The statistical background noise correction is a recently developed approach for producing source descriptions from measured source levels with background noise contributions that are unknown because of variations in background noise across the distinct time spans of source and background measurement. The appeal of the statistical background noise correction is the ability to provide physical source descriptions that envelope the true source and are minimized for measured sources within 1 to 8 dB of the background, while also potentially eliminating the incentives against high precision grade methods inherent in standardized background noise corrections. A present obstacle against the statistical background noise correction is the unknown amount of background noise sampling needed to implement the statistical background noise correction. Accordingly the effect of the number of background noise samples on the performance of the true source envelopes and physical source descriptions produced by the statistical background noise correction are investigated by Monte Carlo simulation. Guidelines for background noise sampling are given based on the results.
展开▼
