Ultrasonic networking: your devices are talking, and you can’t hear a thing

This is cool: Google(s goog) engineer Boris Smus figured out a way to use ultrasonic sounds to let your devices talk to each other —  and he’s using the Web Audio API implemented in Google’s Chrome browser, so you don’t even need to download an app for it.

Smus wrote on his website that the idea behind ultrasonic networking was to figure out alternatives to Bluetooth, WiFi Direct and NFC to let mobile and desktop devices that are in the same physical space pair up and exchange data with each other. One key motivation was that existing technologies often aren’t compatible with each other:

“Because there are so many possible technical approaches, chances are that the pair of devices that you happen to be using don’t have a common language to speak. Even if both devices have Bluetooth, one of them may require a profile the other doesn’t support, or support a different version of the standard. ”

However, pretty much all devices have microphones and speakers, so it’s easy for them to exchange sounds. That’s why Smus built a JavaScript library to do just that. He also mocked up two demo web apps to show the potential of his work: One app simply sends emoticons from one device to another, while the other pairs both devices, and then lets users chat through a cloud-based service

Check out the demo video below:


Smus’ two demos both require an updated version of Chrome on the desktop as well as Chrome beta on your mobile device. And because mobile phones aren’t exactly optimized for ultrasonic sounds, receiving messages on your mobile may only work if you switch to a more audible mode.

And then there’s another limitation: Translating data into sounds also limit your bandwidth. Asked about this via email, Smus replied:

“There’s an inherent tradeoff between bandwidth and reliability of transfer. The longer you spend transmitting each bit of data, the more likely it is to be transmitted without error. (…) The main point here is to show that it can be done on commodity hardware, and that it’s possible to do with the open web. We can do much better, though! I’d expect we can get to something on the order of hundreds of bits per second reliably with more sophisticated DSP approaches.”

However, in the end, you don’t actually need all that much bandwidth for what Smus has in mind. After two devices are paired, everything can be offloaded to the cloud. That approach would also work really well for the internet of things, believes Smus:

“It is a pure web technology that can be used to pair devices together. Because of the ubiquity of web browsers and audio hardware, the combination can be a huge win, even among commodity hardware, without having to wait for Bluetooth and other close-range connectivity technology to become available to the web platform.”

Image courtesy of Shutterstock user Tund.