Stimuli used in research on the perception of the speech signal have often been obtained from simple filtering and distortion of the speech waveform, sometimes accompanied by noise. However, for more complex stimulus generation, the parameters of speech can be manipulated, after analysis and before synthesis, using various types of algorithms to generate stimuli of interest to the researcher. There are several computer-based analysis and synthesis techniques, or vocoders, that make different assumptions about the speech production and reception mechanisms. These include: (1) the channel vocoder, which uses filter banks to split the speech spectrum into discrete frequency ranges measured often enough to keep up with speaker's articulation rate; (2) the homomorphic vocoder, which derives source and vocal tract parameters from the spectrum signal alone; (3) the linear predictive coding vocoder, which generates the speech signal in a manner similar to the two previous vocoders, but measures the vocal tract by fitting the speech waveform in some interval to the output waveform of a linear filter; and (4) the sinusoidal transform coding vocoder, which, rather than treating the sound source as a periodic pulse train or noise signal, characterizes it as a sum of sinusoidal waves that vary with respect to one another. (MSE)