Recording vocals is harder that you might assume.  It’s much more than just throwing up a mic and keep levels out of the red.

Believe it or not, some of the most important parts of recording (or more accurately, CAPTURING) a good vocal have nothing to do with microphones, but rather with your ability to CARE about your singer and their performance.  You are in a unique position from which you can not only elicit the best performance from your singer but also capture all of the important nuances of that performance.  You have to be aware of those nuances and what can affect a performance.

The slightest change in volume, tone, and of course pitch carries expression.  Combine the melodic and rhythmic changes with the meaning of the lyrics and you have a powerful form of communication.  It is your job to protect and enhance that communication.  I often say that I consider vocals sacred, and I strive to capture vocals as purely as possible so when I mix I can then enhance the expression that has been captured.  Unfortunately I often have to mix vocals that are so badly recorded that I am severely limited in what I can enhance.  I find this a frustrating situation that angers me, because it keeps me from being able to serve the song in the best way possible without first salvaging the vocal.

To record vocals the way that I prefer requires an understanding of microphones, mic-preamps, compressors, acoustics, and a little psychology.  Your engineer’s toolkit needs to include empathy along with a good selection of mics.

1. CARE ABOUT THE SINGER.  The first step to recording a good vocal is to care about your singer.  This seems simple enough, but there are many people that use the exact same technique (and even settings) regardless of who they have on the mic. This might work if you are banging out demos or jingles, but if you cannot get your singer comfortable then you are limited in the performance you can get from them.  Granted, there are expressive aspects of a vocal that is anxious or uncomfortable, but this should be elicited intentionally rather than the byproduct of uncomfortable singers recorded assembly-line style.

You need to create an environment that is most conducive to eliciting the best performance from your singer.  Some singers will want an intimate environment, others will want something more open….so what do you do? 

Start with controlling the acoustics in the available space at your disposal.  You want your singer in an enclosed area so that the microphone will not pick up too much ambience from the surrounding room.  While a vocal booth will work, so will putting up sound absorbent partitions around your singer (making a vocal booth within a larger room).  Beware of unwanted reflections off of the floor (a rug will help that) and also unwanted reflections off of any glass in front of the singer, especially when using a small vocal booth.

Now that you have controlled the acoustics you can start to think about singer comfort, which for now will mean giving them the amount of intimacy that they want.  Intimacy does not necessarily mean a smaller vocal booth or closer partitions, which can be claustrophobic.  You can accomplish a great variance of intimacy with light. Florescent lights, lamps, or light that is reflected off of the walls will create different feeling environments, as will candles (used carefully of course). 

2. TRUST.  Trust is the most important thing in a vocal session.  When you record a band together in one room they will communicate with each other throughout the process, but a singer is often alone in the room when recording.  Unless there is a Producer involved it’s just you and the singer.  The singer MUST trust you in order to relax enough to focus on performing.  They have to feel that you are looking out for their best interest, which is something you are going to have to prove the first time you work with them. 

Ask the singer if there is anything they need, or anything that you can do to make them more comfortable.  Then try to accommodate their needs as best you can.  Remember that singers are often nervous and will not ask for anything.  Provide them with water or tea in order to keep their throats moist and comfortable, and do not feel shy about suggesting they have some if they start to sound dry.  This will not only help their performance but will make them realize that you are tuned in to needs they themselves might not realize.

Once you are recording, even though the singer may have already approved of conditions (including the headphone mix, which we will discuss later) you should still ask if everything is ok and if there is anything you can do to make them more comfortable.  Don’t be annoying. Be sensitive as to when is a good time to ask and at what point to stop asking.

Treat the singer like they are the most important person in the world, because for the duration of the vocal session they ARE.  Especially if it is just you and the singer you need them to really understand that you-are-there-for-THEM.  As a result of this the singer will relax a bit.  They will understand that you are taking the time to be considerate of their needs and start to trust you. 

If the singer is nervous or inexperienced then your job will be harder.  You need to be more diligent, more watchful, and more sensitive to subtle cues that will indicate their needs.  I have often gone into the studio or vocal booth under the premise of adjusting something so that I could talk with the singer and put them at ease.  Sure I could have just talked through the headphones, but being in the same room is more personal.

