Statistical sampling in community noise measurement

✍ Scribed by D.J. Fisk

• Publisher: Elsevier Science
• Year: 1973
• Tongue: English
• Weight: 905 KB
• Volume: 30
• Category: Article
• ISSN: 0022-460X
• DOI: 10.1016/s0022-460x(73)80115-4

✦ Synopsis

The statistical theory of community noise sampling is investigated and used to derive the optimum sampling procedures. General expressions are given for the sampling errors in L~o and other percentiles. The correlation between percentile sampling errors is discussed in relation to the sampling error in TNI and to the correct procedure ofdata analysis.

The traffic noise models of Weiss and Kurze are used to derive a sampling theory for L,q. It is shown that, for continuous sampling, the sampling error, AL,q, can be estimated by AL,q ~ 7.4/v'hi dB(A) where m is the number of vehicles passing during the sample. A:" sampling rate of 2 per second is shown to be a good approximation to continuous sampling. The model is used to demonstrate that, when sampling with a sound level meter, using a fixed number of readings, the readings should be at least 5 seconds apart. The use of very short analogue tape samples (microsampling) is shown to lead to improved sampling accuracy on theoretical grounds.