Speech and Audio Signal Processing: Processing and Perception of Speech and Music, Second Edition

CHAPTER 17

SPEECH PERCEPTION

17.1 INTRODUCTION

How do people perceive speech? This question, in addition to being of great scientific interest, is also of central concern to those concerned with building systems for the processing of speech. For instance, engineering choices in the development of speech coding or synthesis systems should incorporate knowledge of what distortions affect intelligibility or voice quality for the human listener. It could also be argued that the human system for speech recognition and understanding is the one known example of a robust speech recognizer, that is, a recognizer whose performance is insensitive to variability over the range of nonlinguistic factors in the speech signal. Although simple mimicry of human mechanisms may not be a good approach to engineering design, it is still likely to be useful to study the functional characteristics of human speech perception.

What can we say about the physiological response of the human auditory system to a speech stimulus? How might this be related to the psychology of what listeners hear for such stimuli? These are the kinds of issues we intend to introduce in this chapter.

17.2 VOWEL PERCEPTION: PSYCHOACOUSTICS AND PHYSIOLOGY

In Chapter 11, we saw that vowel articulation depended greatly on resonances (formants) of the vocal tract configuration. What is the relation between formants and the perception ...