3. KNOW THE SINGER’S VOICE.  You do not want surprises, but rather you want to be able to predict the sounds that you will be capturing.  If there is a scratch (reference) vocal already recorded, listen to it to better understand the different parts of the song and the vocal performance.   Then have the singer sing parts of the song and listen to them IN THE SAME ROOM WITH THE SINGER (no mics yet).  Try to get to know the song and how the singer’s voice sounds when they are loud, quiet, and at different pitch ranges.  Often a person will sound one way when singing softly or at a low pitch, and completely different when they sing louder or higher notes.  Look out for frequencies that peak out at each part of the singer’s range and volume.

This serves two purposes.  First, you will learn what you need to know in order to pick the right microphone.  Second, it will force singers that do not usually warm up to do so.  As you will see below you should be recording EVERYTHING that comes out of the singer’s mouth when they are performing with the track, and if they are warmed up the first take will be better.

4. HEADPHONES.  Although you will most likely have the singer start out acapella to test mics and set starting levels, you will need to have your headphone mix ready to go.  Expect to have to adjust the mix to give the singer more or less of themselves (best accomplished by adjusting the volume of the band as well as the return volume of the vocal track), more drums to help their timing, or more keys or guitars to help their pitch.  I suggest having the band on a group (or going through a pair of faders) to make it easier and faster to make adjustments. 

The first time I recorded Roberta Flack was also the first time I had ever worked on that particular model Neve.  I arrived to the session two hours early and had THREE different headphone mixes ready, two using different sends and the third on a separate monitor section of the board.  That way if the first mix was not right I could easily switch without having to spend time remixing.

When you are creating the headphone mix, use the same headphones as the singer, and the same amp.   If you plug your headphones directly into the board but the singer’s headphones are going through an amp they will sound different, even if the headphones are both the same model.  If you are using different model headphones than the singer be careful not to make changes that might sound good to you but wrong to the singer.  Remember THEY ARE THE MOST IMPORTANT PERSON, and their headphone mix has to be the best.

Some people like to make headphone mixes using sends or separate faders, and other people prefer to simply give the singer the 2-mix (master fader output) coming out of the board.  Using sends rather than the 2-mix will allow you to change the mix in the control room without affecting what the singer hears.  If you decide to use the 2-mix then you have to use PFL (pre fader listen) to solo tracks without affecting the headphones.

You can give the singer some reverb or delay in their headphones, but be careful not to give them so much that they cannot hear their subtle nuances.  Big swimmy reverbs that bury vocals might give the singer a little more confidence, but it might also affect how well they can hear their own pitch and timing.

5. MICROPHONE.  Never assume that the most expensive mic is the right one for every singer.  Test every decent mic at your disposal with the singer singing different parts of the song that show the range of volume and vocal tone you will be recording.  Use the mic that sounds most like when you were in the same room as the singer.

I like a warm vocal mic that is clear but not too toppy.  Air is nice, but airy mics can emphasize ugly frequency peaks when certain singers get loud or higher in pitch.  I am going to avoid listing specific models since mic sound is a matter of opinion, and leave it at “I like a warm vocal mic”.

6. POP FILTER.  I suggest using a pop filter, especially with singers that have limited experience.  Pop filters can be purchased or cheaply made by stretching a stocking around a wire hanger shaped into a circle.  I have heard of people putting a pencil in front of the microphone to deflect the burst of air resulting from pops, but I never achieved good results with that technique.

Some mics have built in pop filters, and in some cases (depending on the singer and the mic) you can go without…but you run a big risk doing so.  Yes it is possible to try to fix pops while mixing (which I will discuss in a different unit) but it is difficult and often does not sound natural.

7. MIC PREAMP AND LEVEL.  The microphone pre-amp will color the sound of the microphone.  Just like mics, pre-amps vary from warm to bright and clear.  The combination of the mic sound and the pre-amp sound will give you the final sound you get (in the same way that your monitors sound is a combination of the colorations of both the speaker and the amp driving it), and you can compensate for an overly bright mic by using a warmer pre-amp.

