In addition to the findings already presented during the last TF meeting in Brussels, a more detailed analysis of the study on detection distances as function of various sounds was shown. All analyzed sounds are suitable to meet the basic NHTSA requirements for detection distance (1.8sec in front of pedestrian @10 km/h), however only three of them were designed in line with the NPRM requirements for number and quality of the 1/3 octave bands (8 bands with individual SPLs). 1 of them actually fulfills the requirements, while the other two miss to fulfill SPL requirements in at least one 1/3 octave bands. This may be a consequence of sound propagation of a sounder fitted to a vehicle being unstable and depending on so far unknown factors.
Subsequently it is likely that sound emitted from the speaker must be produced to include a certain safety margin so to be sure that the signal measured according to ISO 16254 does fulfill all criteria. This off course risks resulting in a higher overall SPL.
Emissions of all three sounds are higher that the emission from an ICE that had also been included in the study.
Actual detection distances are between 8 and 29 meters with sound levels between 39.8 dB(A) and 63,7 dB(A) at the moment of detection and between 52.9 dB(A) and 72 dB(A) in front of the pedestrian. Further analysis reveals that a sufficiently high level of detectability can be reached with signal emitted from only two bands, either according to proposed specification from Japan (2 bands, one above 1 kHz, one below 800Hz, each SPL >43 dB(A) ) or to the proposal from VDA (2 bands, one below 1.6 kHz, individual SPL level for each band) .The detectability does not improve with the number of bands. Also there is no clear relation between the overall SPL from the all 1/3 octave forming the individual sound and the distance at which a vehicle can be detected. Even a lower over SPL (and less energy) produced from less 1/3 octave bands may result in a better detectability.
Analysis of data from ICE vehicles show that the typical sound emission according to ISO16254 is about 60 dB(A) in average (2m / 10 km/h) with sedan type vehicles typically emitting between 55 and 60 dB(A). As sedan are sufficiently detectable in normal urban, it can be concluded that a SPL below 60 dB (A) should be adequate for detectability.
In summary it can be concluded that the frequency range from 160Hz to 5000Hz is a good range and any frequency is suitable for detectability. It can be discussed to blank some bands around 1000Hz, but there is no real need for it. If the sound levels are selected with care, there is no risk to jeopardize traffic noise reduction aims. Two bands are sufficient as requirement; the vehicle will add “natural additional sound” in many cases. An overall sound level per test condition should be specified as well. It is suggested to invite all parties that have generated data, to merge their data together.
The whole database should be re-analyzed to determine, re-assess or validate the given proposals for sound levels again.