SOS 013: Elevating Musicians Through Open Source with Robert Kaye
Robert Kaye is the executive director of the MetaBrainz Foundation, the legal umbrella for MusicBrainz. He got started in the late 80’s and early 90’s hacking on some MP3 projects when most of the world hadn’t heard of MP3. The metadata on MP3s was terrible, so he started creating the database known as MusicBrainz.
Robert talks about his business model for MusicBrainz. As time has progressed, more and more people have access to a laptop and cheap recording equipment. This constant churn of data gave them the ability to play gatekeeper. Their goal was to take that data and make it cleaner, better, and provide context. In 2003 they started a service called Live Data Feed, which allows anyone to set up a copy of MusicBrainz. Turning on Live Data Feed gets you updates to your copy of MusicBrainz. The idea was to take the recognition they had around Live Data Feed and created monetary value from a service around timely and convenient packs of data.
In 2015, MusicBrainz realized that the actual value they had wasn’t in the data, but in the community of people editing the data took. So, they took a radical step and quit caring about code licenses. Now, it is based off memberships with monthly fee. This has worked spectacularly. They have taken to calling their customers ‘supporters’, because if the database is going to stick around then they need their support.
BookBrainz is a similar project to the MusicBrainz database, but applied to books. The project has grown large enough that Robert had to hire a full time engineer to work on it. They deal with disambiguation, deduplication, and conflicts in the metadata so that organizations like internet archives and Open Library can build other tools on top.
For the past 4 years MetaBrainz has also been working on two other projects. AcousticBrainz is machine learning analysis applied to individual songs to determine what music sounds like. It can determine acoustic characteristics such as male or female vocals, presence of certain instruments, and beats per minute. ListenBrainz tracks your listening history, similar to LastFM. In fact, you can import your LastFM history into ListenBrainz and it will do a metadata report on what you’ve listened to. Robert notes that if you choose to learn ListenBrainz your data will be public. These two projects form the perfect basis for building a collaborative filtering algorithm and give you personalized suggestions of what you may also like. They also have a program to work with AcousticBrainz to track what you listen to and the similarities between the songs. They are currently working on compiling the data, but this open source project will enable anyone to come in and create an open source music recommendation engine. When building a recommendation engine, the idea is if there’s a small/medium music label with one computer geek on staff, they can get access to MusicBrainz and download their recommendation engine and start getting their stuff out there, and have it personalized to the listener.
Robert’s inspiration for these databases came from seeing a lot more recommendation engines that are entirely biased and want to push their content. He realized that these recommendation engines were designed to feed money back into the system and keep everyone inside the ‘walled garden’ of music. He got funding for these projects through his good relationships with other companies and because they were giving him the money for MusicBrainz, which is enough money, so the extra money is funneled towards other projects.
The MetaBrainz Foundation emphasizes quality of life for their employees, and Robert and the panelists discuss how he reconciles this quality of life versus the desire to get all this stuff out the door. Robert believes that if you trust your team and empower them to do what needs done, they will do their job. He only really gets involved if it’s legal concerns, monetary issues, or the rare high priority assignment. His company has few deadlines, and he talks about how they organize their work. The panel compares their experience working for other open source companies. They discuss some of the drawbacks of remote work, such as difficulty coordinating meetings and never really being disconnected from work.
The show concludes with Robert talking about where he wants to take MusicBrainz and MetaBrainz. His dream is to create more tools for an improved music listening experience. His hidden agenda is to get the small bands heard so that musicians can make more money, elevating the artists in the world to be able to earn a normal living. He hopes that by applying the concepts of open source to the music industry, it will be cleaned up and all musicians will get the exposure they deserve.
- Richard Littaur
With special guest: Robert Kaye
What is MusicBrainz?
An open source database for the metadata on songs, albums, and artists. Other branches include BookBrainz, AcousticBrainz, and ListenBrainz. The goal is to use these to create a truly personalized recommendation algorithm for music and books..
What is MusicBrainz' business model?
MusicBrainz is supported by a subscriptions and a monthly fee. Their data allows users to create their own music recommendation algorithms.
How do you reconcile the quality of life that working remote provides for employees versus the desire to get things accomplished?
Robert believes that if you trust your employees and empower them to do what needs done they will do their job. He also believes in minimal interference from upper management and few deadlines.
What is the ultimate goal of MusicBrainz?
The goal for MusicBrainz is to create more tools for an improved listening experience. These tools will get the small bands heard so that musicians can be elevated and earn a decent living.