An ANN has been applied to the development of the sound quality index of the booming sound of passenger cars. The 150 synthetic signals are used for the training of the ANN. The ANN used in the present paper is a back propagation neural network with 2-10-1 structure. The input vector of the ANN has two elements and the output vector has one element. The averaged subjective rates for the interior sounds of 16 mass-produced passenger cars are used for the confirmation of the trained ANN being used for booming index. Loudness, sharpness, roughness and fluctuation strength for those interior sounds are calculated for the input of the ANN. It is found that both loudness and sharpness for those sounds appear to have a strong relationship with the averaged subjective rates of those sounds. In particular, the loudness of the signals filtered by a low pass filter is required to get a good correlation. The correlation between neural output and the averaged subjective rate for those sounds is 97.5%. It is concluded that the output of the trained ANN can be used for the booming index for the interior sounds of passenger cars.
Identification of vehicle booming sound and its objective evaluation using psychoacoustic parameters
Identifizierung des dumpfen Fahrzeuggeräusches und seine objektive Auswertung mittels psychoakustischer Parameter
International Journal of Vehicle Design ; 58 , 1 ; 46-61
2012
16 Seiten, 12 Bilder, 4 Tabellen, 18 Quellen
Aufsatz (Zeitschrift)
Englisch
Identification of vehicle booming sound and its objective evaluation using psychoacoustic parameters
Online Contents | 2012
|Understanding Psychoacoustic Parameters
SAE Technical Papers | 2001
|On an Objective Measure of Booming Sound Factor (9630750)
British Library Online Contents | 1996
|Psychoacoustic modelling of sound attributes
Kraftfahrwesen | 2006
|