A top-down approach to recording vocals…
By Joe Albano
For songwriters who record and produce their own music, recording the vocals is probably the most important part of the process. Sure, a good arrangement will help, but it’s the vocal that, for better or worse, is going to carry the song. Getting a standout vocal can be the key to a successful track, whether you’re recording yourself singing in the bedroom with a small multitrack, or producing your songs with a professional singer in your state-of-the-art personal studio. This article takes an introductory look at the overall process of cutting vocals, from setup to mixdown. So, let’s get started!
Who’s going to sing?
You have to make this most basic decision before any technical considerations come into play. If you’re a singer/songwriter and the singer in you is at least as competent as the songwriter, then it may be a no-brainer. But if you’re a better songwriter than singer, you might be better off with a hired gun (or at least a good guest vocalist) to present the song at its best. Competition is huge out there, and you get no second chances.
Get some honest, trusted outside opinions when making this choice. Here’s a hint—that’s what producers are for, not family or friends who either don’t know any better or who love you just the way you are!
If you do bring in vocal talent, it’s more important to get the right singer for each song than the one who can impress with the best name-dropping. The recording process will be largely the same regardless of who handles the vocal chores, though, of course, there will have to be certain adaptations made when it’s a one-man or one-woman show.
Choosing the recording space
For the cleanest recordings, the best option is to have the singer in a dedicated space, isolated from both outside noise and the noise of other people and equipment in the studio. In any commercial facility this is a given, but even in home studios a separate room for the vocalist is often possible. Besides isolation, the other main consideration is the acoustic character of the room itself. For singing, a space with a little bit of ambience can be nice—however, watch out for excessive coloration, which can be common and problematic in smaller spaces.
If the singer has to perform in the control room area for logistical reasons, it’s still possible to get good recordings with a little attention paid to the inherent problems. If necessary, isolate any computer gear or amps (to prevent the mics from picking up the whining of fans) and turn off air-conditioners during the recording. You can create a semi-isolated area within the control room by hanging blankets or arranging absorbent barriers (gobos) around the singer. Finally, make sure to use a directional mic (see below), pointed away from any noise sources. A little ingenuity can often turn a problem room into a workable environment.
Closed-ear (isolating) phones for the singer are a must, to prevent leakage into the mic. The volume balance between the track playback and the live voice in the phones is important. Too much track may cause the singer to strain, or not be able to use their full dynamic range—too little track may result in the singer holding back or going out of time or tune. Experienced recordists might manipulate the monitor mix a bit to coax a certain performance from a singer, like pushing the track deliberately to encourage a more aggressive performance—the trick is knowing how far to go.
Some singers (even good ones), if they’re not used to recording with headphones, may go off-pitch a little with the phones on. A common technique to deal with this (that some vocalists prefer anyway) is to use only one earcup, so they can hear more of their own voice acoustically. But the open earcup will leak into the mic—make sure to kill the monitor feed to that side!
Finally, make sure there’s suitable ambience in the phones—too dry can often be uninspiring for the singer, but too wet can also throw pitch off. Find a happy medium, just for monitoring during recording and playback, but keep the reverb out of the recorded signal. You’ll add “keeper” reverb later, during the mix (maybe some of the same program, maybe not).
The topic of microphones could take up several articles!
The mics you’ll be using for vocal recording fall into three main categories of physical design:
• Dynamic mics are rugged and inexpensive (a diaphragm converts sound waves electro-magnetically to voltage, like a small speaker in reverse), which makes them excellent for live or hand-held use. The most widely-known dynamic is probably the Shure SM58, ubiquitous in clubs and on stages everywhere. The drawback of using dynamics for recording vocals is that their sound quality, while good, is usually not quite as clear and open as the other two types which are preferred for recording.
• Condenser mics use a polarized diaphragm, a design that requires “phantom” power (usually 48V delivered along the mic cable from the mic preamp). This type of mic tends to give a very open and detailed sound, and condensers are the primary choice for professional and semi-pro recording for most applications. Some designs use a tube inside the mic for amplification, which imparts a warm, rich quality—vintage tube mics have become so popular that most companies have reissues and current tube models available.
