How Long Is Enough? Assessing Representative Duration in Military Jet Noise Recordings

Selecting a representative duration for noise recordings from military jet noise is the first step when studying such sources. Current standards and studies struggle to give a uniform answer, particularly for static engine run-ups, where no natural event duration exists. This study investigates the minimum signal duration required for F-16 static run-up recordings to remain representative of the full acoustic event. After a review of current studies and standards, and framing the problem from a physical perspective, the research focuses on the experimental recordings. Long measurements acquired at different thrust-ratings and distances were divided into equal-duration segments. These segments were compared both with the complete recording and with other segments of the same duration using Leq, Lpeak, 1/3-octave spectra, loudness, and sharpness, to guarantee both equivalence to the complete recording and the satisfaction of the quasi-stationary assumption. The results show that the required duration depends strongly on the selected metric, propagation distance, and thrust setting. Near the source, statistically useful segments may be shorter than 1 [s], whereas at larger distances tens of seconds may be required, with the most conservative case reaching 54.81 [s]. Convergence is governed by both the magnitude and nature of noise fluctuations, and the spectral analysis showed that the low-frequency content, as well as mid-frequencies when farther away from the source, play a dominant role. This was further confirmed by the sharpness, which stabilizes earlier than loudness. These results argue against universal recording durations and support criterion-dependent duration assessments in future military jet-noise studies.