Global and local sound quality indicators for urban context based on perceptive and acoustic variables

Invited paper

Catherine Lavandier

Université de Cergy Pontoise

Tuesday 2 june, 2015, 14:20 - 14:40

0.6 Madrid (49)

The aim of this project is to propose urban sound quality indicators based on acoustic and perceptive data. A mobile application has been then developed. After calibration of the mobile microphone, the application proposes to collect, in addition to the noise levels, perceptive data in about 30 places of Paris. More than 3000 measurements have been performed by 60 participants. Each measurement corresponds to a 10-minute recording of sound pressure level (stored each second). Energy indicators such as LAeq, LA10, LA50 or LA90 and event indicators such as the number of noise events exceeding a certain threshold Lα (NNE≥Lα) are extracted from these acoustic measurement. Following the recordings, participants have to answer a short questionnaire: the first questions are related to the sound environment characterization with semantic scales (pleasantness, liveliness, overall loudness and envelopment feeling), the next questions concern the perceived loudness of some special sources such as cars, motorbikes, trucks etc. and the last questions regard the presence time ratio of other sources (traffic, voices, steps, birds, etc.). One global and different local sound quality indicators are proposed, based on multiple linear regression models built on perceptive variables. The overall loudness has the most important impact on sound quality for all models but the presence of some sources influence also this quality on a negative (traffic) or positive a (steps, voices or birds) way. Each influent perceptive variable is correlated to the acoustic parameters. The global loudness is best correlated to LA50. The events over 70 dB(A) seem to characterize the presence of heavy vehicles or motorcycles. A small difference between LA10 and LA90 seems to characterize the human presence. It seems that no acoustic indicator is able to characterize the envelopment which is nevertheless influent for sound quality.

ICS file for iCal / Outlook

[ Close ]