Tracking lead vocals: We demonstrate why the element that demands most listeners’ attention also demands most of the producer’s...
I first came into contact with Thomas Stevenson when he emailed us (email@example.com) with a song he’d been working on for his band Nerve 9: a power ballad called ‘Pray For The Rain’, which set lead singer Kelly Orr against a classic rock backdrop of drums, bass, and electric guitars. Listening to his latest mix on my own system, I agreed that the mix sonics could be made more expansive for this kind of song, but what also struck me was that the lead vocal didn’t yet seem strong enough to carry a production like this. While Kelly certainly had a lovely voice, somehow the melody and lyrics weren’t giving it enough room to shine, and I also felt that more could have been done at the recording and editing stages to achieve that sense of 100-percent emotional commitment that’s so essential for any ballad.
Fundamentally, like many recording musicians, Thomas had fallen into the trap of giving the lead vocals little more production time than any other instrument in the arrangement, despite the fact that the singer generally demands the lion’s share of any mainstream listener’s attention. But he also admitted that he wasn’t confident of how to manage a more in-depth approach (even if time were no object), so I suggested all three of us collaborate on redoing the lead vocals together, so that I could provide some real-time tips and help them refine their results. One large complication was the Atlantic Ocean, which lay between me and the band (how inconsiderate!), so we decided to try using Apple’s FaceTime video telephony service to bridge the divide.
Firstly, I worked with Kelly to refine the vocal part itself, setting up my iPad close to my studio speakers so that I could play her sections of the song as necessary. Thomas had sent me both full-mix and instrumental versions of the song to help with this. After our FaceTime session, we continued batting vocal ideas around between us as MP3 demos via email.
My main concern was that the lyric occasionally felt ‘overcrowded’, with some lines using more words than was necessary. This rushed Kelly’s delivery so that she had less freedom to apply her own vocal inflections. So, for example, we agreed to change her first-verse line “and your smile told me all I was was one day”, to “your smile told me I was one night”, maintaining the essence of the line’s meaning without the ‘bustle’ of the original recording. Likewise, the second and third words of the chorus line “to come and wash it all away” seemed superfluous, so were removed. Many other lines were similarly pared down.
Another principle I discussed with Kelly was taking inspiration from natural speech to make her sung performances more emotionally direct for the average listener. For instance, the line “you won’t be there” forms an important part of the song’s message, but the critical word “won’t” had been skipped over on a sixteenth note. When Kelly spoke the line it was clear that lengthening the central word made more sense.
The final thing I queried was a more aesthetic issue: while the sentiments of the two verses seemed believably personal, the middle section took a more abstract and ambivalent tone that I had difficulty relating to as a listener: “One day there’ll come a time when you’ll want more than you know / When you’re bored and tired from the emptiness you’ve known / And you’ll feel like crying loud / And you’ll feel like trying hard / You’ll be searching for some meaning and the need to find what’s true.” Between us, therefore, we hammered out various changes:
- We brought the action into the present by changing “one day” to “now”, making it a fulfilment of the second verse’s “you’re gonna get what’s coming your way” prediction. This past/present relationship was further underlined by switching “than you know” to “than you’ve shown”.
- Kelly replaced the second line with clearer images of abandonment and humiliation (“There’s no-one waiting. Your empty past is known. There’s nothing left. Regret it hits real hard.”), before closing the section by pointing the finger of blame more explicitly (“You’ve pulled this all upon yourself. You can’t erase it.”).
In addition, we decided to reinforce the narrative progression we’d now introduced between the verses and middle section by altering the final chorus lyric. So where the narrator herself prayed for the rain during the first two choruses, the villain of the song is then put into the same position later on as a kind of poetic justice.
We were now ready to schedule a tracking session. Before the big day, however, I suggested Kelly rehearse the new material thoroughly on her own, so that we could focus more on matters of interpretation while recording. Many musicians think nothing of spending hours practicing an instrument, but don’t apply the same thinking when trying to improve their singing, so I urge vocalists to redress this mismatch whenever I can. (Apart from anything, expanding the boundaries of your own vocal technique can be a pretty grisly sounding activity, full of unintended squeaks, cracks, and gurgles, so most singers won’t push through with it sufficiently unless they’re well out of earshot of anyone else!)
