The more I work with electronic instruments, the more I appreciate acoustic ones-the human voice, in particular. The voice is not only the oldest musical instrument, but the most complex, flexible, and-thanks in part to its connection with language-expressive. What other instrument, whether acoustic or electronic, has such a wide range of timbre and emotion?
By excluding other instruments from the mix, a cappella music highlights the expressiveness of the voice and multiplies it by the number of singers in the group. The results can be greatly satisfying musically-in terms of soul and technique-not to mention a welcome change from the technology-driven soundscapes that electronic musicians often inhabit.
I recently recorded Wicki 6, a sextet from Berkeley, California, and I was reminded of the musical richness, range, and potential of a cappella. But I was also reminded of the many critical considerations and technical challenges that await the engineer attempting to turn out good recordings of vocal groups.
SCRATCH HERERecording a cappella music may seem low-tech, but there's frequently more to it than meets the ear. Scratch tracks are one example. Because an a cappella group has no band to follow, laying down some reference tracks prior to recording usually helps the group stay in time and in tune.
First, determine the tempo of the song and devise a click track. This can be a straight click, a drum pattern, or whatever works best for the group. However, avoid high-pitched or overly bright sounds, such as claves or cowbells, because these are apt to leak through the headphones and onto the vocal tracks. It may also help to EQ the track, cutting the highs and boosting the low mids. Even then, it's important to record a bit of the song and listen back to make sure you don't hear leakage from the phones on any of the tracks.
Next, depending on the group and the complexity of the song, you might want to put down some guide chords. These can go at the top of each bar or musical phrase, especially ones leading into difficult chord changes or modulations. A simple bass line might be helpful, too. Use straightforward sounds for these scratch tracks (a piano patch is usually best), and be sure the parts are in time with the click.
Another point: make your reference tracks long enough to reach the end of the song. If they extend beyond the end, though, be ready to fade (or mute) the reference tracks after the last note is sung to prevent sound from bleeding from the headphones.
I usually create the reference tracks on my sequencer (which is synched to my MDMs), both to avoid using up tape tracks and to readily accommodate last-minute changes. Sometimes a group may want to alter the tempo slightly-or even change keys-and it's much easier to make those changes on the sequencer than to rerecord the guide tracks.
For complex, underrehearsed, or long songs, a spoken guide track can save the day. This could include a count-off, location points (for example, "Verse two," "Bridge," or "Chorus"), and even explicit directions for the singers ("Here come the extra two beats!" and so on).
Of course, for a tight, well-rehearsed group, or for one making a quick demo only, reference tracks may not be necessary. You may even find that a click track inhibits the performance, especially if the group isn't used to working with one. In that case, you should dump it. (Obviously, a click is also ill advised if the song contains tempo changes, rubato parts, or other timing anomalies.)
GROUP DYNAMICIt's possible, of course, to make a solo a cappella record, or for a lone singer to create a group sound via overdubs-a technique Bobby McFerrin has used to great effect on several albums. This article is about recording a cappella groups, the members of which usually sing together. Nevertheless, you could approach the recording as you might a band project: recording the rhythm tracks first and overdubbing the solos, the advantage being that the soloists get to keep redoing their parts until they're happy with them.
I find, though, that a cappella groups are usually well rehearsed-much more so than the average rock band-and tend to prefer a "straight, no chaser" approach to recording. Like jazz groups, they rehearse as a unit, with members feeding off the group energy. In such cases, it makes sense to record the group with everyone singing at once, just as you would a concert performance. That's not to espouse a purist approach, nor to suggest that you forego any "studio magic." You might choose, for example, to thicken the sound by doubling some (or all) of the parts. But for me, the joy is in capturing the communication that can happen only when the singers are doing their thing live.
There are three basic miking techniques for recording an ensemble vocal performance: mono, stereo, and multitrack (miking the singers separately). Each approach has its advantages and disadvantages, and there are variations of each. They sound different, too.
ONE-EARED JACKIt may seem primitive, but if done well, mono recording can sound great for a cappella groups. It has tactical advantages, too. First, it requires only one mic and one track (though a nice trick is to double the performance on two separate tracks-leaving any solo parts off of one of the takes-and then pan the tracks apart). Second, it allows the singers to perform without headphones (which not only frees them up but could save you considerable expense). And third, it allows for full visual contact among the singers (see Fig. 1), encouraging a tight performance.
