It’s great when a vocal performance can be recorded in one cohesive take, but not every session will succeed with this approach. Sometimes, a modular strategy is required when tracking vocals.
Throughout my career as an indie musician, songwriter, and producer, I’ve had the privilege of working with a very talented array of vocalists in different recording settings. I’ve found that capturing the right vocal takes — and crafting the finished tracks — calls for a variety of strategies in the studio.
Sometimes, tactics are straightforward: Let the singer sing. If the vocalist (who is sometimes me) knows the song well, is in good spirits and good voice, and has the confidence to pull it off, things stay simple. We record a few takes all the way through, discuss and adjust words or phrases as needed, and record some more. When we hit a take that nails it, everyone usually feels it at the same time. We may go back and punch in a few words or phrases, or perform an edit or two, and call it done.
It’s wonderful when vocal takes can be recorded this way, but not every project, singer, or song is best suited to this approach. Sometimes, a much more structured, microscopic, and modular strategy is required to get the job done. Here’s an introduction to a strategy I’ve used recently on a number of recording projects.
It started with a difficult song
One track that I produced not long ago was not easy to sing, and the issues became apparent after the very first vocal take. The song was dense lyrically — lots of words and tricky rhythms, almost tripping over each other. The singer definitely had the talent and musicality to pull off the performance in a single go, but she had little time to learn the song before our studio date. It was unlikely that she would be able to internalize its nuances deeply enough to deliver, in a single take, the vocals that we needed in our narrow window of time together.
Going granular
Instead of having her sing the entire song from start to finish — or even full verses and choruses in one go — we went in the opposite direction, asking her to record the song phrase by phrase. In practice, this meant looping the instrumental parts underneath any given vocal phrase, giving her two bars or so of breathing room in between repetitions of the section, hitting record, and letting her sing the same phrase, over and over, until it felt like second nature. Sometimes I would interrupt with constructive feedback; other times, I would let the proverbial tape role while the singer explored the material and got into a groove with each section. This may sound like a tedious process, but it turned out to be a fascinating, fun, and productive way forward.
Benefits of breaking it up
It was intriguing to hear how the singer’s interpretations changed, repetition to repetition. On simpler phrases, or ones that she immediately felt comfortable with, the first two or three takes usually sounded very good and almost identical, and she started to sound bored starting with takes three or four. Needless to say, we stopped tracking at that point. On more complex sections, by the third or fourth time through, she would start to really understand, internalize, and own the phrase and we would get to the sweet spot by take five or six. I remember more than one phrase where there was an almost amazing transformation from take one (“wow, she’s barely getting the notes and words right”) to take six (“that sounds perfect, like a finished record”).
Downsides
Clearly, there’s something wonderfully organic about a singer delivering a performance, front to back, with minimal edits or other digital prestidigitation. Such performances can have a singular cohesiveness and a sense of momentum, flow, humanity, and authenticity that you simply can’t replace. If you go granular and record each phrase individually, you run the risk of sounding more synthetic and “produced” than organic, and your vocal track may lack that emotional or musical cohesiveness that full takes can convey. By going granular, you also run the risk of tiring your singer out prematurely; if you need nine takes each to get the first six phrases of a song nailed, that’s already a lot of singing, and you may still have multiple verses to go. Which is to say, there are not reasons not to track this way.
Mixing it up
When deciding how to structure the vocal session, I made a point of skipping around and tracking the song very much out of order. I started with the most ballsy, aggressive part of the song, which I felt to be the best-suited for the energy of a singer coming in off the street, pumped and psyched to record. There were several parts of the song that mirrored similar melodies and moods, and we tracked all of those at the same time to capture a similar energy. Next, we moved to moderately tender moments, to match how the singer was easing into the session and warming up more to the song. Again, we did all of the parts with similar melodies and vibes in order before switching it up. Finally, when the singer was at her most relaxed and thoroughly in the groove, we hit the quietest, most emotional parts of the song.
Stitching it together
At the end of a session like this, you can end up with a nightmarishly long number of takes to go through, so we made sure to take notes of what we liked and what we didn’t while recording. Jotting thoughts on paper of which takes are preferred, or turning those regions a different color in your DAW, can help document what spoke magic in the moment and save lots of time when it comes to working further on the song.
Do you record vocals phrase by phrase? Tell us about the pros and cons of your experience in the comments below.
Michael Gallant is a musician, writer, and entrepreneur living in New York City. His debut album for the Steinway & Sons label, Rock Rewind, features solo piano reinventions of Pearl Jam, U2, Halestorm, Janis Joplin, Bob Dylan, Radiohead, and more. Read his recent article for the National Endowment for the Arts and follow Michael on Twitter at @Michael_Gallant and Facebook.com/GallantMusic.
Related Posts
10 ways to lay the groundwork for an amazing vocal recording session
Creating a great composite vocal recording
Vocal recording advice for the experimental at heart
Recording vocals in a home studio
Choosing a signature vocal mic for your studio
The biggest challenge I have faced with this approach is maintaining continuity. Many times the takes are at different levels or dynamics, simply because the singer may have been closer or further away from the mic. Then there are breaths or just the delivery itself makes it sound like it’s recorded separately and not organically in the same flow and feel. These are hard to edit and make natural.
Usually, when I am the vocalist, I’ll run through as much of the song as I can in one take. Then go back and re-track phrases or lines that could use some tweaks. Sometimes there is little to clean up. Sometimes there is almost nothing left of my original take. But I feel it helps the balance between natural and polished.
I usually record the chorus first since more energy and consistency is needed. Then I record the verse either complete and then punch-in phrases that need repair to keep the stream of the verse intact.
