Recording vocals

My process for recording vocals for songs is as follows:

0) (Write the song)

1) Record a guide track. This is usually a skeletal piano part that simply plays the chords for the song. And simple means simple – it’s usually block chords, one per crotchet (1/4 note). Sometimes there’s a particular riff or groove that’s integral to the song, and this forms part of the guide track. The guide track rarely makes it to the final mix – its sole purpose is to provide the vocalist with something to inform their pitch and timing. Sometimes the guide track doesn’t represent actual structure of the song – I might record the sections separately and arrange them later while I’m recording the other instruments.

2) Record a scratch vocal part. This is performed to the guide track. Usually I only record one or two takes of the lead vocal melody, and harmonies or backing vocals only if they’re particularly important. This scratch vocal recording only needs to serve as my own guide for writing for and recording the other instrumental parts. As such, it doesn’t need to be a great performance – it might even have some mistakes in it! The bulk of the time in this session is not spent coaching the vocalist – it’s spent working on song structure and melody before the recording takes place.

3) Record other instruments. This is done using the scratch vocal as a guide for song structure and mix placement. For vocal music it’s important to start with the lead vocal and build everything else around that. This is important not only for arrangement, but also for instrument voicing (which notes, how high, etc) and mixing. Producing a backing track without having a vocal to work with will easily result in an instrumental song that sounds great on its own, but will struggle the accommodate the vocal once its added.

4) Record backing vocals. Once the arrangement is worked out and most of the other instruments are recorded, it’s time to record the final vocal parts. It’s important to wait until the track is almost finished, so that the vibe and energy of the music can influence the vocal performance. That way the vocalist can deliver a performance that best suits the song.

I usually have the vocalist record the backing vocal parts first. This is because they’re not as critical as the lead vocal. This allows the vocalist to warm up and familiarise himself/herself with the song and the studio. It gives us both a change to fine-tune the headphone levels and monitoring (usually some compression and treble lift) whilst still remaining productive. It also gives me a chance to learn how to work with the vocalist to get the best results. Some will respond well to relentless pace, some resond better to a gentler approach. Some nail it in the first couple of takes, others need six or eight takes to get it. Some respond well to micro-advice (“That fourth syllable is dragging again”), some need more general encouragement (“That was great, now do it with more energy”). It also helps me learn about the voice. How hard can it be driven? How soon until it needs a rest? Where is the sweet spot between warming up and tiring out? Are there any difficult transitions between chest voice and head voice? These are all important issues to be aware of.

5) Record the final lead vocal. This is critical. The lead vocal is the most important part of the song. As if that wasn’t enough pressure, the vocalist only has a short period of time in the sweet spot at which you’ll get the best performance. You’ve got to know where this sweet spot is, because going much past it will give you worse takes, and demoralise the vocalist. Coming back another day often doesn’t give you a better performance either (unless you screwed up and scheduled the session when the vocalist is hung over or something).

My approach to recording lead vocals  is the result of many years of working with vocalists. Everyone will have their own methods, but this works for me:

5a) Run through the song once for practice. Often I see advice to record the practice takes. I don’t record them, because I know I’ll get better later on. I try to minimise the number of takes I record, because trawling through them later is a chore, and often doesn’t actually get significantly better performances. Some vocalists don’t even need the practice take.

5b) Record one take of the whole song. This is the basic lead vocal. I only record one take of this, unless there were any mistakes (in which case I delete the take and do it again). This is usually a good fallback for syllables to comp[1] in.

5c) Record two takes of each section, in reverse order. After the first take, we record section by section, in reverse order. So, we’ll start with the coda, then the final chorus, then the bridge, etc, finishing with the first verse or introduction. Each section is recorded twice on loop. I’ll only go more than twice if there were any mistakes. Recording each section in reverse order frees the vocalist from thinking about the song structure, and instead focusses her/him in the moment – the section being recorded. Recording each section twice on loop further enhances this focus. This is sometimes the point at which the singer is in the sweet spot, and where you’ll get a good balance between emotional performance and technical correctness. I record only two takes during this phase in order to keep the momentum up (which keeps the singer motivated and interested). By this stage, more than two takes rarely results in a better performance.

5d) Record one take of the whole song. At this point I have lead vocal that’s three takes deep, and is usually of a consistently high quality. I could use this material to put together a decent composite. However, I ask the singer to do one more take – the whole song through. This time, however, I instruct the singer to focus on delivering an emotional performance at the expense of technical correctness, perhaps pushing it a bit more than usual. I might even bump the volume in the headphones by a decibel or two to assist. For some singers, this falls in the sweet spot and results in a great performance. For others, it’s too much, so the take becomes a source for the occasional ’emotional syllable’. If this last take is not suitable, I’ll usually use one of the ‘reverse order’ takes as the base for the comp. Most times, however, it is this finial take that forms the base from the comp.


[1] Comp = Composite. The process of assembling a performance from several takes.

    • Lyrics
    • November 18th, 2009

    I think this is a great method, but does not apply as well for those who write instruments first. I think your process of recording lead vocals though is really well thought out.

  1. For those who write for most of the other instruments before vocals, this would be covered in the first step “(Write the song)”.


    • doozer
    • December 6th, 2009

    Great idea about the reverse takes, for someone like me who already knows the song inside out (solo artist and producer) this could help focus and motivation.

    However the ‘can’t fit vocals to an already made backing track’ thing, I know Freddy Mercury for one insisted on an almost complete ‘backing track’ (all the music done basically) before he’d even consider singing any vox for it as he wanted to become one with the music and use his voice to complement the sound rather than singing to nothing emotional? I understand technically your way is probably ‘better’ but for many singers who don’t just sing (but are very active in the song as a writer or musician) I feel they link to a lot of emotion within the sounds/arrangment which can influence vocal performance in a positive way.

    Just a thought.

  2. You’re absolutely correct Doozer, this is why the final vocal is recorded last (starting at step 5) – after all the other instruments have been recorded. This is exactly because the best vocal performance comes when the singer is able to respond to the vibe and emotion in the rest of the track.


    • doozer
    • December 7th, 2009

    Ahh true. Though most of the time (in my situation only as the ‘only brain in the process’) I have a scant guide vocal early on.. more to remember my melody/idea than to fit music around. Many times I won’t have a guide vocal because I can ‘hear it in my head’ perfectly and adopt the music to that… may sound strange I know but as the all-in-one type it’s like an intrinsic link between my mind and the music. So I could say, yes I have the rough melody idea, maybe some lyrics, sometimes a guide track but all times at least in my head ‘playing along’ with the arrangement. Then when the arrangement is almost done (rarely final as adding vocals will cause adjustments and embellishments to be needed of course) I record my vocals.

    I think it was, for me, the rule of ‘must record vocals first’ that threw me, because I guess in some way I have done that, I just do it in my head instead.

    For a normal full band, or someone with a seperate producer/mix engineer I guess you’d HAVE to record that guide/scratch vocal early to get more of a feel for the music as you went.

    Anyway, great articles I’m working my way through them all as my mixing certainly still needs a lot of help (even after 10+ years of trying) ;)

  3. I think this is a great method for dealing with singers. Me who rap and produce still left kind of stuck. Mainly because I haven’t recorded in a professional studio and don’t know “proper” recording techniques such as background vocals, lead vocals, how many times do I stack the hook, if at all.. Things like that.

  1. August 5th, 2009
Comments are closed.
%d bloggers like this: