Search

The precision of sound power determinations from typical reference sound sources has been the topic of a recent study at the National Bureau of Standards. The 425 m3 reverberation room was qualified for measurements on pure tone sources in the 100-2500 Hz frequency range, according to the procedures of ANS S1.21-1972 [see Noise Control Engineering 7(2), 71-80 (1976)]. The sources which were studied included those manufactured by B & K, Electricite de France, and ILG. The study involved numerous one-third octave band sound power determinations of each source at each of three source positions. The experiment was designed such that variance due to changes in source output with time or position might be identified and that sound power would be determinable to within ± 0.5 dB at the 95% confidence level. The results to be presented include the estimation of the source sound power level and associated variance over numerous repetitions at single source positions as well as the mean source sound power levels and variance over three source positions. [ABSTRACT FROM AUTHOR]

Four community sites were monitored continuously for periods of from one to three weeks. A-weighted sound pressure levels were digitally recorded every 0.1 sec, wind data were recorded simultaneously. Proposed temporal sampling schemes may be efficiently 'tested' by processing this data on a high-speed digital computer, comparing the results obtained by using the sampling schemes with results computed from the entire data set. A priori predictions of accuracy of the sampling schemes are compared with the accuracy actually achieved in this experiment. Results using several familiar sampling techniques are presented. [ABSTRACT FROM AUTHOR]