Speech consists not only of linguistic elements such as phonemes – it also contains features which carry nonlinguistic information. These features, e.g. voice quality and prosody, either occur spontaneously and are related to the speaker's emotional and/or physiological state or are produced intentionally and related to the speaker's attitudes or interaction strategies (even though the latter often include features that are meant to appear as unintentional). Acoustic analysis of the human voice thus allows the identification of vocal correlates of the speaker's emotional and/or physiological state as well as of his social intentions.