ABSTRACT
The present chapter focuses on theories of the perception of oral vowels. Although these are relatively simple stimuli compared to other phonetic categories, many problems remain. Because of their relative prominence in speech-perception research, we focus primarily on formant-based representations and the problems they pose. These include issues ranging from speaker normalization to how formant frequencies may be extracted and interpreted by listeners. We will also discuss formant models in the context of several alternatives that are popular especially in engineering applications such as automatic speech recognition.