For this application, use a large-diaphragm, multipattern condenser mic set to omnidirectional mode. Place the singers at various distances from the mic to "mix" their individual levels, with the lead singer(s) closer to the mic and the booming voices a bit further back. If the singers are wearing headphones, they can move around and determine their own levels. Otherwise, you have to position them.
Omni patterns typically provide accurate and open sound. Provided the room has favorable acoustics, you can make very good recordings this way. You control the amount of ambient sound by moving the singers closer to the mic to reduce ambience or having them step back to increase it. Because the omni pattern is free of proximity effect, you can bring the singers in quite close to the mic without unwanted bass boosting.
If unfavorable room acoustics are spoiling the sound even with the singers close to the mic, try using the figure-8 pattern. This mode has excellent side rejection and relatively small pickup areas, resulting in a tighter, drier sound than omni mode. However, it's hard to fit more than four people (two on each side) into the picture comfortably. As you attempt to add more, off-axis coloration starts to compromise the sound of those not positioned squarely in the pickup areas.
TWO EARS, TWO MICSThe big advantage of stereo recording is that it usually provides the most natural sound-not surprising, considering that we hear with two ears. Also, it takes only two mics, and again, headphones are not required. One disadvantage, though, is that the singers have to face the mics rather than each other, making it difficult for them to see one another. And if they're not wearing headphones, they may have trouble hearing each other, as well.
I'll discuss three stereo-setup options: XY coincident, XY near-coincident, and spaced pair. A matched pair of cardioid condenser mics works best for the first two setups. Cardioids also work for the spaced-pair technique, but I've achieved killer results using small-diaphragm omni condensers as a spaced pair- depending on the suitability of the room acoustics. For all three techniques, stereo sound reproduction is maintained by assigning the two mics to different tracks, which are panned hard left and right during playback and mixdown.
For the XY-coincident pair, position the mics at approximately head height with the capsules at a 90-degree angle to one another (see Fig. 2). Then arrange the singers in a semicircle facing the stereo pair, with the lead singer (or singers) in the center and closest to the mics. Done correctly, this technique captures a good stereo spread with precise placement of voices on the soundstage.
For a broader stereo image, try the XY near-coincident technique (see photo on p. 100). However, near- coincident mic positioning can leave a slight "hole" in the center of the soundstage, so don't use it if you want to create a strong center image, say for a single lead voice.
A spaced pair broadens the stereo image further, creating an even bigger hole in the center, especially with cardioid mics. It also increases the likelihood of phase problems, so when using this technique, check the signals for mono compatibility by panning them both to the same spot.
Despite these slight drawbacks, a spaced pair positioned just right can create a desirable sense of spaciousness. This makes it a cool choice for a doubled part, to add dimension to the mix. For instance, you could record the first pass (including the lead voice) with an XY-coincident pair and then use the spaced pair to capture the ensemble doing the same part again, sans lead. Then you would pan all parts hard left and right, and presto: full soundstage-and a nice chorus effect, too.
TO EACH HIS OWNRecording each singer to a separate track provides by far the most control during mixdown. However, depending on the size of the group, the gear demands may exceed what's available in the personal studio. Each singer requires not only a mic but also a pair of headphones, and probably a headphone extension cord, too. You also need a junction box or headphone distribution amp with enough patches to feed all the phones. Finally, you need enough tracks.
Assuming that you have the gear (and that it's good stuff), the keys to success are mic selection and positioning. Every singer is unique, of course, and there are scores of mics, so it would be unreasonable to make specific model recommendations here. Hopefully, you can make educated guesses based on familiarity with your mic arsenal. From there, it's a matter of trial-and-error until you hit upon the most flattering combinations. (This process can be lengthy for a big group, so you may want to book the testing and tracking sessions for separate days.)
Generally, I use a condenser mic (cardioid, supercardioid, or hypercardioid) on each singer, except sometimes for the bass singer (if I'm going for a particular sound) and the vocal percussionist, if there is one-especially if he or she has worked up a style that includes lots of movement and inadvertent spitting. But mic selection depends chiefly on the type of sound you're going for and how well the mic and voice work together to provide it. Certainly, you can make beautiful-sounding a cappella recordings using dynamic mics alone. What matters most, as with any recording, is the performance.
