Universiteit Utrechtstudierichting Fonetiek

Main collection



Below are some links to demonstrations of the vocoder in pop music. The vocoder is an instrument which splits the input speech into source and filter parameters. In "normal" usage, the original speech signal is reconstructed by generating an artificial source signal (pulse train or white noise, for voiced and voiceless speech sounds, respectively). This source signal is then led through a filter bank which reconstructs the original filter features.

In pop music, the source signal is replaced by a different signal, from a musical instrument. The filter features are then assigned to the instrument signal. The resulting effect is that of a "talking" instrument.
The source signal usually comes from an electronic synthesiser. This source signal is electronically generated (like e.g. a pulse train), but its spectrum differs considerably from that of a pulse or voicing signal. Filter characteristics are usually derived from speech (microphone signal), but it can also be derived from other instruments.

Several readers have pointed out that the audio demonstrations on this page were not vocoded at all. We apologize for the confusion, and refrain from listing demonstrations here before we've improved our homework for this page. Meanwhile, you can browse the hyperlinks below for true vocoder demonstrations.

More information about vocoders:

listen to demonstration listen
technical details technical
text/transcript text/transcript
naar top

Nederlands find home UU home Letteren home Fonetiek e-mail
26 Feb 2003