Last month, Chordify’s own Bas de Haas and Vincent Koops flew down to New York City to visit the annual Conference of the International Society for Music Information Retrieval (ISMIR), the world’s leading research forum on processing, searching, organizing and accessing music-related data. As you know, Chordify is all about music science, so the annual conference was right up our alley.
Co-founder Bas is a real veteran when it comes to ISMIR, having been involved with music information retrieval ever since his PhD research. In fact, our algorithm that automatically extracts the chords of your favorite songs is based on his academic handiwork. The conference features all the latest research in this area and some really big companies – like Spotify, Google, Apple and Pandora – are starting to pick up on that too.
What’s so cool about MIR
Music Information Retrieval may sound a little dry and technical, but the applications are very awesome; you’ve used it many times without probably even knowing it. We’re not just talking about our own algorithm, but things like the recommended songs you get with Pandora and Spotify or using Shazam to get the title of the song you’re hearing, it’s all based on this field of research. “The conference is a great way to keep up with all the technology and research”, Bas explains. “Everything we can currently do, all the latest research, is presented right here.”
“My personal highlight was the talk on the Deep Chroma Extractor”, Bas continues. “Without getting too technical, it uses artificial intelligence and neural networking to recognize chords in music. These neural networks are basically a mathematical model, inspired by the way our own brain works.”
“You can do a lot of really cool stuff with it, like creating an algorithm that recognizes dogs and cats in photographs and YouTube videos, or beat people at that very complex game called Go, like Google’s algorithm recently did. And there’s some really cool art made by A.I. too, like Google’s DeepDream.”
Science and tech companies working together
Speaking of Artificial Intelligence, that’s also what Vincent is involved in. He’s currently working on his PhD thesis at Utrecht University with Chordify, coming up with ways to combine your chord edits in more accurate schematics. “I was a little worried about making it to the conference though”, Vincent explains. “I broke my leg recently and had a couple of metal pins inserted and of course the US is notoriously difficult to get into.”
Fortunately though, the metal pins didn’t set off any alarms at airport customs. “I really enjoyed the conference. I went to a couple of tutorials and workshops, like jazz solo analysis. And a colleague of mine gave a talk about why hip hop is very interesting for scientific research, since most of the genre is sample based, so there’s some really cool applications for that, like recognition software for the original song the sample was used from”, Vincent continues.
“I also really enjoyed the social program, like going to the Brooklyn Bowl, enjoying an IPA while watching the band Soulive perform.”
“What’s also really interesting,” Vincent continues, “is that a lot of big tech companies take notice of what’s happening here. They’re not just here to check out the latest research, but are also actively involved in the official program. Spotify hosted their own Hackday just before the conference and Pandora got the band Soulive to perform at the Brooklyn Bowl.”
The science behind writing down chords
Can artificial intelligence disrupt the music industry?