how to sing them, how to mic them, how to treat them, how to tape them...
Take five and come alive - we show you how
Nothing will help your music more than an excellent vocal track. So why do so many people get it wrong? Paul Freudenberg highlights the pitfalls in a typical vocal dub session, and offers some smart suggestions of his own
You can perform all sorts of fancy tricks on your recordings. But if the vocals don't sound right, you may as well have done the whole thing on paper and comb. There are no short cuts to getting vocals to sound great. A high-quality microphone won't perform at its best if it isn't properly positioned. The signal from that same mic could be ruined if it isn't sympathetically processed. And the best vocal sound in isolation can be a disaster in the context of a finished mix.
The only way you can be sure of giving your singer(s) every chance of getting a good performance on tape is to go right through a vocal overdub session - from microphone selection, processing, and performance, to tracking and mixdown. Which, strangely enough, is exactly what we're about to do in this feature.
So - in with the XLRs, on with the cans, and let's roll...
First, what are the differences between mic types, and how do they affect the final product? The dynamic principle, where the mic is basically the opposite of a speaker, is one of the most common. In a dynamic mic, changes in air pressure cause the diaphragm and connected coil to move within a magnetic field. The movement generates a voltage that then goes to a balancing transformer.
Typically, the membrane of a dynamic mic is around 40 microns thick. Because of the heavy mass of the diaphragm/coil assembly, a dynamic mic is relatively slow to respond to transients, less sensitive (and therefore more capable of withstanding high sound pressure levels or SPLs), and slower to respond to transients. As a result, dynamic mics tend to colour the sound of the vocal, and are chosen for their unique character.
Many so called 'instrument dynamic' mics aren't suited to miking instruments or full-range vocals. For an instrument mic to really reproduce low frequencies, there has to be a chamber connected acoustically to the dynamic capsule.
Condenser mics tend to be more accurate just because of the nature of the condenser principle, their construction, and the thinner diaphragm material used. All condensers require power, whether it's external 'phantom power' (usually supplied by the console), or just batteries.
Large-diaphragm mics, with their one-inch diaphragms, are more popular for vocals, small-diaphragm ones for instruments. This is due to the acoustic and physical properties of each, starting with the mass or the physical size of the diaphragm. The larger the diaphragm, the slower the transient response - the heavy diaphragm takes longer to get moving and react to the changes in air pressure. This results in a slightly 'rounder' or 'smoother' sound.
A small diaphragm condenser with half the mass (a half-inch diaphragm) can react much more quickly, and as a result has a 'sharper', more defined sound, which is technically much more accurate. Also, the physical size of the diaphragm has an effect on the changes in the air pressure around it. The less space the mic takes up, the less it influences the sound around it.
Condenser mics share a similar type of construction: gold 'sputtered' mylar plastic stretched over a metal conductive backplate separated by a small gap. When in silence, the diaphragm is still, and no electricity is exchanged between the diaphragm and the backplate. The moment there is a change in pressure (when there's sound), the diaphragm moves, and electricity jumps the gap between the two conductors.
Large diaphragms provide more surface area to conduct, so more electrons can potentially 'jump ship' on command. This results in a lower noise floor compared to a mic with a smaller capsule.
The final difference between large and small-diaphragm condensers has to do with the relationship between the diameter of the diaphragm and the wavelength of the sound. It so happens that the wavelength of 10kHz is about an inch long. If a condenser mic has a one-inch diameter diaphragm, a build-up of frequencies begins at precisely that point, which results in more directional sensitivity at higher frequencies. In a small-diaphragm condenser, with its half-inch diaphragm, this build-up happens at around 20kHz, so it's far less noticeable.
You can see that a large diaphragm condenser mic will influence the character of the vocal sound a great deal more than a small one. But these colorations often help get the vocal to project out of the mix. One engineer friend of mine uses a large-diaphragm mic for the lead vocal, but a small one for background vocals so that they blend in better with the mix.
When comparing mics to assess their relative suitability for the job, make sure that the mics you're comparing have equal gain. The human ear is very easy to fool - louder almost always sounds better. And when possible, test the mic on the vocalist you're planning on recording.
Never try to judge the sound of a mic while you make the sound. The bone conduction in your head creates such a disturbance that you lose your sonic reference. If you're going to be recording yourself, record the mic to tape and then listen back.
Positioning a vocal microphone depends a lot on the style of performance, the production quality, and the performer. For most sessions, I position the mic so the bottom of the housing is just level with the singer's nose. This ensures that the singer is in a good posture, and you avoid most of the normal pops and breath noises.
"Take the direct output of your console input strip directly to tape, bypassing the unnecessary buss amp. This will provide the most pristine signal you can get"
I like to start with the performer about two feet away, but then move in a bit if the sound is too distant. Some songs demand an even more up-front and intimate sound, in which case I use a really thick 'muffler' pop filter and move the vocalist right in on the mic. Turning the mic slightly off-axis to the performer greatly reduces the dreaded 'pop' noise.
