I've just handed in my final dissertation, entitled Timbral classification. I built a system that can recognise an instrument from a sound recording by quantifying aspects of timbre, the texture of a sound. You can read it here.
I want the computer to become a partner in musical improvisation, rather than just a tool. The first step in this process is allowing the computer to react to sounds in a context-aware manner, so if the computer can recognise the currently playing instrument it can transform its sound in a sensitive way.
Over the summer of 2010, I worked at Google in London as a Software Engineering Intern. It was fab.
For our second-year group project, my team and I put together a user-programmable guitar effects processor for ARM's mBed platform. I was responsible (with Scott Whittaker) for the whole effects processing pipeline and electronics. This project was mega, as I'd never done any serious electronics or embedded programming, and this was my first major piece of C++ work. Still, we managed to pull through and crank out some serious digi-tones with the mBed's 10-bit DAC, giving an overview presentation at the lab's project expo.
Over the summer of 2009, I worked on audio networking applications for Android phones, with the goal of representing arbitrary data as sound. Here's a proof-of-concept audio visualisation tool. You can get freaky with Spectral, which explores a mapping between images and sound. I also released a simple WAV file recorder to the marketplace, Hertz, which has had thousands of downloads already. I consolidated my voyages into a Java library for general audio networking.
In what spare time my crusades allow, I write music. I was Entertainment Officer for Churchill College from 2009-2010, Vice President of the Visual Arts Society from 2010-2011, and I've also been known to take photographs, make films and write articles.
Last updated: 14th May 2011, by Rhodri Karim.