ABSTRACT
One of the aims of musical acoustics is to provide useful technical advice to players, composers and instrument makers. This talk reports some progress in this direction. 'The Virtual Flute' is a web service that offers alternative fingerings, and fingerings for microtones and multiphonics.
Multiphonics are a problem for composers, because not all chords are possible, and quite a few composers are reluctant to write music that cannot possibly be played. They pose a problem for flutists, too, because there is no complete listing of all possible multiphonics and fingerings. Alternative fingerings for standard notes are also useful: sometimes the standard fingerings are awkward (e.g. a trill from F6 to A6), others are difficult to play softly (e.g. most notes in the fourth octave) and many are out of tune (no flutist jokes, thank you!).
The virtual flute uses a database derived from theoretical modelling of all 39,744 different fingerings of the flute. Because each fingering can play several notes and most can play a few chords, the number of possible note/chord-fingering combinations is in the hundreds of thousands. The modelling, in turn, is based on accurate laboratory measurements of the acoustic response of the flute (senza flutist) measured at the embouchure. To get from the physical model to playable notes and chords, it uses an expert system based upon the evaluation of 957 acoustical possibilities by an expert player. Finally it uses a musician-friendly web interface to interact with musicians from all over the world who are now using it.
The talk will give a non-technical overview of how the virtual flute was built, and some demonstrations of what it can do.
* This project was Andrew Botros' final year project in the Acoustics Lab in the School of Physics, where he worked with Joe Wolfe and John Smith. He won the national Siemens Innovation Prize for this work. Andrew currently works for Cochlear Ltd.
Back to 2002 Presentation List