In this article, we’re going to give you some solid, concrete tips for mixing background vocals. Six of these tips will consist of specific techniques; the last two are philosophical. All of them have helped me get my background vocals to gel, and more importantly, to get them sitting in the mix just how I want them to.
Let’s dive in!
As with any technical concept, we’re going to start with the basics: if you want polished background stacks, you’ve gotta edit them! Make sure to align them all so no consonants are out of whack, no one vowel outlasts another, and no single sibilance stretches past the rest. This can be done with software like VocALign, but really, you ought to do it by hand. Not only will you have total control, but you’ll be able to learn a lot about editing in the process.
Editing includes chopping out breaths, cutting out bursts of extraneous noise, de-noising if necessary, color-coordinating, and sometimes—provided the part repeats many times in the tune—flying in good notes from later passages to replace the bum ones. Don’t discount this last tip: it may feel like cheating, but no one’s going to notice and it saves you so much time.
Of course, if the artist explicitly doesn’t want you to replace the errant vocal, you can’t. Either way, you should still edit your background vocals, preferably before you start mixing heavily.
Vocals—even supposedly flawless doubles—are both incredibly dynamic and incredibly sensitive to processing: bad compression is particularly offensive on vocals.
Yet, vocals often need dynamics processing, especially in a stack of background parts. All too often, one wayward vocal will suddenly poke out and draw too much attention. So what do you do?
I use the Auto Level Mode in Nectar 3 Plus. I’d be telling you this even if you magically transported me out of this article, sat me down at a coffee shop, and asked me point blank. I’ve said this before, but I’ll say it again here, in the words of my people: If Nectar only had that tiny Auto Level Mode button at the top of the screen, dayenu!
This button engages just about the smoothest auto-leveler I know of, transparently adjusting the level by +/- 3 dB of the target, which you can set on the input slider. If you sing louder than three dB over the target, it will only restrict the results three dB.
This helps keep the process natural, but it means it won’t be your only line of defense for evening out a vocal—compression and automation will often come into play. Still, it’s a hell of a first step.
Imagine a crowd of vocalists, each one singing, “somebody sold me some substandard soup.” Yeah, those are really stupid lyrics, but they demonstrate the potential of sibilance to harsh your mellow, as that sentence sports a sinful sum of sibilance.
If you were only dealing with one vocal, you might be able to get away without de-essing, depending on the singer, and how the singer was miked. But once you’re dealing with four to six people all singing the same sibilant line, you’ll probably want to reach for a de-esser.
There are many great de-essers on the market, and iZotope makes two that are quite useful—the one in RX, and the one in Nectar. We also have an incredible deep-dive on how to properly de-ess using these tools.
The De-esser in Nectar offers a wideband operation, so it will compress the whole signal when it detects a sibilance, rather than split the vocal into a multiband signal path. I find this to be a more natural approach, as the processor won’t alter the tone of each vocal as it hits the threshold.
The operation itself is simple: loop a passage with some offending sibilance, and click the “ear” icon (the Listen button) in the GUI. Now you’ll be able to isolate the range of frequencies indicated by the detection filter, illustrated below:
Tune the filter until it’s grabbing the most offensive sibilance you can hear, and click the ear icon again to get out of Listen mode. Now, pull down on the threshold control until the esses are to your liking. I like to go overboard on the threshold control at first, so I can really hear what it’s doing. Knowing what “too much” sounds like is key to getting to a goldilocks sound.
De-essing your background vocals this way, naturally, at the beginning of your process, will help them feel more professional, and will help situate the vocals in the mix.
Too much processing will sound overwrought, flat, and terrible by the end of the mix: many are the mixes I’ve heard choked out by an overabundance of processors. So, if you can avoid doing too many things on a track-by-track basis, that’ll be good for the long run.
You can avoid overprocessing by not using compression on your individual vocal tracks, if you can help it. Consider that the effects of compression are cumulative, and there could be a long way to go for your vocal: it may be compressed in the background stack, again in the vocal bus, again on the stereo bus, again by the mastering engineer, and—if you’re lucky—again by the radio station.
You’re avoiding compression here, but you’re not helpless against dynamics control: the Auto Level Mode module will help, and you can also clip gain specific regions of the vocal that are too loud or too quiet. There’s also automation, which you should accomplish with a gain or trim plug-in, rather than with your DAW’s volume fader. You don’t want to be locked in to track-based automation this early in the mix.
If necessary, it can be helpful to EQ each vocal—but I like to not go overboard with this. Never to solo any single background vocal unless you’re actively searching for a problematic resonance, or some other issue that needs intense attention. For congruity, listen to all the backgrounds, panned as you’d like them, while you EQ each one. Force yourself to listen to the effect your one choice has on the entire stack.
In fact, don’t just solo the background stack: EQ the vocals against as much of the production playing as you can muster—certainly keep the lead vocal, and hopefully the rest of the instruments as well. You can lower their levels, sure, but please don’t mute them outright, unless you need to go after a specific issue that’s hard to hear in context.
