A New Random Phases Correction Algorithm for Synthesizing Engine Noise

Abstract

Order noise is an essential part of engine noise and a vital indicator of the vehicle's interior’s sound quality. Based on the Kaiser window function and Short-Time Fourier Transform (STFT), the synthesis techniques are proposed to extract and synthesize engine order sounds. However, using the Short-Time Fourier Transform algorithm to extract order noise is prone to spectral leakage and fence effects, affecting the accuracy of order tracking analysis. In this paper, the pitch-synchronous sinusoidal decomposition method is used to analyze the spectrum characteristics of some essential window functions and their differences. The Kaiser window function short-time Fourier transform, and the synthesis techniques are proposed for the extraction and synthesis of engine order sounds. A random phase correction algorithm is developed to conquer the problem that the phase discontinuity of a single harmonic component during order synthesis leads to the interference effect of the synthesized sound. After all, it improves the coherence of the synthesized sound. Moreover, MATLAB software and engine order-noise are applied for producing a sound modulation software to achieve order slicing, phase extraction, and sound synthesis of sound samples. The results show that: the engine order sound extraction and synthesis method can meet the requirements of synthesized sound accuracy; the proposed random phase correction algorithm can reduce the interference effect and improving the coherence of the synthesized sound.