The midrange driver of a loudspeaker system has the most difficult job: it has to produce the frequencies the human ears are most sensitive to. It is also responsible for handling the major portion of the spectrum comprising the human voice and the fundamental notes emitted by most instruments. Hence it becomes imperative for it to behave well across this entire mid-frequency range. A loudspeaker driver suffers from resonance or ‘ringing’ at various frequencies due to structural limitations. This is the primary cause of artificial colouration of sound or inaccuracy in reproduction of timbre. This project attempts to detect and analyse the inherent resonant modes of a high fidelity midrange dome driver used in Linn loudspeaker systems using Empirical Mode Decomposition. The measurements have been made using a software programmed in MATLAB for accurate identification of resonant modes. The method described in this thesis provides a different approach for accurately detecting colouration or timbral inaccuracies in loudspeakers due to resonance of the transducer components. Limitations of the conventional FFT spectral analysis methods such as spectrogram and Cumulative Spectral Decay have been reviewed with regard to analysis of colouration.