I’ll often high-pass a vocal, and then limit myself to one boost and one cut, all with the purpose of helping the background vocals to sit right as a unit. I might wind up doing more, but this method is a great way to get started:
Try one boost with the idea of helping the vocal stack feel more thick, or more ethereal—more whatever adjective you’re going for. Then try one cut, to make the whole background stack sound less harsh, less dull. Again, pick an adjective you want less of, and use a cut in the vocal to try to get there.
Why do you think Auto-Tune, Melodyne, and other such applications have become so ubiquitous? Many moons ago, I pondered this topic in a reviled Forbes piece, the only piece of writing that ever got me death threats (what can I say, people really love Arianna Grande—and really hate clickbait titles!). But here’s my thinking on why pitch correction is so popular from an engineering standpoint:
You see, pitch correction has a peculiar, ancillary effect—it doesn’t just tune the notes, it also aligns the overtones to sound more congealed. This process can often sound “rounded” or “warm”. Used not for correction, but for effect, it’s almost like modulation, and it can contribute to the sumptuousness of harmonies, particularly if they are pads.
If your background vocals are oohs, ahhs, or pads of any kind—and your arrangement isn’t some sort of rough-and-tumble Americana roots rock—consider using pitch correction on each track as a modulation effect. It can add a wonderful fullness.
Here are some background vocals, barely edited, and completely unprocessed. Now, I’m not saying they’re perfectly in tune, but they’re not horrendously pitchy either:
Now listen to what happens if I give them a touch of slow, subtle pitch correction:
Can you hear what I’m talking about? They’re a bit swimmy, like with modulation, and definitely atmospheric and warm.
It’s been years since I didn’t route multiple background vocals to a shared bus, if only to control their level in one fell swoop. But often this serves a deeper purpose: the compression I told you to avoid on each track? Well, the bus is a great place to play with compression. Not only will you have a chance to apply some dynamics-taming to the vocals, you’ll be able to infuse them with character from the compressor.
By way of example, let me take our background vocal stem and run it through some compression algorithms in Nectar, so you can hear the differences. Here’s our stack of background vocals:
Here they are played through the optical compressor in Nectar 3 Plus.
And here they are played through the vintage compressor.
Each one has a different character, to be sure. And either might be right for the song. Paired with other elements, such as saturation, flanging, delay, and verb, you can get a characterful impact, if you want it:
Because we’re applying an effect to them all, we’re getting a vital glue, rather than the constricted, overly-tightened, stiff sound of each track compressed on its own.
Your last philosophical tip is to make sure to have genre references for how you’re mixing background vocals. This will help guide you to the right place. Nothing works like a reference for focusing your attention.
For instance, I was recently mixing a tune called “Mark Twain Motor Inn” for Travis McKeveny & The Famous Dr. Scanlon Band. Pete Mancini was playing on the tune, arranging it as well. Because Pete and I have worked together for many years at this point, he always knows that I want references—and he was very clear with the background vocals: he wanted the background to sit just behind the lead, the way it does on “Loving Cup” by the Rolling Stones.
Because I had this reference, I knew exactly what he wanted. I knew not to double the background vocal and pan them hard left and hard right, but to keep it centered, just behind the lead.
If you are working on your own music, just pick a genre reference that you enjoy: it will help.
I’ve left this for last, and what a shame, because there’s so many ways to pan a stack of background vocals! Indeed, that’s exactly why I left it for last: it’s so track-dependent that I can’t possibly give you concrete rules of thumbs for this.
Or can I? Come on, of course I can:
Go “asymmetrical” for interesting: things that aren’t quite balanced stand out more. Things that stand out more have the potential to be interesting. If the song you’re working on is quirky, don’t go with a conventional panning scheme: do something quirky: put one vocal up the middle and the rest to the left—it could be just the flare you need to sell the song.
Go symmetrical for atmospheric: note that I did not say “go symmetrical for boring”, because this is not what happens when you pan a vocal to be symmetrical. Instead, when you pan a vocal symmetrically, I find it sinks back into the tune in a manner that influences its atmosphere. Take those backgrounds I showed off earlier, the ones that sounded like pads. They were panned like this:
When I have background vocals that are basically pads—oohs and aahs that are all doubled and layered—I do have a panning scheme I try out first. I’ll relate them as though they were on a clock:
Notice how I’ve balanced some of the doubles: two sets of doubles don’t mirror each other exactly, but their configurations do. I find this is both interesting and atmospheric. Try it out—it could very well work for you for atmospheric pads. Or, maybe it won’t work out, but in trying this scheme, you’ll find another that becomes a good starting point for you. That is also good!
That’s about all the space we have for covering how to mix background vocals in this article. Doubtlessly, there’s more to cover. But for now, this should be enough to get you started in the rewarding process of making your background vocals sit right.