Human voices are extremely nuanced and our ears are highly attuned to listening to them… yes, I can hear my wife calling to me about the dirty dishes I left in the sink!
The most efficient part of the human hearing spectrum is around 1kHz to 3kHz, where much of the clarity in the human voice resides, and where intelligibility will be greatest.
If you throw away all the other frequencies above and below a range of 1kHZ to 3kHz, it may make the voice sound awfully thin, tinny and scratchy, but you will still be able to understand what is being said or sung, all other things being equal.
Background vocal parts come in all shapes and sizes, so no one-size-fits-all recording policy will work here, but there are some issues that are common to all background vocal sessions. For example, you will need at least one microphone!
You also need to be certain that the emotional and musical aspects of the performances are the best you can get, and that they will work with the other musical elements when you come to mix.
Using pitch and/or timing correction plug-ins does work in almost all cases, but you can also tell it has been done in most cases, and, unless you want the artificial aspects of such processing to be noticeable, you will probably get the best results from getting the real performances as close as you can to the desired outcome.
Those tiny mistakes, those imperfections and differences in pitch or timing which are inevitable in a real performance, are easily the largest factor to my mind in determining the quality and musicality of the result. It’s not about perfection, it’s about emotion.
Background vocals are supportive of the lead vocal and therefore usually appear as unison notes (“doubling” or sometimes “tripling” the lead vocal’s melody line) or as a harmony composed of one or more additional distinct melody lines above and/or below the lead vocal’s melody line.
Often they are very effective as hooks in themselves, due to being framed in a counterpoint or “question and answer” way. Think of classic Motown choruses for examples of these latter approaches.
So, recording background vocals really benefits from due care and attention. And it’s not just the sound quality you have to consider. We will deal with sound quality and technical considerations a little later in this post.
To get excellent results, you really do need to take great care with ensuring you capture the correct rhythm and melody of the parts you were planning for. The musical and emotional aspects are paramount, and the ability of humans to hear voices clearly makes any minor technical flaws in your recording not as important to successful background sessions.
Attention to this type of concern is really a production task, rather than an engineering issue, but all engineers will listen for this sort of problem in order to make the best record they can and to support the producer who may be momentarily distracted or listening to a very different aspect of the music at the time.
It’s easy to miss important musical details if you are not very careful and it usually helps to take very good notes as necessary during the recording session. This means you will be very busy and attentive for possibly quite long periods, but you must make absolutely sure that any harmonies sung alongside the main melody of the lead vocal are genuinely effective and will work in a mix at the appropriate levels before you put the microphone away and move on to another task. Ideally, print out lyrics for everyone, including yourself as engineer, and make sure that you annotate the lyrics to make any notes you need to about which takes are good or which words are to be harmonized and so on.
LIGHTS, WATER, ACTION!
Singers also have a time limit on performance ability, which does vary with the singer in question – although they can be prepared to sing as long as they are physically able to, there is a time when the emotional connection required from the performance cannot be maintained any longer and the “vibe” is gone.
There is absolutely no point in recording anything more from them after this point in time. Take a break, try again later or, more likely, try again on another day. Time out is the only thing that will give enough distance to the singer’s mental state such that they can regroup and try again. Don’t wear out the singers if you want them to sing well!
Another thing about singers – they like decent lighting, which means that they may want none at all, or lots! Whiechever it is, make sure you are prepared to give them what they need.
They often need music stands so that they can see words or even a top line lead sheet (a melody written out as sheet music). This requires enough light to see the content, but you won’t want any buzzing lights, so don’t use dimmer circuits. Try clip-on stand lighting, ideally battery-powered.
They also like room temperature water to drink, not ice cold water that would contract their vocal chords. Keeping the environment they sing in at comfortable temperatures is also a smart idea, of course.
