The invention provides an in-vehicle acoustic enhancement method and device fusing determinacy-randomness
hybrid modeling, and the method comprises the steps: collecting a fitted
vehicle acceleration sound sample, extracting audio data, carrying out the analysis, constructing a determinacy component synthesis
algorithm, and associating with a vehicle working condition; the method comprises the following steps: constructing a multi-
physics-field coupled
random noise generation module, wherein the multi-
physics-field coupled
random noise generation module comprises an aerodynamic
noise module for simulating air inlet turbulence, a low-frequency pulse module for representing mechanical shock, a
detonation module for simulating
combustion uncertainty and a
nonlinear distortion module for enhancing exhaust roughness and generating high-order
harmonics; each
random noise generation module is matched with the deterministic component, and mixing of the deterministic component and the random component is realized through
weight adjustment; details are enhanced through dynamic compression, and target sound
waves are generated; through collaborative optimization of determinacy-randomness components, the naturalness and brand identification degree of synthetic sound
waves are improved, the vehicle working condition and driving information can be converted into perceptible sound effect characteristics, and the problem of information
perception simplification in the
electric vehicle driving process is solved.