Wavelet-Based Vehicle Sound Quality Improvement

As people's living standards improved, the material requirements of life are continually improved, In the automotive field the performance is comfort requirements. the car noise is a very important part of comfort, The noise inside the car would take the officers have a sense of irritation,Irritable feeling reached a certain degree it will affect the lives of persons inside the vehicle ride. Sound quality is also directly felt by consumers, The quality of the resulting sound quality will affect consumer purchasing products, This is the increasingly fierce market competition, the designer must pay attention to the problem. Now selected a car, collected noise under different conditions and make noise reduction processing, Major work and the results achieved are the following:First introduced the human auditory system, the various components of the working principle as well as some basic characteristics, Such as the masking effect of hearing,thresholds,pain domain. Describes the sound quality is also an objective evaluation of several parameters: loudness, sharpness, roughness, using the formula in the form of an objective interpretation of the various parameters, but also gives them the computing model. This model can be used for an objective assessment of sound quality.Second, about the wavelet transform algorithm, Systematic exposition of the definition of the wavelet theory, and various forms of wavelet transform, and compare with the Fourier transform, describes unique advantages of the wavelet transform in signal processing. Advantages of wavelet analysis is a good localized nature, both in the time domain and frequency domain. Practical application in the time domain and frequency domain sampling step size automatically adjusted to accommodate the signal analysis of the different frequency components, Solution can not be processed problems in the time domain and frequency domain signal processing. The paper from the perspective of the frequency response demonstrates the application of wavelet transform for signal noise reduction processing is possible, Namely, construct a family of filters with wavelet transform to achieve human cochlea frequency characteristic curve approximation. Illustrated in the text of wavelet transform applied to the actual noise reduction problems, Application of Wavelet Analysis Toolbox Graphical User Interface implementation of wavelet noise reduction processing, In addition, describes the algorithm transform signal decomposition and reconstruction based on wavelet.Finally is the experimental verification link, Layout collection of artificial voice signal in the car,and used Paired comparison method make a subjective evaluation of sound signals, At the same time with the Artemis software acquisition objective analysis and evaluation of sound signals, Obtained loudness, sharpness, roughness and other objective parameters values. The core work of this thesis is the application of wavelet analysis toolbox graphical user interface, interface implementation of wavelet noise reduction processing. Dealing with the voice signal again after the objective and subjective evaluation, compared before and after de-noising concluded that: wavelet transform algorithm significant on noise reduction.Finally reached the following conclusions: The main factors that affect the sound quality is the degree of loudness and sharp, loudness is mainly determined by the low-frequency components, the low-frequency components and the engine speed under the various second-order resonant frequency-related; Sharp degree of influence mainly by the high frequency components, above 16Bark of sharp degree are high. These high-frequency components of the noise formed by the engine combustion decision. These conclusions guiding on the design works.