The noise signal from an individual vehicle or succession/platoon of vehicles within a road traffic stream is highly variable and in reality identifying a representative sample of the population of the noise level maxima of vehicle pass-by for the range of real road traffic noise conditions (combinations of vehicle speed, vehicle type, road surfaces, propagation path) is an expensive and time-consuming measurement task. This paper reports on a secondary analysis of a large sample of maximum pass-by noise data that was collected in Brisbane, Australia. The data set provides the noise level maxima (LAFmax) of the pass-by of some 85,000 vehicles in service on urban arterials and motorways. Concurrent attended monitoring of vehicle type has allowed an analysis of the distribution of the noise level maxima according to some road characteristics and for the range of the different types of vehicle. There have often been suggestions that road traffic noise criteria should include, in addition to integrated energy measures such as Leq, Lden or similar, some metric of the individual noise events (LAFmax, SEL) or number of noise events, generated by individual noisier vehicles within the traffic stream. Apart from selecting such a metric in terms of its relationship to health impacts, such as sleep disturbance, that metric needs to be able to be measured reliably, and preferably able to be predicted. The aforementioned monitoring program identified that, whereas field measurement of descriptors such as Leq have a well-established protocol, there is no similar protocol for the in-situ measurement of noise events. This paper enumerates the issues relevant to the measurement of noise events in road traffic streams.