Laboratory Sound Quality Evaluation Of Transmission

With the continuous development of automobile NVH(Noise, Vibration, Harshness) technology, engine noise, wind noise and road noise are effectively reduced over the past decades. However, the transmission noise like gear whine, masked by other noise sources before. At present,the evaluation of the transmission noise is mainly to control the sound pressure level. In fact,when transmission noise can meet requirements of the sound pressure level,however,people can still hear in the vehicle uncomfortable the transmission of noise.So,people need a new method to evaluate the transmission noise urgently.According to the current demand, this paper studied the transmission of sound quality evaluation method.This paper presents in detail on the characteristics of human auditory system, subjective evaluation and objective parameters. Through a series of experiments, the characteristics of transmission noise are obtained. And a multi-dimensional regression model quantified by psychoacoustic objective parameters is established using mathematical and statistical software, which has a high correlation with the subjective assessment result. The scale of subject evaluation on the sound quality of transmission is obtained through the study on the subjective evaluation indexes which can characterize the auditory properties of sound quality for transmission. Main research contents are as follow:1) Based on hemisphere envelop method, the vibration and sound signals at different positions of the eight transmissions are obtained by using LMS test system. Through spectrum and order analysis, we find that uncomfortable noise is largely affected by the order of howling in particular that it is most obvious in second gear in the speed range of 3000r/min-4000r/min. This provides a reference for subjective assessment of transmission noise.2) Sixteen noise signals of each gear are extracted from the original data to be used as the sound samples for subjective sound quality evaluation. The raw data are edited using audio software. Within appropriate testing environment, 35 subjects are chosen to conduct subjective evaluation experiment using ranking method and 10 adjectives are finally filtered out from a group of sound property adjectives to determine the subjective evaluation scale.3)11 objective parameters, such as kurtosis, the loudness over time, sharpness andjitter are calculated of each gear noise signals, and the subjective parameters are filtered out through P-P diagram and cluster analysis. Then correlation analysis is applied among these subjective and objective parameters obtained before. It is found that the speech articulation, loudness over time, SIL-4 and loudness degree has a great effect on subjective annoyance, which means that they have a high correlation coefficient. Using the level of annoyance as induced variable, and the loudness over time, SIL-4, sharpness and loudness as variables, multiple linear regression function is established and tested by Significance, which can exactly represent the relationship between subjective annoyance and objective parameters.