A vocoder is an audio processor that captures the characteristic elements of an an audio signal and then uses this characteristic signal to affect other audio signals. The technology behind the vocoder effect was initially used in attempts to synthesize speech. The effect called vocoding can be recognized on records as a "talking synthesizer", made popular by artists such as Stevie Wonder. The basic component extracted during the vocoder analysis is called the formant. The formant describes the fundamental frequency of a sound and its associated noise components.
The vocoder works like this: The input signal (your voice saying "Hello, my name is Fred") is fed into the vocoder's input. This audio signal is sent through a series of parallel signal filters that create a signature of the input signal, based on the frequency content and level of the frequency components. The signal to be processed (a synthesized string sound, for example) is fed into another input on the vocoder. The filter signature created above during the analysis of your voice is used to filter the synthesized sound. The audio output of the vocoder contains the synthesized sound modulated by the filter created by your voice. You hear a synthesized sound that pulses to the tempo of your voice input with the tonal characteristics of your voice added to it.