When you set your level, start with the fader at zero and the mic pre at a low setting.  If the fader is lower than zero then you may end up making the mic pre louder than needed and introducing distortion. While the singer is singing loudly bring up the mic pre until you get the desired volume.  Watch out for distortion at the mic pre-amp (unless you are specifically going for a distorted sound).

8. COMPRESSION AND DYNAMICS.  Ah, dynamics.  I personally LOVE dynamics, and often will enhance or even introduce dynamic expression while mixing (I have to take a moment to thank Peter Yarrow for helping me to understand how important vocal dynamics can affect expression and communication).  Your singer will have a dynamic range that can potentially go from a whisper to a scream.  If you have a good level for the whisper, then the scream will overload.  If you have a good level for the scream, then the whisper will be too quiet.

A compressor will control dynamics as well as change the tone and even make the vocal more solid.  Unfortunately too much compression will result in a “squeezed” vocal sound rather than a full one.  I prefer to use minimal amounts of compression (ratios of around 3:1 to 5:1, with the compressor threshold set to ignore the really quiet parts, slightly compress the mid volume parts, and kick in strongly only for the really really loud stuff).

Some people will set their levels and compression so that quiet parts have reasonably good levels and loud parts are compressed to death.  I am not a fan of this and would rather have a loud sound that is not overly compressed, and then ride the fader up for the softer parts to ensure good levels.  Of course this is very difficult if you do not already know the song and how the vocal part is expected to change throughout all the parts.

The most extreme rides to tape I ever did was when I was recording the Amandla album with Miles Davis.  Miles was known for having a wide dynamic range, and I didn’t want to overly compress him just to have a safety net.  I recorded him with no compression at all (no eq either) and performed rides that were in excess of 20db while he played.  Of course I could not predict what he would do, so I listened very closely, watched his breathing, and tried to predict (based on his previous phrases and how much he breath he took in) how loud the next notes would be.  At one point I was at the top of the fader and he kept getting softer, so I had to bring up the mic pre-amp.  The sound was particularly full and rich, and when Miles came in to listen he turned to me and said, “That’s a nice horn sound.”  It was a very good day.

9. RECORDING VOCALS WITH EFFECTS SUCH AS AUTOTUNE.  One word…don’t.  Do that stuff afterwards or on a different track.  The only exception is if the effect is an integral part of the sound, such as the autotune on Cher’s “Believe” or the “Autotuned” sound that is so popular among people that do not like to be bothered with trivial things such as pitch.

10. CAPTURE.  In my experience, it is important to record EVERYTHING.  The first take may be a little rough but might have “magic”.  The golden rule is that if the singer is listening to the song and making any noises you should be capturing what they do.  Even if the singer is just playing around and not really “singing the song” you may end up recording a throw away adlib that becomes an important part.

Musicians tend to have cycles in which they start off cold, gradually get warmed up and more confident, and then finally become tired (after which you are unlikely to get a good performance without a break).  The curve I have noticed in singers is faster and more extreme than musicians.  This makes sense since their sound is dependent on their physical and mental condition rather than their ability to produce sound from an instrument.  When a guitar player gets a little bit tired, they can still hit the string in a way to produce the same sound as before (the guitar itself feels no fatigue).  A singer however will start to sound weaker, and are far less likely to perform the same exact way each take.

In analog days when tracks were limited I tried to make sure that I captured the singer as they were approaching their peak, reaching it, and going a little past it.  If I was using a vocal slave (or digital) then tracks were not an issue and I could just keep recording.  Unlimited numbers of tracks is one of the best things about digital, because you can easily capture the entire range of the performance.

You can help the singer go through the cycle and even extend it by keeping the singer comfortable, being sensitive to their needs (both physical and emotional), and catering to those needs.  For example, you may suggest that they drink if they sound dry, or hold off starting the next take if they seem stuck and frustrated (to avoid immediately repeating what is frustrating them), or encourage them to do another take when they are tired but that tiredness seems to be bringing out a certain desirable vocal characteristic.   Singers are not looking out for themselves, they are too busy trying to sing.  It can be argued that this is the job of a Producer rather than an Engineer, but if there is nobody doing these things are you really going to leave the singer hanging, or are you going to try to fill that important need?

You have to be the singer’s best friend and protector.  You can’t fake this, you can only do it if you really CARE about them and the vocal you are trying to achieve.