An Expressive Synthesis Model for Bowed String Instruments

The paper describes our recently developed model of physical sound generation and playing of bowed string instruments. We apply the waveguide string model and the friction bow-string interaction model for sound synthesis, parameterization of bowing styles, and a special notation for interpretation and expressive playing. Previous results are developed further in three aspects: interaction of the string with the left hand fingers, physically-based modeling of the bow motion, and the notational expression of bowing styles in addition to MIDI data. In order to faithfully reproduce both stationary and transient vibrations, we use a digital waveguide model for sound synthesis. Unlike previous models of bowed strings, we apply at the fingerboard a second scattering junction with moving position and varying impedance. This allows more realistic pitch transitions in legato playing and position changes, as well as production of flautato sounds and other dampening effects by the left hand finger. We simulate the various bowing styles (detache, martele, staccato, spiccato, saltato, etc.) by modulating the bow force and velocity with different envelopes, generated dynamically by a simple physical model of the bow, that partly takes into account the elasticity of the bow and dynamics of the player's right hand.