• Ribbon mics (so named because they use a fragile ribbon element to pick up sound) are known for having a sweet, delicate quality, and are often favored for some vocalists (examples range from classic RCA designs to modern offerings from companies like AEA, Beyerdynamic, Coles, and Royer Labs).
Choose your weapon
There are many models of condenser mics, ranging from under $200 to several thousands for top models like the Neumann U87 or AKG C12. Matching the mic to the singer and song is more important than lusting after an unaffordable mic. As regular readers of Recording will know from the many review articles printed over the years, many budget- and midrange mics can deliver very good results.
It’s always worth spending a little time to try different options, especially if a favorite mic turns out to sound less than ideal for a certain song. For example, one vocalist I recorded a while back generally sounded best with a U87, but on one tune, her voice on the choruses had too much edge—the 87’s high-frequency presence peak imparts great clarity in most cases, but was not the best choice for that particular vocalist on that particular song. As it turned out, in this case a warmer, smoother vocal sound was achieved with a budget condenser—a mic that cost about a tenth the price of the 87, but just right for that particular voice on that particular tune.
Microphones offer one or more of several directional patterns (pickup characteristics, also called polar patterns). Omnidirectional mics (omnis for short) pick up sound equally from every direction (front, back, sides), while Bidirectional (Figure-8) mics pick up sound more or less equally from the front and back and reject sound from both sides (think two singers facing each other). For vocal recording, some type of Cardioid pattern is the most common choice—these mics pick up sound from a wide area in the front and reject sound coming from the rear or sides (Hypercardioid and Supercardioid mics narrow the angle of pickup).
Keep in mind, though, that no mic eliminates off-axis sound from the sides or rear, just attenuates the level, and whatever extraneous sound does get picked up can be uneven in frequency response. This is why two mics with almost identical pickup characteristics for front-arriving sound may nonetheless sound very different in use, due to the varying coloration or tonal balance of their off-axis response.
A few other general terms and issues related to microphone design and use are:
• Proximity effect—this is a boost in low frequencies, heard when the singer is within a few inches of the mic’s diaphragm. This gives radio announcers their deep chesty voices, and it can sound good on stage, but is better avoided in the studio by maintaining several inches’ distance. (Note that omnidirectional mics are not subject to the proximity effect.)
• Plosives—a popping sound, caused by a burst of air from the singer’s mouth, typically on consonants like B’s and P’s. A “pop filter” in front of the mic will usually do the trick—buy one or fashion one yourself with some pantyhose and a wire hanger. Sometimes angling the mic a few degrees will also help.
• Off-axis pickup—keep an ear open for tonal inconsistencies when a singer moves or sways a bit to the music while performing. If necessary, back them off the mic a little further to minimize obvious tonal changes, but keep an ear open for too much overall room tone at that distance.
• Nasal voice quality—this can sometimes result from the placement of the mic. Often, setting the mic a little above eye level and angling it down will not only encourage the singer to keep their head raised, helping to open the throat and project the voice better, but will also pick up head and sinus resonances better, minimizing nasal tonality.
• Unpleasant resonances—not really a mic issue directly, but watch out for things like that metallic music stand—it can reflect sound back at the singer which even a directional mic may pick up as a nasty resonant peak. It’s better to use an open stand or pad it with absorbent material.
Obviously, it’s harder to keep on top of these potential problems if you’re the vocalist and the recordist—in that case take an occasional breather during the session and switch hats for a moment to make sure that everything’s sounding as it should.
The front end
With the room chosen and the mics and headphones set up, the next consideration is getting the signal cleanly into the console and recorder. The output of a microphone is a low-level signal with a low output impedance, and so requires substantial gain along with an appropriate input impedance to transfer the signal properly. Line-level inputs won’t cut it—a dedicated mic preamp is required. These are built into most consoles.
Along with the trend to DAW recording, the use of separate external mic preamps has become popular. They are usually of better quality than the inexpensive ones built into most small mixers, and often employ tube stages (just as some mics do) to “warm up” the signal.