The 'proximity effect' is the build-up of the bass frequencies as the performer moves closer to the microphone. This natural EQ can be quite useful in some situations, providing a warm and upfront sound. The effect occurs in all dynamic and condenser microphones unless they have special porting to counter it. The tighter the polar pattern, the more obvious the effect. Proximity is barely noticeable in omnidirectional patterns, but it shows up more with cardioid patterns, and even more with hypercardioids.
There is also a difference in the natural EQ between dynamics and condensers. Dynamics tend to have a narrower Q (bandwidth, or range of frequencies affected).
Just as different microphones produce different pickup characteristics on-axis (in front), they're different off-axis. This is important when selecting mics. The more expertly crafted the mic is, the more natural the off-axis sound is. Using the mic off-axis can be a particularly effective way of getting a thinner sound for background harmony.
As an engineer, you want a headphone that provides plenty of isolation, so the playback track does not leak into the vocal mic. As a vocalist, you want something that looks good, doesn't mess your hairdo, and won't shake off if you really get going.
Closed headphones, where the sound has no escape hole, are the best choice for isolation. Unfortunately, most of them suffer from bass frequency build-up inside the ear cup and the singer's ear canal. The most common studio headphones use a semi-open design to control the amount of bass build-up and sound much more natural. Open headphones, such as you'd use on a Walkman, are no good for tracking because they leak.
More than other instrumentalists, a vocalist has to contend with the acoustic coupling that goes on inside his or her head. Sound from the vocal chords vibrates the bones all the way up to the ear cavity, and eventually the ear drum itself. That's how vocalists are used to hearing their own voices - and why most of us cringe when we hear our own played back.
When you stick a pair of headphones on a vocalist, you disturb his or her learned monitoring environment. Many singers prefer to have the track - usually without the vocal - played back into one ear, leaving the other uncovered for listening to the singing. As the engineer, you are concerned with leakage, so have the performer keep his or her ear completely covered.
For singers who can't deal with headphones, here's a studio trick that lets you record without them - even in the control room. Position two monitor speakers behind the microphone at an equal distance from it and from each other. Three feet is a good distance to start. Now reverse the phase of one of the speakers (switch the polarity of its cable) and feed the monitors a mono send. The speakers will cancel each other out at the mic, but the performer will hear the sound normally.
Again, most pops and other breath noises can be suppressed by careful microphone positioning. Sibilance ('S' noise) is a different beast that manifests itself in a number of ways.
Sometimes these 'S's' come from the natural sound a vocalist makes, in which case proper mic selection is essential because it avoids accentuating the problem. Choose a mic with a slightly diminished high-frequency response, and avoid boosting the EQ while tracking.
Some sibilance can be caused by high-frequency overload in the mic. This is often just a matter of selecting the -10dB pad to lower the overall level into the preamp. Most of the 'S' trouble I've run into is a direct result of poor recording techniques: too much high-frequency EQ during the recording, and recording too hot.
The trick is to get the best un-EQ'd sound to tape, and then work on the timbre, using the EQ only if necessary. Remember that there are two ways to EQ: additive and subtractive. For example, a shelving boost at 8kHz is sonically very similar to a cut around 150Hz. Always try the lower noise alternative (cutting) first.
"Try an expander or noise gate on the output of your reverb to shorten its envelope more naturally"
If nothing else seems to work, a compressor with an EQ'd sidechain or a dedicated de-esser will help. If you prefer a bit more processing horsepower, read on.
Many people use specialised outboard gear to record the most exposed, upfront vocal tracks. They feed the output of these components directly to tape, bypassing the console while recording. The console is used only for monitoring because there's no reason to run the signal through any more electronics - or even through more wire - than absolutely necessary.
Your basic component record path might run: mic → mic pre-amp → parametric EQ → compressor → tape. If you don't have premium individual components at your disposal, try shortening the signal path by taking the direct output of your console input strip directly to tape, bypassing the unnecessary buss amp. This will provide the most pristine signal you can get.
The outboard EQ can be used at the compressor's sidechain to make it more sensitive to certain frequencies (for example, the 'S's' at around 6-8kHz). Use your compressor sparingly when recording - you can always add more in the mix.
Here's one tried and tested plan of attack for vocal recording. While the vocalist is fresh, get him or her to belt out the most emotional track possible, ignoring everything but the most glaring pitch problems. Using that track as a guide, and build a second track, this time punching in where necessary. Between two takes, you should have more than enough to work with.
Most recording artists don't produce one-pass takes any more. Even the top professionals rely on technology to make their performances flawless. I've seen vocal multitracking used creatively as one artist built a chorus of his own voice. I've also seen it abused with the mega excess of vocal slave reels (dragging a second recorder along to make room for dozens of takes). This generally leads to vocal 'comps' where each word is picked apart and selected for its pitch, performance, and perspiration.
The truth is that how the vocal is recorded is much more important than a few imperfections. However, there are many times when it's necessary to punch-in on one area of an otherwise sparkling take. So make that punch good.
