Vocoders were very popular in the '70s, but their 'talking keyboard' sound soon became cliched, and from then on, their popularity steadily declined. By the time MIDI started to take off, vocoders were all but extinct, with only a couple of manufacturers continuing to make them -- which was a pity, because a vocoder really comes into its own when used as part of a MIDI system. Fortunately, a few multi-effects units now include a vocoder as part of their repertoire, and with a little ingenuity, they can be used to modify sounds in a number of creative ways -- other than producing the classic 'asthmatic who's swallowed a harmonica' vocal effect. But before exploring some of the processing tricks made possible by this unique device, it's useful to take a look inside to see how it works.
A vocoder enables the tonal character of one sound to be imposed on another, quite different sound; the classic talking keyboard effect is produced by using the changing characteristics of the human voice to shape a sustained synth sound. What really happens is this: the vocal signal, which we shall call the modulator, is analysed by a bank of filters that continually measure the signal envelope in each part of the spectrum in exactly the same way as a spectrum analyser does. The more filters in the bank, the more accurate the analysis.
The signal to be modified, known as the carrier, is also fed to a bank of filters, but this time the level of signal passing through each filter band is modulated according to the output from the spectrum analyser section. In other words, the spectral characteristics of the modulating signal are duplicated in the filter bank processing the carrier. Figure 1 shows a simplified block diagram of what's actually going on. If the modulating signal is continually changing in character, as is true of the human voice, these dynamic changes will be passed onto the carrier, giving the synth sound a recognisable vocal quality. So effective is this process that it is possible to pick out intelligible words, even when none of the original vocal signal is present. And because we're analysing the spectral content and not the absolute pitch of the modulating signal, it doesn't even matter if the words are sung out of tune, or even spoken.
Apart from the obvious spectral variations generated by the vocal chords, human speech also includes 'fricatives' -- short, high-frequency sounds present in syllables such as 'S' and 'T' which are formed in the mouth. If these are separated out from the main vocal signal by means of a high pass filter, they can be added to the output to increase the intelligibility of the sound, and because they don't relate to the musical pitch of the vocal, they can be added to any musical output without compromising the tuning. A simple system for adding fricatives is also shown in Figure 1.
The original vocoders were built using analogue technology and the filter design was very similar to that used in graphic equalisers. The more bands the signal could be split into, the more convincing the vocal articulation. Some of these machines used patch cords to link the analyser outputs to the modulating filter bank, which opened up numerous creative avenues. By crossing over some of the patch cords so that one frequency band in the analyser section controlled a different band in the modulator filter bank, the vocal character imposed on the sound could be completely changed. Regrettably, most currently available vocoders don't offer this facility.
The digital vocoder implementation used in the Boss SE50 is built around a seven-band filter bank, and though this doesn't sound like a lot, the results are surprisingly good. I suspect that the filter frequencies have been specifically chosen to cover the human vocal range, which would mean that they are grouped together to provide coverage of the vital mid-range of the audio spectrum. The techniques discussed here were tried using an SE50.
Before trying out any advanced processing tricks, it helps to get a feel for the vocoder by recreating the cliches. The SE50 works at line level, though there is enough gain available to plug a high-Z mic directly into it. However, it's best to take a mic feed from a mixer direct out or insert send and feed this to the vocoder's modulator input (the right input in the case of the SE50).
Vocoding is a type of subtractive synthesis, so the carrier should ideally be a harmonically rich, sustained sound. At any rate, it is essential that the carrier produces sound in the vocal part of the spectrum, otherwise the vocoding effect won't work properly. When you speak into the mic at the same time as playing a sustained musical note or chord, you should hear the typical vocoder effect where the carrier is modulated in both frequency content and level. If you stop speaking, the output will fall silent, even though you are still playing the chord. Similarly, if you speak when there is no carrier present, you'll get no output; both signals have to be present before you hear anything. Essentially, the vocoder is multiplying the modulator signal level by the carrier level in each of the frequency bands, and if the input levels are constantly changing in level, the result can sound quite lumpy. It may help to produce a smoother result if a compressor is used to hold the vocal level as steady as possible, and in some circumstances it may help to compress the carrier signal too.
