Bisping R., Dickson C., Khan S.

The acoustical properties of aircraft sounds dynamically change during the time-history of the fly-over. It is therefore difficult to determine the relationship between the human perception of aircraft sounds and these acoustical changes. One option could be to use time continuous scaling to find relationships between acoustical and perceptual data, e.g. by applying cross-correlation methods. Since aircraft sounds comprise many acoustical features which might change simultaneously this method has a limited range of validity. To overcome such problem in the present study synthesized stationary aircraft sounds were used instead of the natural aircraft sounds. This allows to change experimentally just one feature of interest at a given time. The sounds were generated to represent three different angles of the aircrafts flyover position relative to the observer at 78,7°, 90° and 101,3°. These three positions were found to cover a significant part of the acoustical phenomena which may occur during fly-over, e.g. tonal components, fan noise, low- versus high frequency broadband effects, etc. Synthesis was performed based on the measurements of six different aircrafts and two flight conditions (take-off & arrival). All the sounds were equalized to have the same EPNL-level. The sounds were judged by 25 subjects using a multidimensional scale having five different attributes: loudness, annoyance, harshness, power and pitch. The statistical analysis of the data showed highly significant differences between the acoustical phenomena and all chosen perceptual attributes. The paper will present the applied methods and the results. Auditory examples will be given in order to demonstrate the high quality of the synthesis procedure.