One trick I also suggested to her in this context was to practice singing quieter in general, especially during the verses and middle section, because I felt that her voice was sounding too strident and pure in those sections — the luxury of studio recording is that you can always turn up the fader! To help with this, I suggested that she rehearse with a hard surface (such as a clipboard) right in front of her to reflect more of her sound back to her ears. Not only does the improved acoustic feedback naturally make you sing quieter in general, but it also makes it much easier to hear and control the high frequencies of your vocal tone.
Kelly wasn’t the only one with homework to do before the session. Thomas and I were putting together a comp sheet (to keep track of all the takes) and liaising about how to manage the mechanics of the session. The analogue connections were handled by his PreSonus Monitor Station monitor controller, which created a shared cue mix for all three of us from a combination of backing-track and vocal monitoring signals (from the computer DAW recording system), Thomas’s talkback signal (from the Monitor Station’s talkback mic), and my talkback signal arriving via FaceTime. This cue mix was fed from the Monitor Station’s individual headphone outputs to Thomas’s headphones in his control room, Kelly’s headphones in an adjacent room, and the headphone input of Thomas’s iPad, which fed my headphones on the other side of the Pond. Getting audio in and out of the iPad required a special break-out cable from kV Connection (www.kvconnection.com), which split Apple’s four-conductor TRRS minijack connector onto separate mono input and stereo output sockets, as well as attenuating the iPad’s nominally mic-level input to accept line-level signals.
As a result, Thomas and Kelly could work exactly as they normally would, but with me listening in and offering comments to both of them over FaceTime. The only slight restriction with this setup was that I wasn’t able to switch off my own ‘talkback’ signal, because I was using my iPad’s built-in mic. Fortunately, Internet telephony conserves data bandwidth by muting audio signals when they go below a certain threshold level, so all I had to do was keep reasonably quiet during takes to avoid distracting Kelly. This militated against my using loudspeaker monitoring, but that was no sacrifice given the data-reduced audio stream I was hearing. There was always the risk, of course, that our FaceTime connection might stall mid-session, so we agreed in advance that Thomas shouldn’t interrupt any take-in-progress on that account, and that I’d simply wait for him to hook up with me again from his end at a more suitable moment. We swapped land-line numbers too, on the off-chance that technical Gremlins scuppered FaceTime completely. As it happened, though, we were lucky enough not to have a single connection drop-out during the entire vocal session — although it has to be said that Thomas and I did carry out a short technical dry-run a few days beforehand so as to leave as little as possible to chance in this respect!
Another thing Thomas and I discussed was how to assemble the lead vocal takes for comping. Different people have different techniques here, but I normally prefer to limit the number of takes I keep by restricting myself to recording them on no more than eight tracks, the idea being that I keep rerecording on those eight tracks until they all contain performances of similarly high calibre. This helps keep me in a more critical frame of mind while tracking, and heads off the logistical nightmare of trawling through dozens upon dozens of different takes while comping. Thomas was also comfortable with this approach, so I made sure that we’d be able to audition any of our vocal tracks quickly and easily for comparison purposes, and that we’d be able to punch in on any of them at will — both prerequisites for the smooth running of his approach. Finally, we agreed to face Thomas’s iPad towards his computer screen so that I could see what he was doing, reducing the danger of miscommunication between us while overdubbing.
Kelly and Thomas were accustomed to recording lead vocal parts in sections, building up takes for each one in turn. I was happy to go along with this, as it’s often effective with less experienced singers in particular, naturally encouraging them to spend more time on the most challenging song sections. I’m also a fan of laying down a couple of go-for-broke takes of each section before moving onto the next, making it clear to the singer that they’ve already provided enough material for a decent master take, so they’re completely free to try any idea they like without risk. As it often does, this tactic yielded some really great moments on this session, such as the fragile middle-section ending, and the more elaborate melody line of “for someone to stay” in the final chorus.
The order in which you tackle the different sections when working in this way warrants careful consideration, though. We agreed to tackle the two verses first, because they needed a more restrained and rhythmically fluid delivery than the choruses, and I really didn’t want to pull Kelly out of that emotional zone to do the first chorus, for fear of difficulties returning to it for the second verse. After the verses, we rattled through each of the three choruses in turn before finishing with the quieter middle-section.
