Current developments in standardisation of noise damage risk criteria (DRC) are outlined together with an expression of the outstanding gaps in the scientific knowledge required for formulation of criteria: these concern the effects of impulse noises, intermittencies and strong pure-tone components. The methods of measuring impulse noise and their interpretation are discussed. Repetitive discrete impulses provide the area of greatest uncertainty and information in the literature is critically reviewed. A relationship between DRC noise level and the number of impulses is proposed, which is intermediate between logarithmic and linear and then curves back for large numbers of impulses. The most pressing areas for further research are discussed.