Reducing contaminating noise effects when calculating low-boom loudness levels

Abstract

During NASA X-59 quiet supersonic aircraft community response tests, low-boom recordings will contain contaminating noise from instrumentation and ambient acoustical sources. This noise can inflate sonic boom perception metrics by several decibels. This paper discusses the development and comparison of robust lowpass filtering techniques for removing contaminating noise effects from low-boom recordings. The two filters are a time-domain Butterworth-magnitude filter and a frequency-domain Brick Wall filter. Both filters successfully reduce noise contamination in metric calculations for simulated data with real-world contaminating noise and demonstrate comparable performance to a modified ISO 11204 correction. The Brick Wall filter's success indicates that further attempts to match boom spectrum high-frequency roll-off beyond the contaminating noise floor are unnecessary and have marginal improvements on final metric calculations. Additionally, the Butterworth filter removes statistical correlation between ambient and boom levels for a real-world flight campaign, adding evidence that these techniques also work on other boom shapes. Overall, both filters can produce accurate metric calculations with only a few hundred hertz of positive signal-to-noise ratio. This work describes methods for accurate metric calculations in the presence of moderate noise contamination that should benefit X-59 and future low-boom supersonic aircraft testing.