The Inner Game of Vocal Production
Learn How Top Producers Get the Best Vocal Takes
Perhaps no other part of the recording process requires quite the level of delicacy, tact, experience, and expertise as capturing lead vocals. Think about it: the guitarist may bemoan a fluffed note here and there, the drummer may whiff a fill, but these are fixable, and surely do not bring on the kind of self-flagellation that besets a lead singer when the takes simply aren’t happening, and the clock is ticking.
Here, three top engineer/producers give you some tips and techniques for managing this careful dance of sonic techniques and interpersonal skills you need to coax an inspired, transcendent vocal take.
Meet The Producers
Currently riding high from his production work on Post Malone's smash hit, Beerbongs & Bentleys, Bell has also crafted hits for the likes of Cardi B., Camila Cabello, DJ Snake, and more.
Featuring one of the more extensive resumes you'll ever see, Bendall started at Abbey Road in 1973 and went on to work with everyone from Kate Bush, Paul McCartney, Alan Parsons, XTC, Peter Gabriel, and hundreds more. He's also done extensive soundtrack work, most notably winning a Grammy for The Last Emperor in 1987.
Perhaps best known for his work with Beck, Meldal-Johnsen's production credits include M83, Jimmy Eat World, Tegan and Sara, Paramore, and post-punkers, Moving Units among many more.
Preparing For The Session
Haydn Bendall: Recording a singer is a very private affair. It’s almost like a confessional. The artist is basically naked in front of you. Unmasked. And for that to be comfortable, there has to be a real element of trust.
Ultimately, it’s about being aware and being sensitive to everyone around you, and not only the singers, but to the musicians and studio staff. You must have eyes and ears everywhere, all at the same time.
Justin Meldal-Johnsen: When you start an album project, you’re “learning” how the singer feels about singing: Is it a job? Is it an enjoyable, natural act for them, or is it an overwrought collection of worries and insecurities? I’m in the mindset of getting the singer to optimize their approach to the album much earlier than when they’re standing in front of the mic and the “red light is on.”
I want people to have all their ducks in a row before they walk in there. So they can focus less on the mechanics and more on having fun and being in the moment with the song, and in the mind of the character behind the song. My job is to create the straightest line to that point.
Louis Bell: Knowing the key of the song, and what sweet spots are available vocally, is an extremely important part of vocal pre-production that doesn’t get enough attention. Everyone has favorite keys they like writing in or playing in, but that can also result in being a slave to the limitations of your musical abilities. If you can only get around the guitar or piano in a couple of keys, well, the quality and timbre of your voice is going to be at the mercy of that key, which may not be the most flattering for your voice or the song.
Setting Up The Vocal Mic
HB: Some singers come in wanting to record with a particular mic: a Neumann U 67, perhaps, or a Telefunken 251. But I often suggest that there are other mics that may suit their voice as well, so I set those all up and we’ll shoot them out and choose one. You just have to reassure them that this is a quick job, and you can make a decision after even just a few seconds of them singing into each mic. Then the artist is involved in the decisions about the sound they’re getting. They’re not just a piece of meat being dragged into the studio and being told, “Sing into this.” They’re really part of the team now, so this is important both sonically and psychologically.
JMJ: The whole process with a singer is predicated on physical and psychic comfort. The microphone setup follows those choices: it could be a hand-held mic on the couch in the control room, or a really pristine tube mic with the lights dimmed with lots of isolation in a booth.
There’s also the famous Bono trick. Allegedly, in the control room with playback pretty loud, he’ll use a Beta 58 while standing relatively equidistant between the speakers. Then one speaker is flipped out of phase to minimize bleed. But really, I’ll try any setup. I don’t worry about bleed. I’ll just cut it out later, once I have my comp. I did the vocals with Paramore with Hayley Williams sitting on the couch, holding an SM7. Point being, if that’s what it takes to get the singer feeling relaxed and ready to sing, I don’t care about having the perfect acoustic environment.
"Recording a singer is a very private affair. It’s almost like a confessional."
– Haydn Bendall
Dialing In The Headphone Mix
LB: When the singer hears themselves as close to a finished vocal production as possible, it has a huge effect on their performance. They have the confidence that they’re going to sound great, and that frees them to get way more into every take. If you take the “raw vocal sound” approach, I think that singers are always wondering if they’re going to sound any good in the finished song. They’ll often say, “Are you going to fix this?” You’ve got to prove to them what it’s going to sound like in the end. This also allows the collective vision of the project to be much clearer to everyone working on the session. It’s not like the producer/mixer is holding out on you.
