Spectral and temporal aspects in the annoyance caused by low frequency sounds

About two decades ago, Vercammen proposed a rating procedure for low frequency sounds. To enable refinements of this procedure, the present study addressed a few spectral and temporal issues. The investigation was carried out in the laboratory with 56 subjects ranging between 20 and 60 years of age. To increase our insight into the spectral issue, various triads of simultaneous 1/3-octave noise bands with central frequencies of groups of 40, 50, and 63 Hz were presented at several sound level combinations. It was concluded that an energy-based summation model predicted the annoyance better than the level of a dominant single frequency band did. Especially on the basis of the results in related conditions, where the acoustic conditions were most appropriate to identify the best frequency weighting, it is advised to apply the summation to A-weighted instead of linear sound levels. With regard to unsolved temporal issues, the effects of two types of sound level variations were investigated. Sound fragments, for example those consisting of aircraft departures with silent intervals between the successive events, were presented for rating. For the aircrafts sounds the higher linear sound levels were found in the 1/3-octave bands around 25, 31.5 and 40 Hz. The results showed that when the time between successive sound events matches situations around busy airports or along secondary roads, the annoyance is well predicted by the equivalent sound level. For the annoyance caused by amplitude modulated (AM) low frequency sinusoidal tones and broadband noise (fmod = 1 Hz), however, both overall level and AM depth have to be taken into account.