Song writers, DJs and producers hoping to secure record and publishing contracts will at some time need to master all the techniques necessary to record a singer. The plain truth is that whilst many industry personnel can hear the potential in a mediocre performance of a great song, most cannot and you need to give yourself every chance.
If you are using vocals they will almost certainly be the most important element of your recording. Capturing a great performance is both a technical and psychological challenge. No other part of the recording process is as difficult but with a little care excellent results are attainable even from modest equipment setups. (For the purposes of this article we will ignore spoken word and live recording and concentrate on recording a single singer in the studio).
For those of you wanting to record vocal "sound bites" to be sampled, processed and used in dance tracks all the same rules apply.
Organise your recording day around your singer. At what time is he/she likely to be at their best? Many singers won't perform in the mornings but don't think this means you can keep them hanging around for hours whilst you perfect your Steve Vai solo. Be prepared to record the moment they arrive and leave yourself plenty of time.
If you have a range of mics set them all up and compare the results before resorting to EQ to cure problems. Although your best studio condenser mic will probably work well with its wide frequency response and high sensitivity, allowing it to capture all the details of the sound, a humble dynamic model may suit some voices better particularly when the voice is powerful or sibilance is a problem.
Use a pop shield. Home made ones will never be as effective as professional models but are better than nothing. If popping (Bs and Ps) is a problem have your singer sing slightly over or under the mic but remember off-axis the sound may become duller. Mount the mic on a boom stand, preferably in a suspension mount, to eliminate handling noise.
Sibilance (a high frequency whistling which accompanies Ss and Ts and which can be accentuated by studio mics) is a problem that many engineers struggle to eliminate during mixing by processing the recorded vocal with so-called De-essers (actually a type of compressor designed to attenuate sibilant frequencies). I have always found that prevention is better than cure and worked to eliminate it with mic positioning techniques and pop shields.
Most cautious engineers (including me!) prefer to add processing and fx in the mix. Record the vocal dry and flat with a little Soft Knee Compression. Compression is necessary to restrict the wide dynamic range many singers exhibit and squeeze it onto a recording medium like tape. A Ratio setting of between 3:1 and 5:1, and a Threshold setting allowing for about 10-20db of gain reduction during loud passages usually works well. If your compressor has an Auto feature use it to help control the Attack and Release/Decay characteristics otherwise set these controls to around 1ms. Do not over-compress or Limit the vocal during recording unless you want to acheive an intentionally unnatural effect.
Because of changing vocal dynamics over the course of a session (noise) gating the vocal to tape/disk is not usually considered a good idea. If you have problems with noise or spill, track them down and eliminate them at source.
Use EQ and FX on the monitor signal coming back from tape where you have a non-destructive opportunity to make as many mistakes as you like in the mix. You may find it necessary to boost a little treble on the mic input channel but remember, if you are using a good vocal mic, greater minds than yours and mine designed it that way for a reason. Put some reverb on the monitor signal. Singers love Lexicons but whatever you use make it natural sounding.
There are an increasing number of so called "Voice Processors" available. These bring together several processing components into one box. In general they are a good idea, letting you bypass the mixer and go straight to disc/tape. Use their processing features sparingly, it's easy to spoil the sound of a good mic.
It has long been common studio practice to create a separate mix for the foldback on one (or two for stereo monitoring) mixer Auxiliaries. However if you send the mixer Master outs (or mixer or audio interface headphone socket) direct to the Foldback amp both the engineer and the singer will be hearing a good mix of the backing track with fx in place. A more complete sounding backing track may inspire a better performance. Remember to push the monitor level of the vocal you are recording significantly higher than it will be in the mix to attain a reasonable level in the headphones for your singer.
Using good headphones will help your singer attain an acceptable aural perspective. This is crucial. Singing into a microphone and hearing your voice coming back from outside of your head is a far from ideal way to record but for obvious reasons we're stuck with it. Although a studio standard, Beyer DT100s are not really suitable. They are designed for durability rather than sound quality. Try using a good pair of hi-fi phones. They must of course be a closed design. I use Sennheiser HD250's and they relay the kind of detail needed.
Check that the headphones and mic paths are working correctly yourself and remedy any failings before your singer arrives.
Controlled room ambience is a wonderful thing but you only have one chance to get it right. Unless you are intimately acquainted with the reverberation characteristics of the room you are recording in and are sure of your vocal mix strategy I would recommend selecting a mic with a cardioid response allowing you to record the direct sound only. In theory an omnidirectional response will give you a more natural sound, if this is what you want, but you will need to record in an acoustically dead environment such as a vocal booth.
You will need a gap of between 8-10 inches between your singer and the mic. The best way to achieve this is to place a pop shield about 8 inches away from the mic and tell them to get as close as they like without touching it. If you need to position the mic off-axis the pop shield will act as a visual focus for them and ensure that they do not wander more than 30 degrees off-axis where the sound will change.
