A Rumbling Index Development for Sound Quality Analysis of A Vehicle

초록

Rumbling sound is one of the most important interior sounds of a passenger car. According to the conventional research. Evaluation of sound quality has been focused on the reduction of the A-weighted sound pressure level. However A-weighted sound pressure level cannot give the whole story about the sound quality based on human sensibility. Rumble index is required for the evaluation of sound quality of a vehicle. In order to make rumble index, the artificial neural network has been used. The sound metrics for the interior sounds of vehicles are used for the input of the neural network. The subjective ratio for the interior sounds of vehicles is used for the target of artificial neural network to train the weights of the neuron. For the train of the artificial neural network, 150 signals for rumbling sound are synthesized. The interior sounds of 16 vehicles are tested by using the trained artificial neural network. The output of artificial neural network becomes the rumbling index and has 90% correlation with the subjective results for interior sounds of 16 vehicles