The problem with early vocoders was not so much a limitation of the devices themselves, but of what you could feed into them. Both the carrier and the modulator, usually keyboard and voice, had to be generated together in real time, and that often led to inconsistent or unrepeatable results. Now, of course, we have MIDI controlled keyboard and multitrack tape machines, both of which will produce exactly the same sounds over and over again. The most flexible setup for use with a vocoder is a MIDI sequencer synced to tape, but many of the ideas outlined here may be developed using a just a MIDI synth and sampler.
More realistic choir pads: Synth choir pads or samples are fine, but they don't actually say anything do they? They might go ahh, or ooh, but that's hardly the basis for a great lyric. A more human result can be achieved by modulating the sampled vocal using a real voice, and if you don't feel confident about doing this in real time, the vocal part can be put onto tape or even recorded into a sampler. Because of the way in which a vocoder modulates the level of the sound being processed, it is important not to leave any unin-tentional gaps as you catch your breath; in some cases two or more people singing together can help improve the ensemble effect. And it doesn't matter if you're tone deaf, as long as your timing is OK.
If the result is still too lumpy, try adding the vocoded sound to the unprocessed choir sound, and if you have a mixer with phase invert switches, experiment by reversing the phase of just the vocoded signal. This should change the way it interacts with the unprocessed sound and may produce an interesting alternative. Figure 2 shows this patch.
Cheap Morphing: For this trick, you'll need a sampler capable of performing crossfades to provide the modulating signal and either a synth or sampler to use as the carrier. The first step is to create a sample that slowly crossfades between two sounds with radically different characters, for example, an oboe and a human voice. This is used to feed the modulator input, and whatever is fed into the carrier input will then take on these changing characteristics as the samples crossfade. If the sampler is triggered via MIDI at the same time as instrument feeding the carrier input, the result will be quite repeatable. Figure 3 shows how this patch is created.
Using 'Natural' Modulators: when creating textural back-grounds for instrumental ambient music, the vocoder can be used to modify a synth pad sound using naturally found sounds as the modulator input. For example, a looped sample of cocktail party chatter will impart an almost subliminal murmuring quality to a piece of music. Sound effect CDs or tapes are useful sources of inspiration. Try the obvious sounds such as Wind, Rain, Thunder, Surf and so on, but don't neglect the more obscure sounds, as these often produce the most rewarding results. The patching is the same as for Figure 3.
Merging Synth Sounds: This trick works well if you have access to an analogue synth with a MIDI interface. If this is used as the modulator source, any filter sweeps will be transferred to whatever instrument is being used as the carrier. Don't expect the filter sound to be an accurate representation of the original analogue patch -- what you'll get is a merged sound that has characteristics of both instruments. This technique can be used to create a new sound from any two existing ones, but for the best results, the modulator must change in timbre during the evolution of the sound. Again, the patch in Figure 3 is used, with the modulator input being either a synth or a sampler.
Vocoder Delay: This trick is a variation on the straight vocoder theme. Instead of using the original vocal line to modulate a synth, you use the vocal as normal (Yes, I'm afraid that does mean you have to sing in tune!), and at the same time, use a delayed version of the vocal to drive the vocoder. This means you'll get a conventional vocal followed by a vocoded echo 'sung' by the synth pad of your choice. Figure 4 shows how a vocoder may be used with a delayed vocal.
Vocoding Echoes: Vocoders don't have to be used to modify just synth sounds -- any harmonically rich sound can be used as the carrier. Interesting results can be achieved using long reverbs or multiple delays modulated by vocal sounds. For example, taking another angle on the previous idea, you could use the delayed lead vocal to 'imprint' a vocal onto the reverb tail of itself. This would necessitate setting up a very long reverb time, but as it would only be audible when 'speaking', this wouldn't clutter up the mix. Figure 5 shows a suggested setup for achieving this effect.