Some of these external mic preamps also have eq and compression all in the same freestanding box, making them modular channel strips, comprising an analog “front end” to shape the signal and impart some subtle character to the otherwise ultra-clean recorded quality of the typical modern DAW. In fact, nowadays an array of outboard gear is often used in place of a traditional console, connected directly to the recorder’s A/D (analog-to-digital) converter/interface.
Squeezing the singer
Along with the mic pre and (sometimes) outboard eq, compressors are often employed to control the vocal signal. Compression is a big topic, as are mics, so I’ll try to focus on just the basic aspects related to vocal recording here.
Compression reduces the dynamic range of a musical signal from soft to loud, lowering the louder parts and bringing up the soft parts in level. This is often necessary so that a very dynamic performance will not exceed the limits of the recording medium, or vary in level too much against the track. Since at this stage we can’t be sure how the vocal will sit in the track when it’s done, a good approach is to use compression (if it’s used at all here) simply to insure a good strong, clean signal on the way in, and apply compression again at the mix stage to address issues of musical balance.
The primary controls on a compressor are threshold and ratio. Threshold determines the level above which gain reduction occurs—below this point the signal will be unaffected. Ratio determines how much gain reduction will be applied when the signal exceeds the Threshold. For example, a 3:1 ratio setting would reduce the gain of a signal that comes in at 3 dB above threshold to 1 dB above threshold after compression.
Then there are controls for attack and release settings that shape the level fluctuations over time. After the level of the louder signals has been reduced, output or make-up gain brings the overall signal level back up to where it was, bringing up the volume of the softer portions which otherwise might have been too low to record well.
Gain change in compressors is controlled by either a voltage-controlled amplifier circuit (VCA) , or by an optical (light) element (Opto-compressor). VCA-based models are usually fully adjustable, while Opto designs often have only two controls—compression amount (a single knob that controls both threshold and ratio) and output. Opto compressors not only sound excellent for vocals, but it’s easier to get good results for those inexperienced in the art of compression. Nonetheless, when using a full-featured model, here are a couple of slightly different approaches:
• Set the threshold very low, so even soft signals typically cross it, and then set the ratio also to a very low value, like or 1.5:1 or 2:1. This will subtly compress the entire signal range, imperceptibly bringing the softest and the loudest sounds closer together in level without really sounding noticeably “compressed” or affected.
• Set the threshold fairly high, so only louder passages will cross it, and increase the ratio as well, to anything from 3 or 4:1 to as much as 8 or 10:1 (ratios above 10:1 are called limiting). This will leave the softer passages alone, and compress only when the singer sings hard, potentially a more obvious change in dynamics, but one which can sometimes add a sense of excitement to the recording.
Both approaches are valid—which one works better will depend on the particulars of the compressor in use, the singer’s performance, and the overall style of the music and recording. In general, I tend to favor the former during tracking, because the dynamic control can be more transparent.
One final word about compression: You should take care not to over-compress the vocal at this stage—it can be hard to undo if it doesn’t really work later in the track. Sometimes it’s better to just let the singer “work the mic,” that is, pull back a bit on loud passages and move in slightly on soft ones. If some care is taken to minimize tonal inconsistencies, as above, this can be an easy, natural way for vocalists to control their own dynamics.
Whether you’re taking signal to tape or disc, the final step in setting up is to insure a good level is going to the recorder—this means as high as possible without overload. If the medium is analog, and you’re not experienced in pushing hot levels to tape, it’s probably best to stick to the meter’s 0 VU redline as a maximum. Most readers are probably recording to hard disc with a DAW anyway these days, and here the situation is clear. Maximum signal level in a digital recorder is 0 dBFS (where FS = full scale), and it’s absolute—exceed this and harsh clipping will result.
Okay, so the pre-flight is done and now it’s time to hit the red button and start capturing those solid-gold vocal performances. Let’s consider a few aspects of the actual recording process that come into play at this stage. First off, are you playing the dual roles of engineer/producer and vocalist, or has outside talent been brought in to handle the vocal chores while you man the gear and direct the show? If you’re recording yourself, the best advice I can offer is to try and set up all the technical aspects before you step up to the mic, and once you begin to perform put all your energy into singing, don’t let yourself be distracted by details having to do with the recording process. With that in mind, here are a few basic do’s and don’ts:
• Don’t watch the meters constantly while you sing, it’ll make you gun-shy if you see that levels are a bit hotter than expected, and possibly compromise your performance. Get a good level and then focus on the singing.