HB: Some artists like a more finished sound, but I normally try to avoid that. I find it too dangerous. Many people ask for a bit of the EMT 140 Plate in their headphones or the Galaxy Tape Echo, so I’ll use those. That said, I prefer recording vocals with no reverb in the singer’s cans at all. Mind you, I’ll have a reverb in the chain if they ask for it, so it’s all set up. I also record with quite a bit of EQ, but I never compress or limit on the way in.
JMJ: Everyone’s different. Emily Haines from Metric wants her vocals dry as a bone, very present and very exposed while she’s singing. Other people, like Anthony Gonzalez from M83, will be most productive singing in glorious clouds of long reverb and delay — so you have to be ready for anything. I don’t want to be didactic about recording principles with these people, that’s boring. I want them to tell me what it’ll take for them to have fun and sing well.
Peaks And Valleys
JMJ: I use some transparent limiting at least 90% of the time, about 3dB of gain reduction, unless the singer’s mic technique is so good or the performance itself just has less dynamics. Another thing that can work to keep peaks to a minimum is to have the singer stand two, three, or even four feet away from the mic, and listen to it that way. Especially if they’re really powerful singers who are sort of blowing up the capsule, you can say, “Try backing up a bit,” and you find that at two feet, it can sound really good. I’ll add a touch of gain to the mic, and keep going.
HB: If you’re concerned about peaking, turn the microphone down! That’s what the gain knob is for. Once you’ve compressed on the way in, you can never undo that. You may lose the greatest performance of that artist’s career. If the artist asks for compression in their headphones, I’ll offer them a thousand reasons why that’s a bad idea.
Mainly, I’ll tell them they’re more likely to sing naturally if they’re not fighting the compression. You’re much better off concentrating on the singer’s mic technique, and involve them in the recording instead: encourage them to move around a bit in the booth to find the mic’s sweet spots.
LB: I’m kind of my own compressor. When the singer is in the booth, I actually have my left hand on the Apollo Twin controlling the input gain and I just ride their wave; if they’re singing very softly, I’ll turn the input up, and vice versa. So it’s a good system for tracking very quickly, having my vocal chain all set up.
Sure, there are moments where they catch me off guard with a vocal, we get a peak, and we might do that line again if it’s not a keeper. But for the most part I can sense where they’re going, especially if I already know the melody. I can adjust as we track. I think it helps the final product be more consistent.
Guiding the Performance
LB: Usually we start with a very basic musical idea, just the drums and bass or drums and a chord progression — and I’ll loop that for about 15 minutes. And generally, within that 15 minutes, as Post Malone freestyles, we’ll have the entire song together melodically and phonetically. I have a talkback mic so we can throw ideas back and forth and suggest tweaking this note or that. It’s like he’s the one riding the surfboard, but I’m on an earpiece helping him ride the wave. That first pass, that first listen, is so important, because you can only be surprised by a piece of music once. While the singer or rapper is still in that zone, on that melodic high, you’re just there to give them the proper adjustments.
HB: I will go into depth with the singer about phrasing, and the lengths of notes, where the breaths should be, which words want to get more weight, and which are less important. When I’m engineering, I’m also producing at the same time, and part of that is finding the emotional content of the words. With vocals, there’s often not much of a difference in the manner in which you do it. It’s obviously great if the singer is the songwriter as well, because then you can really ask if that lyric is as powerful as the melody that goes with it. Or, perhaps there’s a lack of connection between the lyric and the melody there, and they can make edits.
JMJ: Keep in mind that the singer may be following their muse, and something may be going down that’s really good, something that you may never get again, and you want to help that direction unfold. So you need to have your technical shit together, but stay out of the way, don’t interrupt the flow, and get it in as few takes as possible.
You do a disservice to the singer/lyricist by demanding an intimate breakdown of the meaning and intent behind every lyric while you track, especially if they’re flowing. Once the lyrics are on the page, the time has passed for extensive analysis. It’s now about being just technically adept enough, on both sides of the glass, to get the desired result.
"I don’t want to be didactic about recording principles with a vocalist, that’s boring. I want them to tell me what it’ll take for them to have fun and sing well."