It is also a good plan to have the singers stand on a rug so that the inevitable foot-shuffling noises are damped down, and be careful to keep cables from getting under their feet while they sing. Ask the singers to remove their shoes (and/or jewellery) if a noise issue is occurring during takes.
They will often have their eyes closed during takes, or else will be making eye contact with the other singers (and, occasionally, perhaps the engineer). This means they won’t be looking at the floor and can easily tangle their legs up in a carelessly laid cable if they are moving around a bit.
Also, use a counterweight and/or sandbags as necessary on larger stands with heavier microphones. It’s easy to tip over bigger stands and a lot of harm to people, instruments, precious microphones or rare equipment can result.
Put some kind of padding or fabric over any music stands you use (obviously, under the music), so that reflections arriving back into the microphone are minimized from the (usually) metallic stand.
Give the singers paper, pencils and erasers if they need them, and have a means of finding accurate pitches nearby – maybe an acoustic piano that is in tune, or a tablet or computer or phone with a software app that can provide accurate notes to pitch too.
It seems obvious that you should make sure all amps and other unnecessary noise-producing equipment in your home studio (a fridge, for example, or a furnace vent) is switched off in the room they are singing in. People still overlook this requirement though, so do consider it before you start doing takes.
Make a note to turn on the fridge/freezer or heating again after the session, or you may ruin a lot of nice pizzas
Once you have recorded all the singing, it’s a great idea to leave everything set up, and have the singers leave the room. Then record a take of one minute length of just the sound of the empty room, with everything set the way it was for the singers. This gives you some “room tone” that can be used at the start and end of the mix if it is needed.
Sometimes the sound at the top of a mix can seem too dry, too abrupt, when the songs starts or ends. This can often be unnoticed until mastering is performed, usually many weeks later than the recording was made. Room tone will save you from this problem.
Room tone allows you to have a simple way to solve that problem. Fade in (or out, as needed) the room tone to cover the abruptness that is bothering you. There are even plug-ins that model room tone, such as the Ocean Way Studios plug-in, but this is not the same thing as using the actual room you have recorded the vocals in. You need to use the same mic and space emptied of singers to get genuinely useful room tone for the purposes described above.
With multiple singers, it really helps them achieve a good blend if they can hear themselves acoustically in the room. Another thing that really helps is to prepare a mix of the music ahead of time so that it sounds as close to complete as possible, including a scratch lead vocal if not the final lead vocal.
Also, using a separate microphone for each singer is very unlikely to result in a good sound. The blend of voices is best done in the room by the singers themselves, and your job is to capture it in a flattering and musical way.
Ideally, you want all the singers to use headphones and you want to remove any chance of spill from those headphones into the open microphone(s) as effectively as possible, but often the best singing will result from allowing them to keep one headphone ear-cup off. If that is the case, it’s smart to make the music bed a mono mix, and sent it only to one side of the stereo. All singers will need to use the same side of their headphones for this to work.
Make sure the silent side of the stereo headphones is the side that singers remove, and there will be no spill from that side.
Where there are two voices, you can put the two singers on either side of a binaural microphone, usually a ribbon such as an AEA R84, or else a suitable condenser that offers a figure-eight pattern setting such as a Neumann U87. Where there are more than two singers, you will get the best results from an omnidirectional pattern instead.
Often, the singers with a lot of experience will use their “microphone technique”, which refers to how they address the microphone in terms of distance and angle. Better vocalists (in a technical sense) will usually back off for loud notes, or change their angle of approach relative to the microphone, perhaps singing across or above the diaphragm rather than directly into it. It takes time to learn these skills and using them can make it harder for the singers to perform without distraction, and with appropriate feeling, until they are used to the techniques, but they are well worth acquiring.
A good average performance distance is about a foot for multiple singers, because that gives them room to move back and forward and sideways to some degree without causing too drastic a change in sound.