• Check levels on playback after the first take in full voice to be sure everything’s okay before you continue—set gain a little conservatively or employ a limiter if you’re performing a song with a highly dynamic vocal part.
• Try to set up remote control of the recorder at the mic position. If you have to run back and forth too much, you’ll not only lose focus, but you may not be able to maintain a consistent position on mic, and the resulting tonal differences will make it harder to combine elements from different takes.
• Don’t try to perform manual punch-ins if you can avoid it—set the recorder to auto-punch whenever this is required (see below).
Method to the madness
Another basic question, especially if you’re recording a guest vocalist, is whether to approach the song as a whole, beginning-to-end each take, or to break it up into parts and work it section by section. Both methods have their merits, and either could work best for a given song. Going for complete takes is good when the song has a dynamic arc, or if you observe that the singer’s interpretation is more nuanced when it builds continuously throughout the tune.
On the other hand, some singers are more effective if they take it a verse or chorus at a time, focus on that area, and fine-tune their performance in smaller increments. The latter is especially true if someone’s just learning the song that day (not the best situation but sometimes unavoidable), or if the tune has difficult transitions that are best approached in isolation.
Sometimes, mixing both approaches is called for. For example, in a recent session at my studio, a vocalist was called in by a songwriter. The singer had been given a music-minus-one CD (the track without the vocal) and a lead sheet a few days earlier (always a good idea!), but just prior to the session the songwriter had changed a couple of sections. After consulting with the singer, it was decided that she’d sing the song throughout as she’d already learned it, and then we’d go back and replace the changed sections piecemeal, once the overall arc of the vocal performance was established.
This turned out well—the singer had prepared a gradually-building musical interpretation which imparted a dynamic shape to the full takes, and provided a clear direction for the feel and vibe of the overdubs. Additionally, the new sections had some difficult intervals transitioning into them, and hitting those lines separately resulted in the tightest performance pitch-wise.
The same considerations apply when overdubbing or punching in smaller sections of a track—some vocalists work better with tight punches, others need a longer lead-in to get good results. Fortunately, with today’s DAWs, it’s easy enough to simply set a cycle between two points and have the recorder automatically repeat a section over and over.
Most systems allow you to specify the punch-in and punch-out points as well as the pre-roll/post-roll (how far before and after the punch the music will play, for the singer to get into the feel and timing). This Loop- or Cycle-recording capability is also a godsend for those who are recording themselves, as the distraction of trying to handle punches while singing is a surefire way to compromise the performance.
Managing the talent
Looking after the comfort of the singer is a big aspect of ensuring the success of the session. The voice is a fragile instrument, and its sound is greatly impacted by the singer’s mood, fatigue, well-being—all factors on which you can have benign influence if you make it a priority. Mood is the biggest factor—it requires positive reinforcement and enthusiastic responses from you (even when they aren’t really warranted—a bit of diplomacy goes a long way!).
Provide for creature comforts, offer tea, turn off some lights, allow for breaks etc.—it will pay off at least as handsomely as renting that expensive German mic. Many a good engineer or producer has been heard to say, after the singer slightly flubbed yet another take, “That was good but I screwed up—one more, please, just for me.” The red recording light can be cruel on the psyche of performers, even experienced ones, so don’t forget the human element in all this!
Here’s a little story to illustrate. Recently, two singers were booked to come on the same day, one after the other, to work on different sections of the same song. The singer scheduled to come in second had already recorded at our studio—she was an accomplished vocalist with a well-controlled voice, who was very conscientious, meticulously working out parts in advance, always prepared. She arrived while the first vocalist was just finishing up. That first singer, as it turned out, had an absolutely huge gospel voice, and was improvising harmonies and vocal flourishes with abandon.
It was clear that singer number two, despite being experienced and having worked for the songwriter before, was intimidated by the impressive performer who preceded her. She wasn’t really on her game until we took a couple of her early takes and processed and mixed them a bit, so she could hear that in the final mix her voice would stand up well against that of the other singer. After that, she settled in and gave an excellent performance.