Why not do the middle section after the verses, given that those sections were closer in character? There were two reasons: firstly, I suspected from Thomas’s original mix that the more nuanced sections of the song would demand the greatest concentration from Kelly, so I didn’t want to hit her with all three of those back-to-back without any respite; and, secondly, I was gambling that Kelly’s voice would have started to fatigue by the time we completed the final chorus, and that this might bolster the authenticity of the middle-section lyric, with a pleasing touch of unfeigned timbral raggedness.
With any overdubbing approach, though, it’s wise to anticipate its inherent dangers. In this scenario, it would have been easy to misjudge the transition into the beginning of each chorus, for instance, so I made a point of playing the preceding section in each case to refresh everyone’s sense of context. I was also watchful for signs that the first chorus might start getting too powerful while we were trying to improve on its earliest takes. I wanted to get the best takes I could, but not at the expense of making the second verse feel like a let-down or by stealing too much thunder from the later, more impassioned choruses.
If you’re recording a vocalist in a separate room with headphone foldback, it’s vital that you keep encouraging and assisting the singer over the talkback to avoid them feeling isolated. I figured that Kelly might also be a little intimidated by this new foreign guy putting her vocal performance under the microscope, so I deliberately tried to keep some kind of commentary going whenever the red light was off, the idea being to minimise any uncertainties she might have in her mind.
A big part of this was, as it usually is on vocal sessions, simply responding naturally to sections that I thought were killer! You’d be surprised how often singers don’t realise when they’ve done something really special, because their mind is focused elsewhere — maybe they’re trying to remember the words, controlling their falsetto, or struggling to nail down an awkward rhythm. If you tell them what you like about a take, and they can hear it for themselves on playback, they’re more likely to trust your opinion, and less likely to go on the defensive if you suggest changes to their performance. Playing back ‘magic moments’ from any take can really speed up workflow too, because it’s very good at focusing the musician on the kinds of results you’re looking for. This was very noticeable on this particular session: although the first chorus took a certain amount of work to nail down, as a result of that groundwork Kelly then aced most of the third chorus in the first take.
It’s also important for you realise that the singer won’t usually know what’s going on in the control room unless you tell them, so try to keep them in the loop whenever possible: “We’re just listening for a pre-roll point. No need to sing along quite yet.”; “Let’s play from the start of the verse and drop in line three.”; “I’d like to listen back to the first three takes. I’m almost certain take eight was miles better, in which case we might want to rerecord some of those.” However mundane such comments may appear to you, they let the singer know exactly what they need to be doing at any given moment, which means they can give their undivided attention to the task of delivering a gripping performance.
And, of course, from an engineering perspective you need to stay alert for anything you can do to help the singer. For example, I normally like to keep vocal sessions moving as swiftly as I can between takes, to help the singer remain ‘in character’, so to speak. But the extended chorus lines on this session demanded a good deal of breath from Kelly, so moving too quickly from one take to the next proved counterproductive — she needed enough time to get her breath back. The solution was hardly complicated (we just doubled the pre-roll duration), but the main challenge with things like this is simply spotting the problem.
Asking vocalists to change something about their performance can be tricky, because singing is such an instinctive activity — you can ask a guitarist to change their picking position or adjust their instrument’s tone controls, but you’ll get a blank look (at best!) if you ask the average singer to increase their sinus resonance or twang their epiglottic funnel less assiduously! Fortunately, though, there are some more tangible things you can ask of a singer which have useful knock-on effects for their hidden vocal apparatus.
One of my favourites here is suggesting that the vocalist try different facial expressions. The most common application is asking the singer to smile, which can dramatically brighten the recorded sound, but on this particular session we actually had good results asking Kelly to sneer while singing some lines, because this hardened the vocal tone, as well as supporting the emotional backdrop of the performance. Urging a vocalist to dance along while singing is another time-honoured trick for adding rhythmic urgency with up-tempo numbers, but here we used this concept more subtly, asking Kelly to move a little in time with just the chorus sections to give the performance a touch more rhythmic propulsion there.
Questions of emotional interpretation can be equally tough to communicate in words, so I often search for shared visual metaphors instead, and characters from film and television can be great for this. When we were grappling with this song’s middle section, for example, I mentioned Natalie Portman’s character in Luc Besson’s classic film Léon. Although this wasn’t actually a film Kelly knew, the ensuing discussion of similar characters in the media nonetheless moved us closer to the particular ‘sadness suppressed by bitterness’ mood I was hoping for.
