An Efficient Modal-based Approach Towards Guzheng Sound Synthesis

TL;DR

This paper introduces a novel, efficient modal-based physical model for synthesizing authentic guzheng sounds, enabling precise control and improved accuracy over previous methods.

Contribution

It presents a new computational approach that accurately models the guzheng's string and resonant body, enhancing real-time sound synthesis capabilities.

Findings

Improved sound accuracy compared to existing models

Efficient real-time synthesis with acceptable computational complexity

Successful synthesis of the 21st string of the guzheng

Abstract

Encouraged by recent interest in traditional Chinese instruments this work proposes a computational sound synthesis model for a traditional Chinese instrument, the guzheng. Digital waveguide model and modal synthesis are the most popular physical modelling methods. However, the excitation signal is usually over-simplified and the design of some digital filters is inadequate for real-time synthesis. We propose an accurate and efficient approach to synthesise the string signal of the guzheng. Our proposed model enables accurate parametric control of the pluck because the parameters are based on physical model of the instrument. The sound of the guzheng is then computed as the convolution of the synthesised string signal and the impulse response of the resonant body. As proof of concept, the 21st string of the guzheng is synthesised using the proposed model. The sound computed using our model shows improved accuracy with an acceptable computational complexity.