We all know the sound: the robot voice that seems synthesized but is singing real words. Kraftwerk’s “The Robots.”  “Mr. Roboto” by Styx. Daft Punk’s “Harder, Better, Faster, Stronger.” “E=MC2” by the late, great J. Dilla. Herbie Hancock making ’80s kids get their robot moves on with “Rockit.” These and countless other classic hits owe their sonic identity to the vocoder, a unique and often misunderstood piece of music technology.


Vocoder custom designed for Kraftwerk in the early 1970s. Image public domain.

In the simplest terms, a vocoder superimposes your words onto another signal such as a musical instrument (often a synthesizer). Getting more technical, it uses a number of bandpass filters to analyze and break down your voice into separate frequency bands. The changes going on in each of those bands are then applied to the instrument sound through another group of filters. Your voice is the modulator and the instrument is the carrier — and both need to be present. 

Some vocoders included a built-in carrier signal generator or even a full keyboard instrument, while others assumed you would have a synth on hand.

There’s no rule that says the modulator signal has to be a voice. Material such as a drum track can turn a drone or pad into a rhythm loop with a sonic character all its own.




Peter Frampton’s talkbox. Creative Commons license via Wikimedia Commons.

One of the most recognizable robot voice sounds in music comes from the talkbox. Made famous on guitarist Peter Frampton’s “Do You Feel Like I Do“ and Roger Troutman’s “More Bounce to the Ounce” with Zapp, it was basically a small powered loudspeaker inside a pedal. A plastic tube ran from the speaker to your mouth, and the words you mouthed as you played your instrument literally shaped the sound. In an interview in Keyboard magazine, P-Thugg of neo-funk duo Chromeo once described the sensation as “like having a speaker cabinet in your mouth.” The talkbox achieves the same musical goal as a vocoder in a way that’s more analog than analog but, strictly speaking, vocoder refers to a device that does this electronically. Companies like Rocktron sell new talkboxes today, and Antares’ plug-in Articulator is designed to create the talkbox sound.


Schematic drawing by Homer Dudley from his article “The Carrier Nature of Speech,” Bell System Technical Journal, October 1940. Creative Commons license.

Homer Dudley, an engineer working for Bell Labs, was working on a way to send the human voice over copper telephone lines while conserving energy and bandwidth (and thus cost). He had the genius idea to break up human speech into frequency bands, transmit only the bands needed to make it understandable, and reconstruct everything on the listening end. 

Dudley showed off the proof-of-concept at the 1939 World’s Fair in New York City. Called the Voder (Voice Operating Demonstrator), it didn’t use a microphone for the modulator signal. Instead, a skilled operator worked a series of keys, a wrist bar, and a pedal to turn its built-in carrier signals into vowels and consonants and change the pitch. The results were crude, but crowds marveled at a machine creating intelligible human speech and singing.



SIGSALY station in the NSA Museum. Image credit: National Security Agency (public domain).

The U.S. Army Signal Corps (which would eventually give birth to today’s NSA), adapted Bell Labs’ technology to encrypt sensitive communications. Most famously, it was used for conversations between Franklin D. Roosevelt and Winston Churchill over the trans-Atlantic cable during World War II. The system, called SIGSALY or simply “Project X” (how Tom Clancy is that?) added noise to the vocoded signal using a special record played on a turntable. The station at the receiving end would then play a “decoder” record and filter out the identical noise band for the high-level listener’s ears only. It turns out a pair of turntables isn’t just for DJs!



Siemens synthesizer in the Deutsche Museum, Munich. Creative Commons license.

The first musical machine to incorporate vocoder technology was the Siemens synthesizer, built by the German tech firm now associated with ATMs in the late 1950s.
In 1971, synth designer Bob Moog and composer Wendy Carlos combined various synthesizer modules to improvise the vocoder that can be heard on the soundtrack to “A Clockwork Orange.”

Electronic music legends Kraftwerk are synonymous with early vocoder trailblazing, as on “The Robots” and many other songs.


But if one modern artist stands out for making the vocoder sound a staple of popular music, it’s Afrika Bambaataa and Soulsonic Force with the breakthrough hit “Planet Rock.


This gets controversial because in fact, Bambaataa did not use an actual vocoder but achieved a robotic vocal effect via a Lexicon PCM-41 digital delay. But there is no doubt that more than any other, the song led artists to seek out the same sound and therefore made the vocoder a going concern among musical equipment manufacturers.




Image: EMS.

The EMS 5000 was developed by Peter Zinovieff’s Electronic Music Studios, more noted for the VCS3 synth of “Dark Side of the Moon” fame. In 1975 it was the most comprehensive stand-alone vocoder yet made, with a built-in two-oscillator synthesizer and 22 frequency bands for unprecedented detail. You could even change the gender of the output voice, foreshadowing today’s vocal tract modeling in plug-ins such as AVOX Throat. EMS vocoder line would eventually include the 1000, 2000, and 3000, all of which are highly desirable on the vintage gear market today.


Image: EMS.

Harald Bode’s model 7702 vocoder featured 16 frequency bands. The more bands (filters) a vocoder has, the better it is at making the speech aspect sound clean and intelligible. Bode’s design was so good, it was licensed and rebranded by none other than Bob Moog in 1978.


Image: Roland Corp.

One of the most popular and sought-after models is the Roland VP-330, released in 1979. This is because it wrapped up a vocoder and string/choral ensemble synth into a portable keyboard as suitable for gigging as it was for studio use. French electronic/ambient artist Jean Michel Jarre also raised its profile.


Image: Korg USA.

The tiny and affordable microKORG made its way into seemingly every indie band’s keyboard rig beginning in 2002, and sold so well that it’s still in production two decades later! This virtual analog synth featured a vocoder mode and integrated gooseneck mic, and compact synths such as the Novation SuperNova followed suit.




Several plug-ins have tried to bring the vocoder into the realm of computer-based music production over the past 20 years or so. The latest and most comprehensive of these is Auto-Tune Vocodist from Antares. A lot of companies have made a lot of great vocoders over the years, so Vocodist features a selection of models of 18 coveted classics (including Homer Dudley’s Voder) plus two original algorithms that offer exceptional clarity. There are also artist presets curated by P-Thugg of Chromeo, producer Buddy Ross, Rachel K. Collier, and others. 

Vocodist can work with an external carrier signal such as your favorite synth, but that’s not necessary—an 8-voice, dual-oscillator synth is built in. In addition, you can control the pitch of the vocoded signal either with your voice or by playing a connected MIDI keyboard.

As the name implies, full Auto-Tune functionality is on hand, as are key and scale detection via compatibility with the Auto-Key app. Vocoder tracks may therefore be subtly pitch-corrected, or combined with the classic “Auto-Tune Effect” at more aggressive settings. Additional integrated effects include chorus, tube saturation, and a ring modulator.

Antares offers Auto-Tune Vocodist as part of a subscription to Auto-Tune Unlimited, the ultimate vocal production suite.