Music Box – Voice Recognition

April 2011.
A voice driven music-player app.
My role: Sole programmer. 

One of my favorite projects during my time at Hungama was the development of a voice triggered music player app.

I was motivated by that geeky satisfaction of peeking under the hood of a fascination piece of technology accompanied with the immensely satisfying prospect of speaking to my phone and having it respond intelligently (this was pre-Siri); not to mention the fascination a developer experiences when watching a real time Fourier Transform dynamically change to the changing inflection of voice.

Since this was an experimental prototype for the company I set out looking for multiple ways to accomplish voice recognition, and in my travels across the interwebs I stumbled upon the Open Ears voice recognition API for the iPhone by the swell folks over at Carnegie Mellon University. Open Ears uses the open source CMU PocketsphinxCMU Flite and CMUCLMTK libraries.

Using the Open Ears API is fairly simple and I had a working prototype up in a few hours. Here are a few screenshots from the app courtesy the uber-awesome Ram Wadaye.