– Justin Meldal‑Johnsen
What To Do When Confidence Wanes
JMJ: If I need another take or two, I just need to read the person, and either be a full-on cheerleader, or be subtly and quietly encouraging. They don’t always need someone being all “rah-rah!” about it. This is more about paying attention to the nuances of human behavior than following a script. You have to respect the viewpoint of the singer, and depending on the song, that can be anywhere from super weighty and deeply considered to a singer who just wants to casually throw it out there, y’know, “let’s just sing the damn thing!”
HB: What I won’t do is have them carry on until their voice gets tired, which is pointless. I’ll reassure them that it’s not about being lazy; we can carry on until 3 o’clock in the morning, but I promise you, we’ll get nothing. If you take a break now, even for a half an hour, or maybe till tomorrow, you’ll come back stronger mentally and physically, and you’ll sing it better.
And the thing is, even if the takes aren’t great, you’ve still learned a lot from doing it. You’ve invariably made great progress with the phrasing, you’ve gotten used to the balance in the cans, you’ve connected with the song in different ways, and now everyone in the studio understands what it is that we’re all after. It’s a balancing act — for some people it could cause a real confidence crisis if you don’t put it in the right words.
Keeping Track of the Takes
JMJ: You really need to keep detailed notes. I have an Excel spreadsheet, which I customize on a song-by-song basis, with all these different cells, and across the top I have the take numbers, 1-10, or whatever, and on the left I have rows of lyrics, of lines of the song, broken out as succinctly as I can. These can be up to four pages long, and I’ll print them out and put in symbols for various things, like the best takes, fix a word, whatever, and maybe an asterisk for specific notes about each line.
HB: I’ve got the lyrics, or the lyrics and music, printed out. And I take lots of careful notes about which takes are the good ones, what may need to be fixed, etc. And this isn’t just for me; it’s a big confidence boost for the singer if they know you’re really on top of it and taking careful, very accurate notes. If you’re able to say that singer, with the notes to back it up, “I promise you, I guarantee you we’ve got this.” And then say, “Now that we’ve got this, let’s try a few more and see what happens.”
Mind you, I’d never tell them “we’ve got it” if we haven’t got it. But for me, one of the most damaging things a producer can say is, “Give me one more like that.” What’s the point if you’ve already got one like that? There needs to be a good reason to do it again, and you need to communicate that to the artist.
"While the singer or rapper is in that zone, on that melodic high, you’re just there to give them the proper adjustments, and stay out of the way."
– Louis Bell
Playing it Back For the Singer
JMJ: If a singer is prone to self-criticism, I might dress up the vocal comp so it sounds a bit more finished before I play it back for them. Perhaps I’ll spend a few minutes on some clever effects design and do more dynamic control, rather than just using simple verb they were listening to in the cans.
In getting to know the singer, I’m going to want to learn how sensitive they are when they hear themselves back. And if they have insecurities, it may be better for them not having to imagine what it could sound like with a good mix, but to hear it in a 75% finished state.
HB: For continuity, I like the final comp to have big chunks of the original takes, rather than a word-by-word edit. But certainly if I have a great line, but something like the end of a particular word is irksome, I will go in and do that more surgical edit. I care very much about tuning, but I very rarely use Melodyne or Auto-Tune, or any of those things.
I’d rather spend five hours with the singer getting the takes we need, than a half-hour with the singer and five hours trying to tweak Auto-Tune. It just requires more direction. You can spend weeks if you want tuning and timing a vocal, and it doesn’t make the vocal any better! It only makes a bad singer sound more in time and in tune, but it has no emotional gravitas at all.
LB: You’ve got to believe in the lead vocal. You don’t want to do too much trickery to it. You can lose some of its personality if you start subtracting too much with EQ or dressing it up too many with effects.
The worst is when a singer says, “Let me take this home and listen to it.” That’s the worst. You really don’t want a singer taking a rough mix home and listening to it a million times, because they’re only doing that in order to find something wrong with it. And they will. There has literally never been a time when a singer went looking for something wrong and came back saying, “You know what? It’s perfect!” Never.
— James Rotondi
How to Record Auto-Tune without Latency
Learn how to get the sound of modern pop and hip-hop vocals with near-zero latency using Auto-Tune with Apollo interfaces.
Best UAD Plug-Ins
Here’s our list of best-selling UAD plug-ins, and the top 10 that you should own in 2022.
The Best Compressor Plug-Ins
Learn the best compressor plug-ins for each element of your mix, with insight from the industry’s top producers and engineers.