When you’re trying to home in on an appropriate performance, taking a cue from natural speech patterns can once again be a useful aid. For instance, I noticed on one occasion that Kelly seemed to be stressing the second syllable of the word “baby”, which you’d rarely ever do in natural speech, so we changed the way she delivered that line to get a more suitable result. You should keep your ears open for weak enunciation of lyrics too, which is one thing that mixdown processing is all but powerless to correct. Here, for example, the words “you” and “your” in the song’s final chorus were central to the lyric’s change of perspective (from “I” to “you”), so I deliberately asked Kelly to bring out the ‘y’ consonant a little more clearly.
Another crucial variable is how forceful the singer’s delivery is (in terms of both physical effort and pathos), and whether that’s actually appropriate at each moment in the song. With Kelly, the trickiest moments to manage in this regard were the first verse, where it took a little while for her to get used to a gentler, more vulnerable singing style; the first chorus, where we wanted to keep some firepower in reserve for later choruses; and the middle section, which needed an affecting performance, but with enough restraint that the final chorus could still arrive with a bang. In all those cases we deliberately stripped away some of the part’s melodic ornamentation, particularly in the middle section, where some of the lines moved towards an almost childlike one-note-per-syllable feel. We also tried to connect the end of the first chorus back into the mood of the second verse by encouraging Kelly to thin out her vocal tone gradually towards the end of the final line (“you won’t be there”), a tactic that also worked quite well for the transition between the end of the third chorus (“but no-one cares”) and the subsequent plaintive guitar solo. By the same token, I was keen that she gave the “now I” chorus lead-in a certain degree of rhythmic emphasis, despite retaining a fairly breathy vocal timbre, to subtly prefigure that aspect of the chorus vocal’s attitude.
All in all, it took us just under three hours to get all the material I felt that we needed, which is fairly typical in my experience. Thomas then sent his consolidated Cubase project file over to me so that I could edit together a preliminary comp, and it didn’t take long before we had master take that we were all much happier with.
So, if you find yourself struggling to achieve a truly satisfying lead vocal recording, why not try out a few of the vocal production tricks we used here for your own production process? It might just give you the extra edge you’re looking for.
One of the best reasons for any vocalist to learn their lyrics by heart is so they can perform with eyes closed, which can really help connect a singer with the mood of what they’re performing — as it did on this session while we were tackling the complex territory of this song’s middle section.
There’s a lot to be said for the engineer closing their eyes too, though, because one of the best ways to judge a performance’s emotional immediacy is by shutting your eyes and picturing the singer telling you their story. Somehow the lack of other visual input seems to make it easier to decide whether they really mean what they’re singing about.
Head over to the SOS site to hear audio examples demonstrating some of the issues discussed in this month’s column. You can also compare Kelly’s original vocal performance (in it’s original mix context) with the comped master take from our collaborative re-recording, as featured in the final mix.
Kelly Orr: “My main ‘takeaway’ from the whole production process was the importance of having a vision of how I wanted the song to sound, and holding true to that. Initially, I was trying to make my performance have ‘attitude’, but Mike focused things more on the vulnerability of the lyric, which better matched the mood of the song. On stage, it’s easy to get into the right character, but that was hard for me in a recording environment, so having someone ‘remember’ the goal and keep pushing at it over and over again was a key success factor.
“On a practical level, I learned to quit yelling at the mic! I love to ‘power sing’, but Mike’s respectful suggestion to let the mic do more of the work helped let in the emotion. Practicing with a clipboard in front of my face also really helped to demonstrate that less can be more. I’m the daughter of a choral singer, which means I’m used to singing in a very controlled way, so realising that it was OK to let in expressive breaks, cracks, and warbles was another ‘Aha!’ moment for me.”
Thomas Stevenson: “The tracking sessions with Mike really introduced us to new workflows for tracking vocals. In the home/project studio environment, it’s easy to be casual about these details, but Mike’s discipline of making some keep/discard decisions during tracking, thereby restricting the number of vocal tracks, definitely made the comping process easier. This meant that we could concentrate our editing decisions more on personal preferences, rather than having to shortlist takes according to execution merit first. It was also educational to hear which parts Mike preferred, especially on those occasions where he highlighted emotion over execution. Overall, it’s been a wonderful learning experience for which we’re both grateful!”