A method, the ‘read speech normalization’ method (RSN), is proposed by which the variability of prosodic parameters (rhythmic durational and intonation) can be compared across different conditions of spontaneous speech. In an experiment using a customized variety of the map task method, spontaneous speech was elicited from two speakers of English in a formal and an informal situation. Sentences from the spontaneously spoken formal and informal situations were afterwards read by the same speakers. The formal and informal conditions were then compared in terms of their differences to the read speech version (read speech normalization). Results showed that meaningful differences could be observed between formal and informal speech from the read speech normalized sentences that could not be obtained by comparing the two spontaneous speech conditions directly.