Disc brake vibrations are a cause of concern for automobile users and manufacturers. The user can find the noise of brake squeal annoying and sometimes worrying. The manufacturer suffers the immediate costs involved in vehicle warranty claims and may lose sales if a particular vehicle model develops a reputation for noise problems. The measurement of disc brake squeal and its relationship to audible and structureborne vibration measurement is difficult to determine without influencing the vibration by the measurement technique. Ideally, measurements would be taken under 'real' conditions. The complexity and cost of 'in-car' testing was prohibitive. Instead, the approach taken was to use a real car brake system, but mounted in such a way that the support structure carrying the knuckle was as rigid as practically possible. This allowed an array of non-contact displacement transducers to measure the vibration at the disc surface. This experimental approach was used to investigate how the displacement signals produced by a squealing disc could be processed to determine the mode shapes of the disc and the behaviour of the waves that generate the audible squeal noise. The two methods used in the analysis of the squeal data were the fast Fourier transform of each data channel and the least-squares analysis of the time domain data. The least-squares approach was the more difficult to achieve, but successfully produced detailed analysis of the behaviour of the wave motion in the disc from the available data.A finite element analysis of the brake structure compared favourably with the frequencies and mode shapes obtained by both modal hammer tests and squeal tests. The use of this type of dynamometer, measurement system and analysis method has proved successful in providing a deeper understanding of the mechanisms involved in disc brake squeal.