Transient Phenomena in Single-Reed Woodwind Instruments

The function of musical instruments has always been under study by physicists, engineers, and scientists in general, not only due to their apparent role in music production, but also because of the compelling physical effects that govern the process of sound generation. To this cause, music acoustics has developed as a science field that aims to study the underlying physics and to carry on asking further research questions, towards the understanding of the observed phenomena. One issue that has seen increased interest in the past few years is the analysis of transient phenomena in wind instruments. The rapid variations that take place during tone transitions pose several problems to experimental and theoretical investigations alike. Therefore, even though the behavior of wind instruments during steady state sounds has been already well comprehended, the same can not be said concerning transient oscillations. The emergence of such transients during expressive performance in single-reed woodwind instruments will be the focus of the proposed research project. The use of different articulation techniques available to the player, allows various ways to control the oscillations of the instrument and give expression to the sound. Recent research by the author has used experimental measurements to observe the differences in the resulting sound pressure. During the course of this project novel measurement techniques will be developed for the analysis of transients. Furthermore, a physical, time-domain model of a single-reed woodwind instrument will be formulated, suitable for the synthesis of transient phenomena. On the basis of both the experimental data and the physics-based analysis, an inverse model will be designed that can estimate physical parameter values from sounds recorded under real playing conditions. This way it is possible to extract information regarding, not only the musical instrument, but also how the player is controlling it in order to produce the desired sound. With the use of inverse modeling it is possible to calculate the value of certain physically meaningful parameters that are impossible to track under real playing conditions. Furthermore, the model parameters can be associated to certain sound features and performance actions. It is thus possible to determine which of the underlying physical effects, that take place during transients, are significant for the generated sound. The development of an accurate and efficient physical model, which is able to capture transient phenomena, is thus invaluable for both sound synthesis and sound analysis applications.

The function of musical instruments has always been under study by physicists, engineers, and scientists in general, not only due to their apparent role in music production, but also because of the compelling physical effects that govern the process of sound generation. To this cause, music acoustics has developed as a science field that aims to study the underlying physics and to carry on asking further research questions, towards the understanding of the observed phenomena. One issue that has seen increased interest in the past years is the analysis of transient phenomena in wind instruments. The rapid variations that take place during tone transitions pose several problems to experimental and theoretical investigations alike. Therefore, even though the behaviour of wind instruments during steady state sounds has been already well comprehended, the same can not be said concerning transient oscillations. The emergence of such transients during expressive performance was the focus of this research project. The use of different articulation techniques available to the player, allows various ways to control the oscillations of the instrument and give expression to the sound. Focusing on single-reed woodwind instruments (such as the clarinet and the saxophone), three different research axes have been followed. (I) Measurements with real performers. (II) Measurements using an artificial playing device. (III) Modelling of the physical phenomena that take place using computer simulations. During the course of the project, an artificial blowing machine for single-reed woodwinds has been constructed, able to imitate sound generation by human players. A non-intrusive experimental setup was also developed in order to carry out measurements with real players, in order to analyse their performance and to obtain reference signals. These signals were used to validate both the artificial player, as well as the physical model that was formulated. With the use of inverse modelling it is possible to calculate the value of certain physically meaningful parameters that are impossible to track under real playing conditions. Furthermore, the model parameters can be associated to certain sound features and performance actions. It is thus possible to determine which of the underlying physical effects, that take place during transients, are significant for the generated sound. Combining the information gathered by the above approaches, it has been possible to improve the accuracy and efficiency of physics-based synthesised sounds and to shed some light on hidden players' actions during performance (mainly regarding the use of their tongue, as well as the way they adapt their vocal tract in order to control sound generation). These findings have direct implications to the sound generation in virtual environments and may be also used by performers and music teachers during practice, by allowing instrumentalists to receive visual feedback regarding their actions.