Managing the session
As the session progresses and the takes accumulate, managing all this raw material becomes a key issue (especially with large HD-based systems where nothing need be thrown away). Naming and cataloguing takes as you go might add a few seconds each pass, but it’s a very good idea—a little time spent on this during the recording will save a lot of time and energy in the editing stage.
Some DAWs help out with organization, letting you record multiple takes to the same track, showing only one at a time as you go, but allowing you to view and edit all the takes conveniently later. For systems that don’t have this exact capability, the recordist can usually set up the same thing, using multiple tracks assigned to the same recorder channel, or using subfolders to accumulate and organize repeated takes. When possible, color-coding is also a great way to do this (it’s one of my preferred techniques)—use a family of colors for each different part, like blue for lead vocal, green for harmony, with different shades denoting preferred takes or specific problem areas.
Editing and comping
When recording is done, you’ll likely be staring at a healthy accumulation of vocal takes, overdubs, punch-ins and various and sundry snippets of audio. If you’re very lucky, one of those complete takes will be just perfect, and you can pack up and go home happy. In the real world, though, as often as not you’ll need to create a vocal “comp,” a composite track made up of the best pieces of the various recordings at your disposal.
This is where all that labeling and cataloguing during the session will be a real life-saver. Fortunately, nowadays it’s relatively easy to graphically arrange sections of audio into a new track in a DAW’s editing window. Additionally, the use of crossfades can smooth the transitions between different snippets of audio butted together, where there would otherwise be a click—this can really increase editing speed and efficiency.
From a musical standpoint, when comping pieces together from different takes, you need to listen for a good match not only in technical quality (mic position, room tone, level) but also in performance (similar vocal quality, inflection, interpretation). If a section needs to be replaced and there’s no good match from an alternate take, you can often splice in a section from earlier or later in the same take. An experienced editor can even sometimes create a missing phrase word-by-word, and it’s not unheard of to splice in individual syllables or even consonants!
Of course, as with anything else, the law of diminishing returns can eventually set in—sometimes it’s better to leave a little technical imperfection for the sake of feel and excitement, rather than edit everything to the nth degree and lose the soul of the track in the process.
When the mixdown stage arrives, the full range of signal processing tools is often brought to play to make the final vocal comp sit well in the overall mix. As mentioned earlier, eq and compression are often used here (again), along with reverb or ambience, and the use of delays and doubling of the vocal line (to enhance a thin or tenuous vocal—aren’t you glad you saved all those takes?).
Tools for the modern age
Besides the traditional mix processing techniques mentioned above, the DAW environment offers a few additional tricks for fixing up vocal tracks. Modern software plug-ins and outboard effects offer a wide range of useful and amazing vocal utilities, like transposing notes or phrases, generating harmonies automatically, or matching the rhythm and timing of one vocal line to that of another.
One of the most widely used types of processing is pitch correction. By now, everyone is probably familiar with plug-ins like Antares’ Auto-Tune, which can take a vocal track with a few slightly off-key notes, or even one that consistently veers sharp or flat, and pull it into tune automatically. This tool (and others like it) have become so popular that many engineers just throw it onto a track without even going through the song carefully to get a feel for the singer’s overall pitch—some singers even request it!
However, it’s probably still a better idea to use pitch correction tools only when needed. Despite the relatively transparent quality of today’s vocal correction algorithms, processors like this can still impact the vocal in subtle ways, ever-so-slightly reducing the degree of nuance and musical gesture. So on passages that don’t really need to be “fixed,” bypass it, or split the corrected vocals to a separate processed track, to preserve as much subtle musical detail as possible. In fact, that advice goes for any type of processing—depending of course on the style of music, very often less is more, and the best recordings are frequently the simplest ones.
Well, that’s it for now. We’ve covered the entire vocal recording process, and pretty much any of these topics could be the subject of a whole article, but hopefully this broad overview will serve as a decent introduction to the many aspects of recording and producing killer vocal tracks.
One last bit of advice: whatever level your singing and recording chops, don’t let yourself get sidetracked by the technology and logistics of the recording process, just go ahead and do it! Remember, at the end of the day, it’s all about the song.