Vocoder

A vocoder is a device for analyzing and synthesizing human voice. The term itself comes from the words 'voice' and 'encoder'. It has many uses but its main function is in secure voice transmission. The vocoder is useful for voice encryption which ensures that only another person with the voice decoder can listen to what has been said. Thus, a vocoder is basically a voice encryption or encoding device.

Interestingly, the vocoder has been around since 1928 (and has been widely applied in the music industry around the '70s). While technology has done much to improve the device, its basic principles and operation remain as they were almost a century ago.

How a Vocoder Works

The key element to remember about the way the vocoder operates is the fact that the human voice, as it is perceived, has varying harmonic content. Human speech starts as sounds (in a periodic waveform) produced in the vocal cords which pass through the nose and throat. In effect, the original sound is filtered through a complex resonant system before it leaves the mouth.

The filtering process has an impact on human speech. The process of resonation enables humans to create a wide variety of sounds which ultimately comprise human speech. Humans are also capable of forming voiceless sounds (those sounds produced with non-vibrating vocal cords) and plosive sounds (those that are produced by manipulating the flow of air while speaking).

The vocoder records a person's speech. The program pinpoints the base frequency of the voice and then proceeds to measure the variation in harmonics. These measurements are coded in numbers that represent a certain level of frequency. These are then recorded prior to transmission.

The result of the process is a dramatic reduction in the amount of bandwidth needed to store or transmit the words. Rather than a complete recording, what you have is an identified carrier wave and a series of numbers corresponding to the harmonic frequencies which is far, far smaller than the original speech. This process is somewhat comparable to what happens to WAV and MP3 files – one is the 'original' speech while the latter is the 'vocoded' version.

A musical vocoder operates in much the same way, except that the musician or technician can set the frequency of the carrier wave (such as a sound range using a musical instrument) while the characteristics of a speech or song (especially the key sound range or frequencies) are identified, split up and later on imprinted on the base wave. The musical vocoder therefore adapts a singer's voice onto a musical carrier wave.

Voice Decoding

Recreating the original speech is a matter of reversing the process: an oscillator is used to create the primary wave, which is then passed through a stage where the other variations are incorporated based on the originally identified numbers. The result is, at the very least, a robotic-sounding voice since the vocoder tends to wash away some of the harmonics that communicate emotions and other sensations in normal speech.