Here's how. First, make it clear to the performer where the punch will take place. Choose the natural end of a phrase or line, if at all possible. Have the vocalist 'sing into' the part (sing along with the take before the punch point), and continue singing all the way until you stop the tape. This will ensure that breaths taken between lines match, while pitch and level problems due to 'cold starts' are kept to a minimum. It will also prevent dynamics processors (like compressors) from causing level matching problems due to their attack times being slow.
Vocal compilations, or 'comps' for short, are often made by taking bits from two (or maybe three) performances and putting them together to form a complete take on an open track. To do this, take the vocal tracks off your console's main mixdown buss and assign them all to one multitrack buss. Then patch that buss to the desired track you wish to comp to. Remember that adjacent tracks feed back when bouncing in sync (on analogue machines), so make sure you have at least one track separating the nearest playback track from the record track.
Monitor the tape only from the output of the comp track - not the original tracks. Use the console mute switches to select the track that goes to the comp track. Keeping the levels equal is very important at this stage, too.
For a bigger sound, you might choose to double the vocal in certain sections, such as the bridge, to fill out that part. If so, pick the best take as the main one, then bring the double in underneath.
At the end of all this, you should have one perfect take and be ready to mix.
"The human ear is very easy to fool - louder almost always sounds better"
With few exceptions, I recommend that digital effects units not be used for these or any other applications where the entire signal is processed.
Analogue processors are better suited for compression, EQ, gating, and so on; use your digital effects boxes for reverb, delay, and other applications where you mix just a little of the effect in with the dry signal. Use your digital multi-effects units on auxiliary sends, and return them through dedicated returns or empty console channels; reserve the in-line processing (channel inserts) for analogue devices.
Here are some of the ingredients to use for vocal tracks:
Compression and limiting - These two dynamics processes allow the engineer to 'package' the vocal track. The compressor works to bring the level up and keep longer duration levels even, while the limiter waits on top to catch any sudden peaks that might pop out. Patch this device (it's usually a single unit) in line with your vocal to keep it just where you want it in the mix.
Slap delay - The most effective device to support or strengthen a single vocal track. Mix a delay of around 70 to 120 milliseconds in with the vocal. Try using this delay return as the send to the main reverb. This separates the vocal from the room, and simulates the effect of direct sound, first reflection, and diffuse room ambience. If you don't have a processor available, bear in mind that even a three-head analogue cassette deck can produce a very effective slap echo.
Reverb - Selecting a reverb with a slightly different character from the main one can help to draw attention to the vocal.
Shorter reverb times that simulate smaller spaces put an environment around the singer, and are good for faster-tempo songs.
However, this type of space can be the hardest to simulate on today's cheaper digital reverb units. Try an expander or noise gate on the output of the reverb to shorten a reverb envelope more naturally. Select longer reverb times, with good predelay, to simulate large, concert-hall performances.
Harmoniser - Very effective for concealing 'pitchy' vocals or strengthening a weak singer. Just a notch or two over or under the original vocal lends support. This creates a very 'modern' vocal, but too much produces an unnatural 'beating' effect that can be annoying. Harmonisers can also be effective for spreading out the background vocals for a full, wide stereo sound field.
Double tracking - In the old days, when there was an extra track (on the four-track), the lead vocal would be 'doubled'. This would produce a support track with random imperfections, due to the differences between the two performances. The late Harry Nilsson was an expert at doubling his vocal, getting the slightest phrasing nuances so close you could barely discern the two parts. Today, a similar effect can be simulated by using a very short delay time (45 milliseconds or less) and a slow, chorus-type modulation. This changes the pitch and time ever so slightly, but just enough to make a subtle difference. In any case, mixing the support vocal track, or modulating delay beneath the lead, creates a nice centre for the rest of the track to hang onto.
I haven't said much about equalising the vocal track up to this point because I am a firm believer in leaving the track 'naked' until the mix. From the beginning, try to visualise where the vocals will fit in terms of frequency response and dynamics. Leave room for them by arranging the other instruments' timbres around a 'hole' where the vocal will sit.
This frequency 'hole' is at about 1-5kHz. While the vocal frequency range is wider than that, those are the frequencies that have the most to do with intelligibility. You'd be surprised what the slightest EQ bump at 5k can do in this respect. Also, the bass end of the vocal can step on the guitar and other lower-mid instruments. You can roll-off some of the low end and still keep the vocal loud and on the top of the mix.
"Use your compressor sparingly when recording - you can always add more in the mix"
Stay away from graphic equalisers - they often cause more harm than good. Graphic EQs have many filters that result in more phase shift over the audio spectrum than parametric circuits. They also have a fixed bandwidth, and bringing up adjacent bands usually creates a ripple, rather than the smooth curve the front panel suggests.
Ultimately, while getting all the technical aspects of your vocal dubbing session in order can make a big contribution to the sound, nothing is more important than a relaxed atmosphere and a positive vibe. The arts of the host(ess) are as important as those of the engineer.
Accentuate the positive in your vocalists' performance, give them refreshment breaks, and don't get too technical. An easy mental picture, like "throw the sound out", will get a better result than blinding them with science.
The vocal is your recording's crowning glory. So - at the risk of creating a primadonna - let nothing interfere with your singer's freedom or comfort.
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!