The nearer they get to the microphone, the more your recording will be impacted by changes in how they address the microphone. This is yet another good reason to choose an omnidirectional pickup pattern for backgrounds. Omni patterns do not exhibit bass-boost (“proximity effect”) characteristics as you get within a few inches of them, unlike cardioid-based patterns, which do.
It’s also a good plan to avoid drastic height differences between the singers. If one is much taller than another, you should try to find boxes or benches for the singers to even up their height differences. The microphone capsule is of course at a fixed height position, and this can cause problems when singers’ heights vary significantly.
When there are lots of singers, say five or more, you will get the best results from two microphones, both in omni pattern, but you can also use a stereo spaced pair, if you are careful with positioning singers.
Alternatively, the horseshoe arc is a good shape for a lot of singers placed around an X-Y positioned stereo microphone pair.
A useful approach to the process of doubling or tripling background parts, often mentioned in regards to Michael Jackson’s producer Quincy Jones and his engineer Bruce Swedien, is to have the singer(s) change position relative to the microphone with each new part, while leaving the microphone position alone.
Obviously, getting too far from the microphone will result in increased levels of room sound relative to direct sound, so bear that in mind as you decide how far back to move singers. Real room sound is also more effective than artificial room sound, which is another nice reason to try this technique. It sounds different than keeping the singer in one place for each take and varying the applications of reverb on each part in the eventual mix.
After the first pass is completed, have the singer(s) move back a foot and sing the same part a second time at this new distance. Leave everything else the same, other than any necessary gain structure amendments. You may not need to change gain structure at all, and ideally will not. However, the “noise floor” usually is an issue when recording at home, and for this reason you may need to bring up the input gain a little for the most distant takes.
The different balance of “room sound” (ambience) in relative proportion to the “direct sound” (vocals) present in the two parts will show in the recorded result when they are combined, and give a greater impression of space and depth, and help keep the vocal parts relatively distinct within the blend. The added complexity of the natural differences in spatial character will help keep the listener interested in what they are hearing.
There is a point where the levels of ambience are equal to the levels of direct sound, and it would be most unusual for you to veer too far into the ambient territory, since the vocal (the direct sound) is of greater interest than the ambient sound field in the room. After the room sound is more than 50% of the total sound, things are likely to be too cavernous and a little too indistinct.
Another good thing to do is to stick with the section of the song you are working on as you go through the parts, because the tracking will be much tighter that way.
For example, if you are tripling a chorus to add harmonies, then stay with that specific chorus and that specific harmony, until you have successfully captured what you need with all necessary takes. Once you have three “keeper” takes of the low harmony, move on to capturing three “keeper” takes of the mid-harmony, and then three of the high harmony.
The singers will find it easier to lock onto what has been recorded, and to their own parts. usually, they’ll be louder when they start to track a part they already captured once. Be prepared to lower the gain a little on the microphone for the second and third takes on a particular part as the singers grow in confidence and sing louder.
Once you have the all-important choruses nailed, only then move on to the verses and any other sections needing background vocal parts.
It’s tempting to record one chorus perfectly, and “fly” that into the other choruses in the song, but there are two disadvantages to this approach. First, the sound of unique choruses is more interesting to any listener. Secondly, there might be a key change (a modulation) during the course of the song, and that would make it much more problematic to fly in earlier parts, because you would have to pitch-shift the parts you are flying in to suit the new key.
Thankfully, you will know ahead of time if there is a key change, since you will be familiar with the song, so you can be prepared to record exactly what you need in each part of the song, and not be caught out by thinking you will fly in vocals to the last chorus and remember too late (after the singers are gone) that the parts are all a minor third higher than they were in the earlier choruses. It happens.
Sibilance can build up in multiple tracked vocals quite easily, so ask singers to understate their “esses” by turning very slightly away from the microphone as they sing the “ess” sound if it’s very noticeable. This may distract them from giving their best performance, but experienced singers usually have no trouble helping in this way and still delivering a great take.