I found this article to be a good strategy since most singers will have some sort of error when trying a take from top to bottom. I agree that the most demanding parts should be done first and doubling usually fits the prior energy levels of certain parts!!! Thanks—–EARLY A.M. Productions / So. Calif……….
If one has to do a Song in Phrases then there is something wrong or maybe lacking of training or knowing the song well enough to record.
.
True Professionals knows one has to learn the song forwards and backwards so that when they come in they are not wasting time & money.
.
Someone told me once in the Prof. World of Music said, “One has to go through a song, at least, 100 times before they have the song memorized.” I would add to this that it takes more than that to style the song the way you want it.
Charles of “Higher Call”
This is how I record vocals every time. I like the way it allows you to focus everything on just that one phrase or line without having to worry about remembering the next. This works even with songs I’ve been performing for some time.
When I first started recording, I thought I was a bad singer because it was difficult for me to sing my songs all the way through without a mess up somewhere. So, I learned by doing. I would sing the song through 3 or 4 times and then comp. the best parts together into one track. As I started learning more and reading articles like this one, I realized that even professional musicians record this way. The thing that I’ve learned is that if you are going to do multiple takes on a song or punch in and out it is better to do it all in one session. I’ve found that when I come back in to do this kind of recording a few days later that my tracks don’t usually sound cohesive. Thank you for your article. I look forward to reading more of your writing.
I like to record my vocals in sections, from beginning to end. I’ll do each song 3 times recording verse 1, then pre-chorus, then chorus separately.
Then verse 2, pre-chorus 2, chorus 2. And so on. This way there’s some continuity to the vocal but the vocal energy is always up.
My preference. Your mileage my vary.
I have worked with writer singers who had wonderful songs and warm expressive voice, but zero concept of time, meter, and performance skills.
From way, way back in the tape days I became a master of punch ins, line by line recording, conducting the vocalist, and adding harmonies.
A lot of very good records were made this way.
For many years digital equipment was too limited and difficult to do this kind of work. A lot of it still is. You can hear it.
I have always considered it to be an essential engineering skill.
In today’s world (on radio) almost every professionally produced artists vocal track I hear is processed heavily to the point where it sounds done on purpose, as part of the “art” and song, as opposed to simply tweaking pitch or timing. Also, I doubt that “any” song that actually goes out to the world by a known artist, is still sung all the way through in one take, and are most likely done in many small parts, and edited later. With the exeption of a live recording I suppose. Anyone putting out a (look how great I can sing) one take record is going to find that they were off key, or of time more than they thought. Those tracks sound like bad songwriter demos to me.
I find this method very useful when recording my own background vocal harmonies. First I’ll do several takes until I figure there are a couple of good ones, then without pause I’ll do several takes of a harmony line, repeat as needed for additional harmonies. Because of the quick repetitions, the phrasing between takes is very consistent and the takes blend well when stacked.
I’m a professional vocalist with a home studio, and I track vocals for multiple producers and projects. I often don’t have time to study the material so I use this technique quite a bit. My trick is to sing-along for a bit with the end of the previous segment just before the punch, which helps each section to have continuity with the one before it…making them all seem more cohesive with each other. I’m also careful to pay close attention to the breath. Cutting off or cutting into the breath at the beginning or ending of any of the phrases can really take away from the organic sound of the vocal.
One comment I would like to add is that in my recordings using the modular approach, one needs to stay aware of the breathing of the singer. If you are not careful, you can create an artificial sounding track by comping takes together without the natural breath sounds of a singer. I know this from my own mistakes.
I self-produce my own stuff, I do it all myself from start to finish so I am also the vocalist as well as the arranger and engineer. I routinely stitch bits together. I will sing multiple takes of the complete song and while each take sounds similar, there will be some moments which I love more on one than the others, usually to do with the tone or how a long note sounds and the vibrato on it, or an inflection in a phrase where the subtle emotion comes through better, or even sometimes when a note cracks and does something very unexpected and unplanned in the moment but sounds perfect. Sometimes a whole section will be the strongest of all the takes and I will use that, but I will go through all of them and pick the best for that section and keep using it until I come to a part which is better performed elsewhere if one exists. I usually record about 6 complete takes before I even start listening back and then pick the best bits and join them together. On the opposite side, I will sometimes need to take out something which didn’t quite sound right which can often come just before or just after one of the best bits of all the takes, so while the couple of words before it were the best versions, what comes after it I might see as the worst version of all of them, so I will snip out the bad ones which follow and replace them with the best version of that phrase from another take which in most other areas could have been the most mediocre compared to the rest. I was happy to read this article btw because I always felt like I was cheating lol.
Especially vital to do vocal tracking this way if you are going for a more theatrical performance. The tone of certain phrases such as “Into the fire” – “back from the ashes” – need to sound like you mean what you are syaing. As well as being in pitch, and in time. We go for full verses and listen back. If it is good we keep. If not good after a couple takes – we start doing it line by line. Same with chorus. If the singer gets worn we come back and do it again another day. Bottm line is – take time to do whatever need to be done to make sure it is great.
That’s a good idea if you’re not getting anything usable from recording the entire song. This seems to be the case when the singer doesn’t know the song very well. I’ve had that situation, too. Somehow we muddled through, punching in the weak parts over and over until guy that hired us was happy enough with the final track. It’s so much better when the singer has done their homework! Then, they can nail the track and the extra time could be spent giving different choices for phrasing parts of the song.
I like your writings. You summarize all the significant points. Thank you Michael. I have been following your posts.
we have tracked vox line by line at Blodren Studios on several occasions
this was (as you mentioned) when the singer* wasn’t really familiar enough – or didn’t like what he had prepared
it can be very beneficial under those circumstances
as a general rule tho, we go for at least two lines linear, which tends to give a more natural/cohesive feel to finished product
RBG
ilikebox.com
*he says he is