studying the thing that brings us all here, the thing that raises your hairs when you hear the music that you love, the thing that makes you fall in love with music. We are studying the most human element of music.”
Moving forward, the team is working to scale the dopr platform. This means expanding the genres that can be analyzed, increasing the number of song features looked at, refining the audio ingestion process and enhancing UX design and implementation. Rosen also says that this tech could eventually be applied to areas like advertising and branding, and even eventually in educational and clinical environments.
The key tech used in the platform analyzes the audio signals themselves in order to decipher their most basic features like harmony, melody or rhythm. Additionally, the software analyzes music ahead of its publication directly based on these musical features, rather than categorizing or ranking songs based on post-publication criteria like streams and social metrics.
Secret Chord Labs is in the middle of a $1.5 million seed round of fundraising and has raised $355,000 to date. The dopr platform is still in the R&D stage and is currently in beta testing, and there is a waitlist on their website for testing out the software. The Secret Chord Labs team has already been in talks with industry partners, whom they said they can’t name yet but would classify as some of the biggest names in the music publishing business, and they've been testing the viability of songs for commercial success by analyzing thousand-song catalogs.
“While we cannot guarantee that one particular person will enjoy a given song, we are able to predict preference up to about 25%,” Owens said. “Even if we’re pushing the needle 6%, given the $6 billion a year that the music industry spends on A&R and audience seeking tests, 6% is still a number that makes everybody happy.”
Added Rosen, “When we leave the world of science and just talk to people out in the world, they really care about this. There's something about the concept that is intriguing enough to make it a worthy pursuit.”
