ABSTRACT
Speech sounds are generally classified broadly as vowels and consonants. Vowels are voiced sounds whose spectral characteristics are determined by the size and shape of the vocal tract. Changing the shape of the vocal tract changes its filtering characteristics, which in turn changes the formant structure, that is, the frequencies at which the speech signal is enhanced or de-emphasized. Categorical perception has been explored in many studies using a variety of speech and non-speech materials. Categorical perception was originally observed for speech sounds, but not for non-speech stimuli. Duplex perception refers to hearing separate speech and non-speech sounds when the listener is presented with certain kinds of stimuli. The perception of sine wave speech has often been associated with a specialized speech mode because it reveals how the same signal can be experienced as either speech or non-speech. Several points about the power of speech sounds are noteworthy prior to a discussion of speech intelligibility.
