Techniques for the analysis of expressive gestures in musical performance

Classical music is traditionally studied from notation; but music sounds, and how it sounds depends on performance style. The project developed techniques to show what constitutes a performance style. Expressive gestures in sound that characterise personal styles of playing and singing were identified and analysed in detail, using computer visualisation techniques for sound analysis. Their deployment and function in different musical contexts were examined. The process of style change over 100 years of recorded music can be seen as resulting from the changing constituents of personal styles.

Principal investigator
Professor Daniel Leech-Wilkinson
Principal project staff
Professor Daniel Leech-Wilkinson
Start date
Monday, April 1, 2002
Completion date
Saturday, March 1, 2003
Source material
All the sound files derive from original shellac discs held in the King's Sound Archive at King's College, London: They were transferred to digital format by Daniel Leech-Wilkinson, using a range of stylus shapes and sizes, a variable speed turntable, an Elberg equalising pre-amp, a Flying Cow A/D converter, CEDAR de-clicker and de-crackler noise reduction units, and a Lynx L22 digital sound card in a Novatech PC. Conversion to mp3 was done in Steinberg Wavelab 4.0 software. Selected label scans were also provided.