CN-122021078-A - Physical modeling computer simulation method for saxophone sound

CN122021078ACN 122021078 ACN122021078 ACN 122021078ACN-122021078-A

Abstract

The invention relates to the technical field of computers, and discloses a physical modeling computer simulation method of saxophone sound. The method comprises the steps of dividing an acoustic cavity of a saxophone into a plurality of acoustic subsections according to geometric parameters and fingering configuration of the saxophone, modeling by adopting a one-dimensional non-uniform digital waveguide, a two-dimensional axisymmetric finite difference time domain or a three-dimensional boundary element local resonance unit according to geometric features of each subsection, constructing an interface coupling model meeting sound pressure continuity and volume speed conservation, synchronously calculating the sound field state of each unit through a multi-rate time domain recursion algorithm by combining a reed excitation model based on a nonlinear Bernoulli equation, and outputting high-fidelity audio data streams. According to the invention, through heterogeneous modeling and multi-rate strategies, the simulation efficiency is obviously improved while the fidelity of tone is ensured, and the cooperative optimization of precision and real-time performance is realized.

Inventors

ZHOU XUEWEI
LV XIAONING
XIAO YUFENG
PAN YONG

Assignees

杭州智歆科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260414

Claims (10)

1. A physical modeling computer simulation method of saxophone sound, comprising: Obtaining geometric parameters, material properties and fingering configuration information of the saxophone; Dividing the saxophone acoustic cavity into a plurality of acoustic subsections according to the geometric parameters and fingering configuration information, wherein each acoustic subsection corresponds to an independent acoustic propagation path and comprises one or more combinations of a straight pipe section, a conical expansion section, a bent pipe section or an open hole section; Selecting a corresponding physical modeling unit type according to geometric form characteristics of each acoustic subsection, wherein the physical modeling unit type comprises a one-dimensional non-uniform digital waveguide unit, a two-dimensional axisymmetric finite difference time domain unit or a three-dimensional boundary element local resonance unit; establishing an acoustic interface coupling model between each physical modeling unit, wherein the acoustic interface coupling model is used for describing sound pressure continuity conditions and volume speed conservation conditions of adjacent acoustic subsections at the junction; Applying a reed airflow excitation model at an excitation end, wherein the reed airflow excitation model is solved based on a nonlinear Bernoulli equation and a reed displacement dynamics equation in a combined way, and a sound source signal changing with time is generated; Sequentially calculating sound field state variables in each physical modeling unit through a time domain recursion algorithm, and realizing cross-unit data transmission and synchronization by utilizing the acoustic interface coupling model; and finally, collecting the synthesized sound pressure signal at a receiving end and outputting the synthesized sound pressure signal as an audio data stream.
2. The method for computer simulation of physical modeling of saxophone sound according to claim 1, wherein the step of dividing the saxophone acoustic cavity into a plurality of acoustic subsections comprises: Identifying geometric abrupt points in the main body of the saxophone, wherein the geometric abrupt points comprise pipe diameter change points, bending turning points and button opening positions; dividing the whole lumen into continuous acoustic subsections which are not overlapped with each other by taking each geometric abrupt change point as a dividing node; For each acoustic subsection, its start coordinates, end coordinates, cross-sectional shape function, wall impedance characteristics, and the presence or absence of lateral openings are recorded.
3. The method for simulating the physical modeling of saxophone sound according to claim 2, wherein the one-dimensional non-uniform digital waveguide unit is suitable for a straight pipe section or a conical section with a pipe diameter that changes slowly along the axial direction, the internal sound field is represented by a pair of counter-propagating traveling wave variables, the traveling wave variables are recursively updated by a discretized one-dimensional wave equation, and the wave speed and the characteristic impedance are dynamically adjusted according to the local cross-sectional area and the equivalent sound velocity.
4. A method according to claim 3, wherein the two-dimensional axisymmetric finite difference time domain unit is suitable for a curved pipe section or a large caliber expansion section with a significant transverse sound field gradient, the control equation is an acoustic wave equation under a cylindrical coordinate system, the sound pressure and the particle velocity are spatially and spatially separated by adopting a staggered grid finite difference format, and the time step satisfies Ke Lang-friedrichs-lux stability conditions.
5. The method according to claim 4, wherein the three-dimensional boundary element local resonance unit is adapted to a local region including complex open-cell structure or cavity coupling, calculates an acoustic impedance matrix of the local region in a frequency domain by solving a helmholtz integral equation, and converts the acoustic impedance matrix into a time domain convolution kernel by inverse laplace transform, for updating a sound pressure and normal velocity relationship on a boundary in real time.
6. A method of computer simulation of physical modeling of saxophone sound according to claim 3, wherein the construction of the acoustic interface coupling model comprises: defining the sound pressure value on the public node as the weighted average value of the output sound pressure of the two side units at the interface of two adjacent sound subsections, wherein the weight coefficient is determined by the characteristic impedance of the two side units; Meanwhile, defining the total volume speed passing through the interface as the sum of the unit volume speeds at two sides, and forcing the total volume speed to meet the law of conservation of mass; For interfaces where lateral openings exist, an opening radiation impedance term is additionally introduced, which is determined by the opening diameter, edge correction factor and surrounding medium density together and added to the equivalent load impedance of the main channel.
7. The method for computer simulation of physical modeling of saxophone sound according to claim 6, wherein the construction of the reed airflow excitation pattern comprises: Establishing a nonlinear feedback relation between reed displacement and sound pressure in a blowing nozzle cavity, wherein the reed displacement is constrained by reed rigidity, damping coefficient and static pretightening force; Calculating the volume flow entering the pipe cavity according to the relation between the instantaneous air flow speed and the pressure difference of the inlet of the blowing nozzle and combining a continuity equation; the volume flow is input as a sound source term to the excitation port of the first acoustic subsection.
8. The method according to claim 7, wherein the time-domain recursive algorithm uses a uniform time step to update all physical modeling units synchronously, and uses a multi-rate sampling strategy for two-dimensional or three-dimensional units with higher computational complexity, i.e. uses smaller time steps to perform local iteration in the units, and only outputs the result of integral multiple of the main time step when data exchange between units is performed, so as to maintain global timing consistency.
9. A method of computer simulation of physical modeling of saxophone sound according to claim 8, wherein the spatial step size of the two-dimensional axisymmetric finite difference time domain unit And Are all set to 0.5 mm, time step Satisfy the following requirements And performs local iterations internally at twice the subdivision rate of the main time step, Is the speed of sound.
10. The method according to claim 9, wherein the time domain convolution kernel of the three-dimensional boundary element local resonance unit is fitted by a linear combination of exponentially decaying sine functions, and recursive computation is implemented by a plurality of second-order IIR filters.