Place a music stand with the lyric sheet on it at eye level and if possible in such a way that your singer does not have to turn their head to read it. In general performance sung from memory are better than those read but unless your singer has written the lyrics themselves this may not be possible.
So the singer has arrived, the kettle activated, and loose jangley metal objects removed from their person. If you have not done so already and are using sequenced backing you need to find the right key. A word of warning. A singers range can expand dramatically as they warm up so get them on the mic running through the song. It is intensely irritating to be continuously interrupted when warming up so let them sing through the song three or four times while you make the coffee (what do you mean your sequencer doesn't have cycle mode!). Don't worry if the performance sounds awful to begin with, it nearly always does.
By trial and error find the ideal key. A change of 1 semitone +/- can have a dramatic effect on the overall feel of the vocal. In general your singer will have more problems pitching if the key is too low. Inevitably changing the key will present problems for other musicians but it's a question of priorities.
Solo the vocal and check you are getting no headphone spill. Continually check the tape level until they are fully warmed up. When setting the record level be conservative. Ensure that the mic amp is not being overloaded, then the mic itself and use the mic or desk pad switch to attenuate the level if necessary. Distortion is horrible and often goes unnoticed until it's too late. If you are using a digital recorder err on the sound of caution, once those 16-bits are used up you've had it. Make sure your peak signal rises no higher than -2dB.
Be sure you are both using the same terminology. Rise or Bridge?, Doubling or Tracking?. Don't overburden your singer with comments. Discuss and solve one problem at a time. At all times you must express enthusiasm and voice encouragement even if things look bleak. Singers need to feel good to perform at their best. Another problem is knowing exactly what their best is. Of course you will want to push them as far as you can but decide exactly what is too far as soon as possible.
Decide whether to record lead or backing vocals first. Starting with backing may allow them to warm up but could also exhaust them prematurely.
When multitracking backing vocals, pan completed takes away from centre and lower them in the mix.
You may want to compile a finished lead vocal from a number of separate performances or try to record one complete track by dropping in and out. Record the less stressful parts first. You may find it useful to start with all the verses followed by the bridges and so on. Don't over tire your singer with belting chorus lines and ad-libs early on in the session. They will perform these better when fully warmed up later on.
When dropping in, monitor the singing live prior to the drop in point so they are not coming in cold. Give them just enough lead in time, not too little and not too much. Use the programmable Drop-In/Out and Pre-roll features of your recorder for consistency. It will help you to place a strip of masking tape near the transport controls and write down the tape counter locations for all parts of the song if you are using a tape recorder.
Allow your singer plenty of time and freedom to get it right. Don't pressure them. Seek their opinion and play back the takes when asked. Invite them into the control room for playback every 20 minutes or so. Monitor at moderate to low levels so you don't tire your ears and miss mistakes.
If you decide to compile a finished vocal and are not using a hard disk system (you have my sympathy, I've been there many times!) bypass the mixer if possible and bounce directly between recorder tracks.
Q. What if the singer sings continually flat or sharp?
A. Tell them. Check the backing instrumentation is well tuned. Check the monitor vocal balance. Check that you have a musical sound sufficiently high in the mix for the singer to pitch too. Suggest they place one earphone half on/half off. Pitch/Speed adjust your recorder without telling them. If all else fail forget the headphones and record live in the control room facing towards but as far from the monitors as possible. Reverse the phase of one of your speakers and switch your mixer or amps monitor output to mono. Freddie Mercury recorded like this with a Shure SM58 and so can you with a little care.
Q. What if the singer sings continually out of time?
A. Tell them. Raise the level of the drum kit and percussion but remember to re-check headphone spill.
Q. What if the recording level jumps erratically?
A. Ask your singer not to move around the mic and pull back only slightly for loud passages. Adjust your compressor so it compresses a little all the time.
Take care not to undo all your good work in the mix. Be aware of the dangers of over EQ-ing the voice to make it cut through. Leave it as natural sounding as possible and fit everything else around it. If recorded well it will act as a guide for the entire EQ and gain structure of your mix. Pan backing vocals and instrumentation away from centre to leave space for the lead vocal and lead instrument when it takes over in the solo. Well arranged instrumentation will make it easier for you to place the vocal in the mix. Refer to your favourite CD's often and monitor at low levels to establish the correct balance. Try out a rough mix in the car or on your personal stereo. In general a vocal mixed too high is better than one which is low and indistinct.
To sing well under studio conditions is without doubt the hardest task of all and so we must be sensitive to every detail of the process. Always remember, you may have the finest song, arrangement and instrumentation money can buy but without an excellent vocal no-one's going to notice or care.
None at present