Part III: Measuring hearing-critical job tasks 127 In this study, we focussed on the spectral elements of the warning signals, although the temporal structure of a warning signal can influence detectability (Misdariis et al., 2013). We observed distinctive differences in temporal characteristics between the two warning signals. The ATP signal decays over time and the DSD signal does not, whichmight facilitate discrimination between the two signals (Edworthy, 1994; Graham, 1999). Moreover, the ATP signal is presented once if a change in speed is required and three times when the change in speed is accomplished. If the maximum speed is exceeded, the ATP is presented with a longer duration of approximately three seconds. The ATP signal might be easier to detect if it is presented three times or with a longer duration. However, a detailed analysis of the temporal structure of the warning signals was beyond the scope of this study. Some study limitationsneed tobementioned. First, the criteria fromtheDetectsound model and the ISO are based on warning signals that can occur unexpectedly. In a train cabin, the auditory signals occur often and are thus expected and very wellknown by the driver. However, because it is vital that a locomotive engineer does not miss these warning signals, even in very tense situations, the detection models used are deemed relevant. Second, since the acoustical measurements took place in 2006 and 2007, the measurements in this study do not cover all train types that are currently in use. Also, all Mat64s have been decommissioned in 2016. The result that the detectability of the DSD and ATP signal was critical in the Mat64 is therefore less relevant for current practice. The train equipment of the included trains has not changed, with the exception that it used to be possible to open the window in some locomotive cabins, but at present these windows cannot be opened anymore. This does not influence the applicability of our results, since all measurements were performed with closed windows. Third, the experimental design did not allow full control over the selection of the railway tracks. Consequently, the number of measurements and the measured driving speed differed between the train types. Fourth, it is not possible to use these measurements to obtain an accurate estimate of the daily noise exposure of locomotive engineers. However, since the average measured sound levels were much lower than 80 dBA, it is reasonable to assume that prolonged driving on these trains does not exceed the current Dutch and European first action level of 80 dBA averaged over an eight-hour shift (STB10053, 2006; Directive 2003/10/EC).
RkJQdWJsaXNoZXIy MTk4NDMw