Clio Searches for Music—Using Music

April 11, 2012

In 2007, Clio founder and principal scientist Dr. Greg Wilder saw a fundamental problem with music search engines: they all required words. So he set out to build an engine that could not only find music based on other music, but actually compare tracks to one another to find something similar—a music-based analysis, search and discovery platform.

“It’s similar to what Shazam does, but on steroids,” says Monte Zweben, the company’s executive chairman. “[Shazam] makes a fingerprinting algorithm to match music. What we do is make a musical model that understands music.”

Back in 2007, Shazam gave consumers (well, at least smartphone users) a way to identify live music in real time. With their app, users sitting in a coffee shop could hear a new song over the loudspeaker, hold up their phones, hit a button, and bam, know the name of the song they want to add to their iPods. Easy.

But five years later, Clio is offering something that goes a step farther—a music search engine good enough to help professional media producers.

Wilder wrote a piece of software that can break down a “musical sentence” into its grammatical parts–rhythm, harmony and melody. He analyzed those parts and constructed a set of numerical values for them, building a code. Then, he created a search engine that could find a musical match based on both attributes.

“There are a variety of tech people building apps to recommend music, but all of them are based on people trying to describe music and categories with words, tags and search engines for these tags,” Zweben says. “We see ourselves providing something much more specific and powerful from a tech standpoint.”

In terms of organization, and the ability to find similar songs, Clio comes close to what Pandora‘s Internet radio does when it gives users a station based on a particular band, say The Rolling Stones or Led Zeppelin. But Pandora relies on a team of musicologists to build those connections, which then get coded into the software “Most companies can’t afford to do that,” Zweben says, but Clio’s back end does it automatically. “It’s a very cost-efficient way to make your own searchable catalogue.”

Clio also has a different business model from music search engines that cater to consumers. Instead, Clio’s clients are professionals, from major labels to music aggregators to anyone who tries to license music and make it searchable.

“There’s a gazillion new videos being produced, a lot of films, and unbelievable amounts of visual imagery exploding because of the Web,” Zweben says. “Every mobile video, game, advertisements traditionally requires music.”

And it’s expensive. Directors often will stick a track onto a piece of video, Zweben says, and then music directors are expected to license it, or find something similar that fits into a budget. Finding anything through a traditional word-based search would be tough, but finding a similar piece of instrumental music can be particularly hard.

Though the company isn’t setting out to be a consumer-focused recommendation engine, like iTunes’s Ping, “that is an opportunity for Clio,” Zweben says. “We do have people knocking down our door for rec engines. We have customers building them with our technology.”

Zweben himself is new to the company; he took on the executive chairman role last month, coming most recently from SeeSaw Networks, a digital media company he co-founded. Before his entrepreneurial career, Zweben served as deputy branch chief of the NASA Ames Artificial Intelligence branch. But, as a four-time entrepreneur and a long-term guitar player, the position seemed like a good fit.

Wilder developed the analysis and search functions in 2007, but the 10-person company is “in the earliest stage you can be as a company generating revenue,” Zweben says. “We’re in that really exciting phase where we’ve released our product, customers are paying and we’re out there growing the company, seeking the financing.” Clio is starting its series A now; the company was initially funded by angel investors in its early seed round.

For now, Clio’s main product is licensing the technology as a service to other businesses. The company plans to release other products soon, but sees its search engine as its strongest point. There are a lot of other platforms that offer music search, but Zweben isn’t aware of any competitors that can offer the same kind of technology.

“I am very excited with our progress. Stay tuned for more product announcements,” he says.

Original Source: Xconomy