Research Seminars

Exploration and Mapping of Sound Synthesis Feature Vector Space

Martin Roth and Matthew Yee-King

Wednesday 11 February 2009, 14:00, Room 105

Abstract

A report on recent work concerning the exploration and mapping of sound synthesizer feature space is presented. The feature vector space for different sound synthesis algorithms was investigated in various ways. Large scale maps of this space were generated, to provide an insight into the space of all possible sounds for different sound synthesis algorithms. First, a data set is generated by rendering many sounds from a sound synthesis algorithm, each with randomised settings. These sounds are then transformed into feature vector space using feature vectors such as MFCCs and perceptually weighted power spectra. This high-dimensional data set is translated to a low-dimensional space for human exploration using Multi-Dimensional Scaling (MDS). MDS aims to make a lower-dimensional map of a larger space whilst maintaining relative distances between points in both spaces. An interactive tool was developed which allows the rapid exploration of 2D MDS maps of feature vector space by a user.
Different sound synthesis algorithms and audio feature vectors are compared using this tool. Finally, a fine grained analysis of the inter-point feature vector space is carried out in order to compare the smoothness of this space for different feature vectors.

Biography

Matthew Yee-King is an electronic music composer and computer science academic as well as a reasonable drummer. Musically, he is currently involved as a performer and producer with the new London Jazz scene and has previously released music on Aphex Twin's Rephlex Records.
Academically, he is finishing off his DPhil in Computer Science and Artificial Intelligence at Sussex University where he is researching machine intelligence applied to sound synthesis.
http://www.yeeking.net/

Martin Roth received his PhD from Cornell University in 2005 in the field of mobile wireless ad-hoc networks. He designed distributed stochastic control algorithms, inspired by the behaviour of social insects such as ants, bees, termites, or wasps, applying them to the problem of robust end-to-end routing. Thereafter he completed a two year post-doc at Deutsche Telekom Laboratories in Berlin, Germany where he developed a theoretical performance analysis of such algorithms. This time also saw an interest in the heavy-tailed statistics of human mobility models in wireless networks. In 2007, Martin joined Google Inc. in the mobile-focused London, UK office where he worked on cutting-edge mobile web applications (e.g., new versions of Google Mail and Google Talk), in addition to analysing application usage and behaviour on the first consumer Android mobile phone, the T-Mobile G1. As of February 2009, Martin is a freelance researcher working in the areas of machine learning and audio processing.