It's been said that the human voice is the most important source for recording or amplification. However, even when vocal tracks are beautifully recorded with excellent equipment, there are still lots of reasons to process them.
Vocal processing works on the basic properties of level, timing, timbre, and pitch, but also includes reverberation, spatial placement or movement, and just plain mangling. Most vocal-processing tools can be used either correctively (to fix a problem) or creatively (to achieve an effect).
Loud, Louder, Loudest
FIG. 1: Classic de-essing is accomplished through frequency-sensitive compression, which is achieved by inserting equalization into the compressor''s sidechain. Credit: Chuck Dahmer
Level control comes in two forms: dynamics processing (compression, limiting, expansion, and gating) and automation. The most common vocal-processing task of all is probably compression. Distant-miked vocals, such as a choir or chorus, may not need compression, but close-miked vocals, whether in performance or recording, almost always need some.
Compression is often done in the analog domain, as part of a live or recorded signal path. Being able to get good results from a compressor in a vocal recording session is a key skill in vocal processing. Some applications call for a compressor that is reasonably transparent sounding, while others need the sound of a compressor colorfully squashing a vocal. A compressor working on a vocal can suggest power, as it sounds like the singer is really belting it out. A good rule of thumb when recording vocals is to shoot for 2 to 6 dB of gain reduction on a regular basis.
Stereo compression is a great tool for vocal groups, such as harmony vocals, as it helps to attain a smooth, even blend. Compression or limiting can also be useful following heavy effects processing, such as intense flanging, to keep peaks under control. In fact, it is common for vocal chains to use both compression and peak limiting.
Automation is an extremely powerful vocal-processing technique, whether it is created with a fader or by using breakpoint envelopes to do surgical dips and boosts. Automation is also good for balancing multiple vocal tracks, EQ or effects automation, and more.
De-essing, usually achieved through frequency-sensitive compression, is a crucial vocal-processing task. Close-miking most often creates the need for de-essing, as the full high-frequency pickup that occurs close to the mouth combines with the rising high end of the condenser mics used in most vocal recording to create sibilance. The compressor is sensitized to sibilance by using an equalizer set to emphasize sibilance frequencies in the sidechain (see Fig. 1). (Dynamics processors have two signal paths: the audio path and the sidechain, which derives the signal for controlling the gain from an audio input.) Multiband compression can be even more effective for de-essing when configured correctly, and EQ automation is another excellent de-essing method when the problem occurs only in isolated spots.
Time for a Change
FIG. 2: Hard-panned micro-pitch-shifts are an effective way of widening the image of a vocal track. Using delays of a few milliseconds is another alternative. The two techniques can even be used in combination to widen the image further. Credit: Chuck Dahmer
In the time domain, doubling and chorusing are two very common vocal effects. Most doubling or chorusing processors (software or hardware) have presets for vocals in which the effect is fairly intense and pitch movement is accomplished with a sine or triangle LFO. For a subtler effect, don't be afraid to use a very low wet/dry mix, even as little as 6 or 7 percent wet signal. For a more natural and interesting effect, use a modular synthesis environment like Cycling '74 Max/MSP or Native Instruments Reaktor to construct your own chorus effect that uses slow random modulation instead of sine or triangle waveshapes.
Of course, DAW editing makes you the master of time with your vocal tracks. Fly choruses around, nudge one line, or fix a spot that has rough rhythm. Beware, however: it is easy to overprocess vocals until they no longer sound natural, and that applies in particular to moving things around in time. It can be a great effect but also a slippery slope.
Time compression and expansion have become popular vocal-processing tools, as their audio quality has gotten quite good and the interface for invoking them simple. Today's time-stretching can even be used with confidence on a lead vocal. A good example of an application for time-stretching is when harmony vocalists end a note at slightly different times.
Tone to the Bone
Timbre covers a lot of ground; it's essentially everything that's not level, timing, or pitch. The most fundamental timbral tool is the equalizer. From a corrective standpoint, EQ is often used to deal with problems resulting from vocal quality, equipment (mic, preamp, or compressor) characteristics, and the interaction among them. Reducing boominess, boxiness, sharpness, dullness, or sizzle are all common EQ uses. EQ can also be used for effects, such as the ever-popular telephone filter. EQ is often vital to giving a vocal its own place in a mix, including among other vocals.
Traditional graphic and parametric equalizers are still found in abundance, but there is also a new generation of equalization plug-ins with uncommon approaches to EQ. Among the traditional tools, parametric EQ is the more difficult and, by far, most useful choice, offering the capability both to fix many problems and to shape the sound. Really learning how to use parametric EQ on vocal tracks gives you awesome control over their sound.
Here's the Pitch
Last but not least is pitch correction. Programs such as Antares AutoTune have become very popular for a variety of pitch-related uses. But although AutoTune is an outstanding tool, it is often used as a substitute for singing ability. Pitch correction is best used sparingly, for individual problem notes. If you have to use it heavily, what you really have is a performance problem (though pitch correction might just save your bacon in that situation).
Tuning aside, a classic pitch-shifting application is to widen the image of a single voice by hard left and hard right panning two versions of a vocal with very small pitch-shifts above and below the original pitch, while the original stays centered (see Fig. 2).
In the last several years, formant-based pitch-shifting has become readily available, opening a number of new doors, from “gender reassignment surgery” (making a male voice sound female, or vice versa, by shifting pitch and formants) to making corrective pitch-shifts less audible by also shifting the formants.
Wrap Your Dreams in Echo
When all other processing is done, you will usually want to put some sort of space around the vocal with echo and/or reverb. In thick mixes, you may find that using simple echo on vocals is sufficient and that reverb may actually smear the vocals into the mix to the point that they are no longer distinct. On the other hand, in a sparse setting, a lush reverb can be gorgeous.
Watch out for excessive boominess or sibilance in the vocal, which gets emphasized by reverb. An EQ in the reverb-send path can be most beneficial for protecting against these. Boominess can also be reduced by making the low-frequency decay time shorter than the high-frequency decay. Though this is contrary to almost every natural acoustic situation, it works well in pop music.
Mutants and Space
Spatial placement for vocals usually means panning the lead to center and backgrounds somewhat off center, but this is often combined with other techniques, like putting different processing on the backing vocals than on the lead. Because the vocal is typically the primary focus of a song, it is unusual to subject it to dynamic spatial movement, but on occasion, flying a vocal around can be a very dramatic effect.
There is a tradition, dating back at least to the Beatles, of using transformative processing on vocals. The Beatles introduced processing vocals through a Leslie cabinet, but the use of vocoding, talkboxes, or flanging (among many other devices and processes) on a vocal can immediately cast an entire mix in a different light. Taste aside, the challenge of using radical processing on a voice is retaining intelligibility. Mixing some amount of unprocessed vocal back in can help when a vocal gets so twisted by processing that it can't be understood. It's worth pointing out that vocal recording and live performance commonly call for real-time processing, frequently provided by dedicated outboard processors. Compression and equalization, especially, are often best performed by a good-quality analog compressor or EQ following the mic preamp.
Though creativity always offers additional options, mastering a relatively small set of vocal-processing skills, such as compression, de-essing, EQ'ing, and adding ambience, can be the third best way to get great-sounding vocals — the first two ways being having a good singer and having a great mic-and-preamp combination.
Larry the O is a contributing editor and has been writing for EM for more than 20 years.