騒音レベルと音の大きさのレベルに関する統計的研究 : 騒音スペクトルに関する統計的研究第I報

DOI Web Site

書誌事項

タイトル別名
  • Statistical Examination of Sound Levels and Loudness Levels of Noises
  • 騒音スペクトルに関する統計的研究-1-
  • ソウオン スペクトル ニ カンスル トウケイテキ ケンキュウ 1

この論文をさがす

抄録

About 170 octave band spectra of actual noises were gathered from the papers is "Noise Control", Vol. 3 (1957) and Vol. 4 (1958) and inspected statistically. All of their sound levels (S. L. ) were estimated by using 4 weighting responses, C (flat), B (DIN-1), A (DIN-2) and DIN-3, and also their loudness levels (L. L. ) were computed by the S. S. Stevens' method. Tab. 1 shows an example of calculations where the band pressure levels of C are given. In Fig. 1, the ordinate is the difference of the number of L. L. and S. L. . The correlation between L. L. and S. L. of B (or DIN-1) seems to be most excellent in levels higher than 70 phons, but through all ranges, that of A (DIN-2) is best. Tab. 2 shows the fluctuation or accuracy numerically. For example by A or DIN-2, L. L. (phons)=S. L. (A)(dB)+14±5 in 97% accuracy. Fig. 2 is the average octave band spectra of noises. The left figure (1) shows them in pressure levels, and the right (2) in A weighting and in sones. Generally speaking, the sound of higher levels has more higher frequency components. To determine the importance of bands, one new concept "band weight" was introduced. Weight number "2" is given to the band which has the largest level. Say it Lm. Weight number "1" is given to bands, the levels of which are smaller than Lm and larger than (Lm-10)dB. "0" is given to the remaining bands, because they contribute almost nothing to the resulting level. In case of loudness, number "2" is also given to the band of the largest sones, and "1" to bands whose number of sones are larger than the half of the maximum sone. In Tab. 1, these numbers are also tabulated, and in Fig. 4, the ordinate shows the total number of band weights of each band. It shows the degree of relative importance of bands at every response. By A and sone (loudness), bands at middle frequency region are of great predominance, showing that the accuracy of measurement of middle frequency region is most important. The figure (2) of Fig. 2 also shows the same result. As conclusion, the author wants to propose the following points: (a) The sound level measurements by sound level meter should be made only by A response; (b) The tolerances of A response of sound level meter at higher frequency up to 5000 c/s should be a little smaller than those used at present.

収録刊行物

詳細情報 詳細情報について

問題の指摘

ページトップへ