With the spread of transportation systems on the road, the effects of traffic noise have become one of the most serious problems in our daily life and some epoch-making counter-measure has earnestly been waited for. We have to measure the traffic noise first in order to control it. Present experiments were designed to investigate what kind of estimated value of traffic noise is most suitable to predict the loudness of traffic noise. All the stimuli used in these experiments were the traffic noises recorded on magnetic tape. In Exp. I the loudness of 5sec traffic noise, generated when only one vehicle passed was obtained, using the method of subjects' adjustment. That is, subjects were asked to match the loudness of pink noise with that of traffic noise by controlling the level of pink noise using a remote control attenuator. The results show quite good correlation between loudness and any one of estimated values of traffic noise (dB(A), SIL, NRN, LL, PNL, L_<eq>). So long as the traffic noise generated by one vehicle is concerned, any estimated value seems to be suitable. Among them L_<eq> appears to be better since its standard error is small. In Exp. II the loudness of 1min traffic noise, during which various kinds of vehicles passed, was obtained. As in Exp. I the method of subjects' adjustment was adopted. From this experiment all the estimated values, except maximum level, TNI and L_<np>, were found to have high correlation with loudness. Among them L_<eq> had the smallest standard error. In Exp. III the impression of traffic noise used in Exp. I and II was measured using the semantic differential method. Subjects were asked to estimate the traffic noise using 13 scales of adjectives. As a result of factor analysis, three factors were extracted: powerful, pleasant, and metallic. Coefficients of correlation between estimated values of traffic noise and factor scores show that the upper values of traffic noise such as L_<eq> or L_5 have a high correlation with the "powerful" factor. From these experiments it might be concluded that L_<eq> is the most effective value as a representative of traffic noise since it is highly correlated with subjective estimation of traffic noise and for practical purposes it can be easily calculated. Our previous experiments concerning the loudness of level-fluctuating noise ascertain this fact. Further, it is said recently that it is the upper value of traffic noise that corresponds well with the annoyance of residents.
It is well known that the sound pressure level at the receiving point is given by an integral of a certain function over the plane space surrounding the plate. Maekawa has already derived the estimation procedure of noise reduction by a rectangular barrier, based on the Fresnel-Kirchhoff diffraction theory. However, we have no useful solution for the estimation of noise reduction by various shapes of structures. In this paper a practical method of estimating the noise reduction by various shapes of finite structures with knife and/or right-angled edges is proposed. In order to easily obtain the integral over the plane space around a polygonal plate, the plane space is divided into several spaces by the extended lines of each side. But in case of complicated shapes in a structure, the structure also is divided into adequate sizes and numbers. The integral for the plane space around a structure is obtained as the total of an integral over each divided space. The integral over each space is obtained by means of the practical charts each for a knife or a right-angled edge shown in the references 2) and 3). The noise reduction by a thick plate also is easily obtained by applying the "Method of Equivalent Sound Source. " The results of several kinds of scale models set up by knife or right-angled edges were compared with calculated ones. The calculated values shown are approximately equal to an average of measured values for the test frequencies from 2. 0 to 8. 0kHz (see Fig. 9). Also, it is found that the sound attenuation by a circular plate can be well approximated by a regular polygonal plate with many sides. This solution may conveniently be applied not only to the estimation of noise reduction by a single structure but also to multiple structures, and as long as the sound attenuation attained by every edge is found, the noise reduction for more complicated structures may also be estimated by the procedure mentioned in this paper.
Since 1971 several social surveys have been carried out to investigate annoyance resulted from noise and vibration by aircraft, high speed railways, road traffic and factories or plants under commission of the Environment Agency of Japan. Throughout these surveys, the author has taken part in planning the surveys and statistical analysis of the investigation. In this paper, several significant characteristics of individual/social attitude to noise and vibration are discussed from the common standpoint through these surveys including methodological aspects. As the research objectives treated in this paper were to quantify the relation between annoyance response to noise or vibration and properties of respondent including physical factors such as noise exposure etc. , samples collected by the social surveys and physical measurements were analyzed by multi-dimensional analysis (quantification theory of qualitative data proposed by Dr. Hayashi). The results clarified in these analyses are as follows: (1) Disturbances by noise exposure are classified into three distinguished groups such as social activity disturbances (conversation, telephoning and listening to TV & radio), human activity disturbances (sleep, consideration and others) and physiological affects (headache and others) as shown in Fig. 2. Social activity disturbances have a large correlation with noise exposure as a physical factor in scaling annoyance rating. On the other hand, significance of the attitude can not be found in the case of physiological affects by noise exposure below the noise level of 100dB(A). (2) Noise sources which generate noise intermittently such as aircraft and high speed railways give no significant difference to annoyance as shown in Fig. 7. (3) The attitude to vibration has a clear distinction from that to noise exposure. Through the most significant factor to decide the attitude to annoyance by noise is noise exposure itself as a physical factor, in the attitude to vibration, noise annoyance or experience of exposure to vibration as a psychological factor masks the physical factor for decision of attitude. As the result, the accuracy of analysis is less in scaling disturbances by vibration with physical factors than in the case of noise. (4) It may be generally said that the probability distribution of annoyance with noise level as the variable is fit well for a normal distribution which has the standard deviation of about 5dB. On the other hand, the standard deviation exceeds 15dB in probability distribution of attitude to disturbances by vibration.