Sound Design Workshop: Vocoding With Nonsense Syllables

Image placeholder title
Image placeholder title

FIG. 1: Antares Articulator is ideal for vocoding diverse material with scat vocal snippets. Use the formant controls in the middle to tweak the vocal analysis.

One- and two-syllable snips from scat vocals make good fodder for the modulator input of a vocoder. These nonsense syllables produce talk-box-like effects, and sequencing or arpeggiating them gives you rhythmic control. For the source audio, you can use material as diverse as the total mix, various submixes, and individual rhythmic and ambient stems—almost anything works as long as you tailor the modulator sequence to the material.

The first step is to extract an assortment of scat vocal clips and fashion a sampler instrument or drum machine from them. Choose a cross-section of lengths, pitch ranges, and phonemes. For my examples, I''ve used vocals from Montreal singer Julie Hamelin, available in Julie''s Scat Collection ($9.99) from Les Productions Zvon.

The Setup
How you configure your system depends on your vocoder and DAW. In particular, DAWs have different ways of routing audio to a plug-in''s sidechain input, which for some vocoders feeds the modulator, and for others the source. If possible, you''ll want both vocoder inputs to be routed pre-fader. That lets the level of the source audio in the mix be independent of the vocoder feed. It also lets you audition the scat clips (the modulator signal) while setting up but leave them out of the final mix, where they''ll almost never fit.

The most important feature of the vocoder, aside from having an audio input for the carrier (not just an internal synthesizer), is high resolution: Use the maximum number of bands and use FFT mode, if available. The BV512 vocoder in Propellerhead Reason, Ableton Live''s Vocoder plug-in, and Antares Articulator in the Avox Evo bundle all fit the bill. Articulator is the most talk-box-like, dispensing with band controls and incorporating Antares throat-modeling technology (see Fig. 1). I''ve used that in my examples.

You can trigger the sampled scat clips in real time or from sequences. Playing a single note or chord produces a one-shot sound effect, whereas looping a sequence or playing into an arpeggiator gives you a rhythmic pattern (see Web Clips 1 and 2). To create rhythms, I use an arpeggiator and make the scat clips audible to audition patterns, then I create MIDI sequences from the patterns that work—step sequencing is a viable alternative. Depending on the tempo and material, I use arpeggiator rates of quarter-, quarter-triplet- or eighth-notes; play from two to four notes; and set the arpeggiator''s pattern to the order played. The play order can make a huge difference in the outcome (see Web Clip 3).

In Practice
One of my favorite uses for scat vocoding is to thin busy drum and percussion mixes. That''s especially useful when you''re working with limited resources from a construction kit or remix. The scat sounds determine which percussion frequencies are emphasized and which are suppressed, and the placement and spacing of the trigger notes influence how busy the track is (see Web Clip 4).

For sustained sounds, such as pads and ambient washes, put a stereo delay after the vocoder with different and longish delay times for the right and left channels. Use a single-syllable scat sound with a strong attack and trigger it only once per measure, or less. That can add some subtle motion to the pad without interfering with the rhythm tracks (see Web Clip 5).

When the effect is applied to a full mix, you''ll probably want to keep it in the background. Using a less-active scat sequence with mostly short syllables makes a big difference. I like to follow the vocoder with a long-tailed reverb set to 100-percent wet and to automate the reverb level as suits the piece (see Web Clip 6).

All these examples use repeating scat patterns, but one-shots are also handy for adding accents derived from the material already in the song. That''s a good way to repurpose sound effects (see Web Clip 7). In either case, experiment with audio effects plug-ins after the vocoder.

Len Sasso is an associate editor of EM. For an earful, visit his website,