Mr Chris Cannam

I work with researchers to produce finished software that people want to use.

My job is to make, package, publish, and maintain end-user applications based on research work, and produce new tools to help researchers carry out that work.

I help maintain the SoundSoftware services and a number of other servers for the Centre for Digital Music, providing infrastructure for audio and music researchers who need to develop software. I also provide and coordinate development effort for specific research projects, offer development support for new research grants, and am occasionally involved in industrial partnerships.

Projects

• Sonic Visualiser, a widely used application for viewing and analysing the contents of music audio files.
• Sonic Lineup, a tool for quick comparative visualisation and navigation of multiple related audio recordings.
• Piper, a project to explore audio analysis using browser technology, Javascript, and network protocols. (code project)
• Tony, a tool for machine-assisted annotation of pitch and note information from solo singer or monophonic instrumental recordings. With Matthias Mauch and George Fazekas. (code project)
• A Python host for Vamp audio feature extraction plugins. (code project)
• A Vamp plugin implementation of a shift-invariant latent variable model for multiple fundamental frequency estimation and polyphonic note tracking, from Emmanouil Benetos and Simon Dixon. (code project)
• A C++ library and Vamp plugin for quick online constant-Q transforms. Based on work from Christian Schörkhuber and Anssi Klapuri. (code project)
• Mechanisms for embedding machine-learning models trained using typical research workflows into native-code end-user applications.
• The Vamp plugin SDK and API.
• Sonic Annotator, a batch tool for feature extraction and annotation of audio files using Vamp plugins.
• The QM Vamp plugin set.
  
• EasyMercurial, a simple user interface for the Mercurial version control system.

Things I'd like to do
But won't necessarily get the time for

• "Mini Edition" of Sonic Visualiser, with instant access to common visualisations and touch-friendly interface.
• Library and plugin of analytical features for recordings of singing (vibrato rate and extent, vowel classification, etc). Ideally just collecting some existing methods.
• Media player that does a sensible job organising recordings from the traditional classical canon. I made this dataset (related code project) while thinking about this once, but it didn't go far. I'd like to revisit it.