The Beatles would famously “flash” their esses, by putting their hands in front of their mouths very briefly as they sang “esses”, but this is a difficult technique that few have really practised enough to make it as effective as simply turning away a bit. The same goes for the plosive consonants like “p” and “b”, which can easily cause loud bass “bangs” at the microphone input.
Using processing and DAW automation can remove most or all of these after the fact, but it is not always easy or even possible to repair these issues, so preventing them entering the mic is the best plan, and this involves the singer turning away a little at the time they make the sound.
The human voice sounds great on most microphones, but for backgrounds you probably want either a condenser (one that has an omni pattern and/or figure-eight pattern available) or a ribbon microphone (which will be figure-eight). Dynamics are rather unusual in this application.
The ribbon is generally warm and smooth, due to less high-frequency sensitivity, which is a problem with background vocals, whereas the condenser is typically bright and sounds way more intimate since it captures the breath sounds and lip-smacks and the like with far more clarity than a ribbon. Ribbon microphones are also very sensitive to blasts of air and need more care in use to avoid damaging them, and you won’t want to get too close to a ribbon for these reasons.
The mic of choice, then, is likely a condenser. Pick one with an omni pattern selector, and it will usually also have a figure-eight pattern option. Either of these can be useful with background vocals, and the more singers there are, the more likely you are to use the omni pattern. Omni patterns don’t suffer from proximity effect, as noted above, so maintaining distance to singers is a little less of a concern.
Bright microphones also pick up more of the room sound, which is a good thing in sensible amounts. You might choose a small-diaphragm condenser, but they will be more sensitive to the angle of attack of the vocalist than a large-diaphragm microphone. On the other hand, the smaller diaphragm makes it less sensitive to “boxy” or “hollow” sounds when the singer does move noticeably off-axis. Personally, I always use a large-diaphragm condenser such as a Neumann U87, and I love the results I get.
It can be a very good idea to use a different microphone for the background vocals than the one you choose for the lead vocals. Often, this will assist in keeping the individual vocals more distinct, and helps to clarify the lead vocal when it is supported by harmonies in the mix. Most of the time, the lead should remain the lead even when blended in with a lot of backgrounds.
A pop shield or windscreen may seem like a good idea, but I find they darken the sound of vocals more than I like, and if background vocals are performed from a foot or so away, they are likely to be entirely unnecessary. Better to address a microphone slightly off-axis, in my opinion. Clarity is reduced by a pop shield, and if proximity pops are not an issue it makes little sense to use one.
You may find you get the best results from the “Starbird” type of large boom microphone stand, or similar stands. These allow you to suspend the microphone from above the singers, usually hanging the microphone in front of their mouths about a foot and up slightly above eye level, looking down at their mouths at a slight angle. This minimizes pops, and allows for good eye contact between singers. Again, make sure the stand is weighted down so it can’t fall.
BACKGROUND VOCAL POST-PRODUCTION OPTIONS
Before finishing up, I should mention that there are plug-ins that can align the timing of multiple background vocal parts, under automation, such as VocAlign software. Others, like Melodyne or Auto-Tune, can correct the pitches. These techniques can also be applied manually, one track at a time, for greater control of musical expression.
Naturally, the best results are probably obtained by getting it right at the time of recording, rather than tweaking after the fact, but it’s great to know you can save the day later with software, since that increases the confidence at the time of recording that you needn’t be too freaked out if something is proving difficult to match perfectly.
Syllables can be adjusted in timing and pitch later on if the singer simply can’t “get it” to match his previous take(s). It’s not the end of the world these days, and artificially adjusted parts can be virtually indistinguishable from accurately sung parts!
Well, that’s about all for today, so do join me again in a few days for my next post on “Recording Lead Vocals”. Until then, I recommend listening to your favourite recordings to make mental notes on how the artists handled the arrangement of their backing vocal parts. A lot can be learned from careful listening to those who do it really well.
Join me again soon!