From the technical, signal-oriented point of view, the production of speechis widely described as a two-level process (Koreman, 1996:1). In the firststage the sound is initiated and in the second stage it is filtered on thesecond level (Fant, 1960; Titze, 1994:136-168; Stevens, 1997:463). Thisdistinction between phases has its orgin in the source-filter model of speechproduction (Fant, 1960).
The basic assumption of the model is that the source signal produced at theglottal level is linearly filtered through the vocal tract. The resultingsound is emitted to the surrounding air through radiation loading (lips).The model assumes that source and filter are independent of each other. Althoughrecent findings show some interaction between the vocal tract and a glottalsource (Rothenberg 1981; Fant 1986), Fant's theory of speech production isstill used as a framework for the description of the human voice, especiallyas far as the articulation of vowels is concerned.
From the linguistic phonetic point of view, the production of speech is regardedas a superposition of initiation, phonation, articulation and prosodicorganization processes (Laver, 1994:116). An overview of the physiologicalconstraints on speech production will be given later in this section. Theorgans involved in the production of speech are depicted in the Fig.1.
Figure 1. The production of speech sounds -lungs, glottis, vocal tract (from:Bordenet al.,1994:64).