After determining which mics go to whom, the tracking process is the same as for an individual singer. A pop screen on every mic is a good idea, but if you don't have enough to go around, at least make sure the lead singer has one, as well as the percussionist (that is, if you are using a condenser mic) and anyone else generating excessive plosives. (For more in-depth information about recording vocals, see "Recording Musician: Keeper Vocal Tracks" in the January 1999 EM.)
You should have no problem recording all the singers at once in the same room, especially one with good acoustics. Just be sure to use the rear rejection afforded by the mics' polar patterns to create as much isolation between signals as possible. This is usually best done by positioning the singers in a circle (see Fig. 3), which also lets them see each other. Some bleed is inevitable, but, kept to a minimum, it shouldn't harm the final mix. If anything, it will make for a more natural-sounding blend.
Some dynamics processing may be called for, especially on the lead voice, which is more likely to span a broad dynamic range. Opt for a transparent compressor (one known not to color the signal much; probably a tubeless, VCA type), and use the lightest setting you can get away with-say, at a ratio somewhere between 1.5:1 and 4:1 and a threshold around -5 dB, depending on the part. The point is to control unruly dynamics while leaving as much dynamic range as possible, to maintain a natural sound. (Compression may also be helpful for the mono and stereo recording applications detailed earlier.)
Although recording each singer to a separate track affords extensive control in the final mix, don't forget that a cappella music is generally best au naturel. Don't go crazy equalizing and processing the tracks; use the control to unify and enhance the overall sound, not to revise it. Assuming the parts are well recorded in a decent-sounding room, it should take no more than a few EQ tweaks, a touch of compression, and a bit of reverb to make the final mix shine.
Of course, the naturalist approach doesn't always apply. The voice may be the oldest musical instrument, but some a cappella groups today make pretty wild sounds, especially when the voice is used to simulate other instruments. These signals are prime contenders for more extensive tracking or processing techniques.
SHH BOOM BOOM BAPVocal percussion is fairly commonplace in a cappella these days, but some groups take the concept to new heights. Check out the Latin group Vocal Sampling, for instance, which makes the sounds of the various percussion instruments typically heard in a salsa band-including guiro, shekere, cowbell, and timbales.
No matter how bizarre the vocal simulation, though, the basic trick is to treat it as you would the actual instrument. For example, for a kick/snare/hats-type groove, try recording the performance in a brighter- sounding, reflective space (a tiled bathroom or stairwell, for example), possibly using additional room mics, to create some ambient attitude. Then, in the mix, pump up the kick with some boosts between 60 and 150 Hz, the "crack" of the snare at 3 to 5 kHz, and possibly the hats with some high-frequency shelving. Finally, to add punch and clarity to the rhythm, squeeze the performance fairly hard with a compressor, say at a ratio between 5:1 and 10:1 with a -10 dB threshold.
THE LOWDOWNThe bass track may also cry out for processing. Say you want an electric bass sound. First, record the voice through a tube preamp to fatten the signal. Next, spend some time with a tweakable compressor exploring different attack and release settings until you find those most complementary to the groove or bass style that you have in mind. Slower settings and moderate ratios (between 2:1 and 4:1) typically work better for slower, ballad-type lines, while faster ones with higher compression ratios suit funkier, popping lines.
Next, equalize the signal to taste, perhaps patching in a sonic enhancer to clarify (or cut) highs and to broaden or tighten the low end. Last, dial in the effects common to the bass style you have in mind-for example, a chorus and big reverb for a Jaco-type fretless sound.
TAKE THE LEADOne of my favorite vocal simulations is Bobby McFerrin's electric-guitar solo on "Sunshine of Your Love" from the album Simple Pleasures (EMI Manhattan Records, 1988). To simulate electric guitar, start with a cheap microphone-one of those Radio Shack models that's designed to plug into a boombox can work wonders, or else try a mic typically used for blues harp, such as a Shure Green Bullet or Astatic JT 30.
Next, run the signal through a guitar rig-for example, a Marshall tube amp with a distortion pedal or wah wah, or whatever is appropriate. Mic the cabinet with a Shure SM57 and compress the heck out of the signal. Then tweak it some more in the final mix, perhaps rotary panning it to create movement. As long as the singer gives you a good guitar simulation to start with, the final effect should be hella convincing.
Brian Knave is an associate editor at EM.