Methods for assessing noise from high energy wideband impulsive source
s currently use either linear peak descriptors or ones based on freque
ncy weighted r.m.s. values. Significant errors can be introduced if su
ch assessments ave made using data obtained from instruments with limi
ted bandwidth or with ill-defined filter or cut-off chacteristics. Of
particular importance are the errors introduced when extrapolating emp
irical equations for the prediction of impulse noise levels to differe
nt size sources and different distances. This note considers the effec
t of using band limited data when making predictions over hard and sof
t surfaces and concludes that such predictions will be move reliable i
f full frequency waveforms are used from source to final location.