What’s that song? Just ask your Mac: Music recognition specialist Shazam launched an OS X app Thursday that automatically listens to and identifies any music played in the background. It’s the first time Shazam has ever targeted desktop computers after exclusively focusing on mobile apps and services for more than a decade.
The new app installs a small icon in the OS X menu bar. Upon opening, it asks users for permission to access the microphone and listen for music, and then automatically tries to recognize each and every song played within reach of the computer. Successfully recognized songs are added to a list, giving users the option to learn more by visiting a song page on Shazam.com, or switching to iTunes to buy the title right away.
But the app doesn’t just listen to your Mac’s microphone. It also keeps tab of music played by the computer itself, which can be particularly useful when watching YouTube videos. “People shazam YouTube all the time,” explained Shazam Chief Product Officer Daniel Danker during an interview Wednesday. Oftentimes, people would watch a viral video and wonder about the song in the background, he said.
Shazam first started automatic recognition of music with its iPad app a little over a year ago, and then brought the same feature to the iPhone in December. This makes Shazam’s apps one of a number of examples for a new trend to always-on microphones that analyze audio for hot words or other key values. Danker admitted that the company has to be especially careful to address privacy issues with this kind of technology, which is why Shazam never actually records audio, but simply calculates audio fingerprints to compare against its database. The app also offers users a way to turn off listening at any time, Danker said.
I had a chance to play a bit with the app Wednesday and found that it worked decently well with music I was playing from my phone. Shazam for OS X was able to recognize most, but not all tracks I subjected it to in just a few seconds. Danker said that the database for automatic recognition is smaller than the full Shazam database that gets queried when you manually start song recognition with your phone’s Shazam app, but declined to go into specific numbers.
However, I found that using it with YouTube was a lot slower. Recognizing a Jay-Z track from the official music video took close to 50 seconds. Recognizing songs in noisy environments was also hit-or-miss. In a busy coffee shop, it recognized only one out of a dozen or so songs. Still, when it works, it’s actually a pretty nice experience, with desktop notifications alerting you to every newly-recognized songs.
However, the bigger issue I had with Shazam for OS X was that it didn’t actually keep a persistent record of my discoveries. Users can’t log into the app, and recognized songs aren’t synced with the service’s website, so you can’t really go home at the end of the day and check out all the songs that you were subjected to by your office colleagues.
A Shazam spokesperson said that this would be added soon, and Danker also told me that the company is thinking about making all those song discoveries more contextually relevant — which is when things could get really interesting: By taking the different devices you’re using as well as their locations and other sensors into account, Shazam could one day be able to tell the story of your day in sound. Think Moves, but with music instead of step counts. Danker didn’t want to tell me what the company is actually going to do on this front, but said that his team is already “thinking hard” about contextualization.