A procedure to measure industrial impulse noise was developed that provides information on both the equivalent noise level and impulsiveness. Impulsiveness of the signal was defined as the difference betweeen the A-weighted peak level and the root-mean-square level. This difference, also known as a crest factor, varies in time, so the cumulative distribution function was used to describe the probability to achieve a certain value of impulsiveness. A telemetric sampling system carried by the worker was used to take measurements at a shipyard. The sample duration of 10 minutes was considered sufficient to cover welders', platers', and grinders' different work phases. The total number of 10 minute samples was 123. The welders were exposed to the most impulsive noise, whereas the grinders were exposed to the highest equivalent noise level. The present paper describing a procedure to measure industrial impulse noise comprises part of a follow-up study on the effects of noise on hearing.