Description

Physical modeling computer simulation method for saxophone sound Technical Field The invention belongs to the technical field of computers, and particularly relates to a physical modeling computer simulation method of saxophone sound. Background With the rapid development of digital audio technology and virtual musical instrument simulation, the synthesis method based on physical modeling is widely applied in the fields of music production, education and man-machine interaction due to the advantages of high fidelity and expressive force. The physical modeling simulates vibration, propagation and radiation mechanisms inside the acoustic musical instrument through mathematical equations, and aims to reproduce dynamic response and tone details in real performance. Waveguide synthesis (Waveguide Synthesis) is one of the mainstream technologies, and a time domain recursion structure is used for efficiently simulating the propagation process of one-dimensional sound waves in a pipeline, and has been successfully used for simulating string instruments, copper tubes and woodwind instruments. However, this approach presents significant challenges in handling instruments with complex geometries and nonlinear airflow excitations. Saxophones are a typical cone-cylinder hybrid resonance cavity whose sound generation is highly dependent on the nonlinear interaction of the reed-pipe coupling system with multiple order resonance modes. Conventional waveguide models typically reduce the acoustic propagation path to a single or limited number of linear channels, and it is difficult to accurately characterize the dense distribution of overtones in the tube, cross-modal coupling, and the "growling" and other unsteady acoustic phenomena that result from strong airflow excitation. In order to improve simulation precision, the existing scheme approaches complex resonance by increasing the number of waveguide branches or introducing a high-order differential equation, but the improvement causes exponential increase of computational complexity, and severely restricts the instantaneity and practicality. In the prior art, when saxophone sound is simulated, the problems that a large number of resonance paths and mutual interference effects thereof need to be tracked simultaneously in fine modeling, and the traditional serial waveguide architecture cannot process multipath superposition in parallel, so that the problems of phase distortion, spectrum fracture or too high delay occur in a high-frequency overtone area or transient special effect simulation are generally existed. Particularly, in a real-time playing scene, the system is difficult to synchronously maintain the integrity of a low-delay response and high-dimension resonance structure in a limited computing force, so that tone quality is dry and dynamic response is delayed, and the severe requirements of professional-level virtual musical instruments on expressive force and interactivity cannot be met. Therefore, a new computing architecture is needed to break through the efficiency bottleneck of the traditional waveguide model in complex resonance simulation while maintaining physical rationality. Disclosure of Invention The invention provides a physical modeling computer simulation method of saxophone sound, which aims to solve the technical problem that the traditional waveguide model has low calculation efficiency when simulating complex resonance and cross mode coupling. According to the method, a mixed physical modeling architecture based on combination of a segmented non-uniform transmission line and local modal decomposition is constructed, a saxophone acoustic system is decomposed into a plurality of sub-structural units with definite physical boundaries and coupling relations, and a special numerical solution strategy matched with geometric characteristics and acoustic behaviors of each sub-structural unit is adopted, so that the acoustic simulation precision is ensured, and meanwhile, the calculation efficiency is improved. The invention provides a physical modeling computer simulation method of saxophone sound, which comprises the following steps: Obtaining geometric parameters, material properties and fingering configuration information of the saxophone; Dividing the saxophone acoustic cavity into a plurality of acoustic subsections according to the geometric parameters and fingering configuration information, wherein each acoustic subsection corresponds to an independent acoustic propagation path and comprises one or more combinations of a straight pipe section, a conical expansion section, a bent pipe section or an open hole section; Selecting a corresponding physical modeling unit type according to geometric form characteristics of each acoustic subsection, wherein the physical modeling unit type comprises a one-dimensional non-uniform digital waveguide unit, a two-dimensional axisymmetric finite difference time domain unit or a three-dimensional boundary eleme