These few ideas should be enough to demonstrate that there's more to the vocoder than 'Mister Blue Sky'. With a couple of multi-effects units now including vocoders, the price of experimentation has never been lower -- indeed, you can pick up a second-hand SE50 for around £200, around half the price I paid for my first vocoder ten years ago. Now that we have MIDI to ensure a reasonable degree of repeatability and control, the vocoder has become a valid studio processor rather than the temperamental live performance gimmick it started out as. With so few effects coming onto the market that are actually new, doesn't it make sense to explore something that's been here all along, yet has rarely been exploited to anywhere near its full potential?
Never originally intended as a musical device, the vocoder (VOice enCODER) was originally developed in the 1930s by one Homer Dudley at Bell Labs in the US, during research into reduced bandwidth speech transmission over long-distance phone lines. But, as with much early 20th century electronic gadgetry capable of processing or producing sound, the vocoder moved out of the lab and into the electro-acoustic studio. In the days when electronic music meant manipulating raw sound on tape, the vocoder was one of the few real-time processors available. Early examples of vocoder-based effects can be heard in Disney cartoons and feature films; Radiophonic Workshop-composed BBC theme tunes have often featured vocoding, and that's not to mention tracks from the likes of Pink Floyd, ELO, Laurie Anderson and Devo, amongst others.
The vocoder has enjoyed a long-standing, if fickle, flirtation with the music industry; many have been made and forgotten, though some notable units are still available on the second-hand market and rather sought-after. Here are some of them. (Note that 'new' prices come from 10 or 15 year old price lists!)
BARTH Musicoder (16 filters)
Barth's Musicoder was a comprehensive device developed in the late '70s/early '80s from a Bell Labs-like research tool; Mike Oldfield's studio was reported to contain one of these at one time. Barth have since moved into the niche market of broadcast station controllers.
ELECTRO HARMONIX Vocoder (14 filters)
This famed effects pedal company, who also made a compact synth and a couple of small samplers -- one of which evolved into Akai's S612, also produced a19-inch rack vocoder which also included a compressor on the mic input. It's a simple device, and cost around £400 when new.
Vocoder 2000 (16 filters)
Vocoder 3000 (16 filters)
Vocoder 5000 (22 filters)
EMS, manufacturers of the legendary VCS3 and Synthi A synths, also made vocoders: the 5000 is the most visually striking (matching the synths perfectly -- it was even called the Synthi Vocoder in one version), and is the most fully-specified of the range. It also features a comprehensive patching system, enabling the analysing and synthesizing filters to be connected in any order. Incidentally, the 3000 and 2000 are still available new from EMS, the 3000 at £3000, and the 2000 at £995. Contact EMS on 0726 883265.
DVP1 Digital Voice Processor
This unit couples vocoder facilities with harmonisation, pitch shifting and other features aimed at vocalists, all under MIDI control.
This is likely to be the easiest vocoder to find and has the style and feel of Korg's MS10/MS20 semi-patchable mono-synths, with a sloped front panel and a 32-note keyboard. It's simple to use, and can process either mic or electronic instrument inputs.
16 Channel Vocoder
This was an expensive (£4400-ish), lab-spec, music-specific device, and is trés desirable, featuring patching and many facilities that make it easy to use live, in spite of its complexity. Rare.
Along with the Korg VC10, the SVC350 is likely to be one of the most visible vintage vocoders on the market. This simple to use, rugged rack unit did retail for an affordable £500 at one time, though it might cost close to that to obtain a good example now.
VP330 Vocoder Plus
In spite of its preset-like simplicity, this has become something of a classic: it easily produces typical airy vocal vocoder sounds and features a built-in keyboard to boot.
VP70 Voice Processor
Similar to Korg's DVP1; offers vocoder-like facilities, but really scores as a pitch to MIDI and harmonising system.
This £6000 plus device was a truly pro machine, and is unlikely to have made its way into domestic settings. What you can't do with the VSM201 probably isn't worth doing.S
Syntovox 202 (2 filters)
Syntovox 221 (20 filters)
Syntovox 222 (10 filters)
The 202 was designed for stage use, and had a sub-£300 price tag and an unfussy front panel making it very easy to use; the 221 was a £2800 powerhouse (complete with filter patching matrix).