5 tips for comping vocals
What is Vocal Comping?
Vocal comping describes the process of combining multiple vocal takes into one “supertake” that has the best parts of each. This is called a “composite track,” or comp for short. Many instruments can be comped on studio tracks, but vocals are the most common since they are the most central aspect of most popular songs.
Comping vocals has become essential in popular music because popular music must sustain repeat listens, which requires a near-perfect vocal take with many interesting dimensions. This is a tall order, so it normally takes multiple takes to get all the right pieces, which then need to be fit together.
In this article, you will learn common practices for comping as well as some more advanced tips for getting the best sound.
Artifacts from Comping
Before we get into more advanced tips for comping, let’s take a moment to understand the artifacts that can surface when comping is done poorly, and the basic techniques used to do comping well.
Clicks and Pops
The main artifact you get from comping is clicks/pops. You get this artifact when you introduce a sudden break in the waveform through comping.
Speakers produce sound when a cone moves backward and forwards, creating air pressure waves. This cone moves in response to an electrical signal (the waveform) which is telling it which position it should be in. For any organic waveform, this wave is going to be continuous and the speaker will move smoothly without producing any artifacts.
But, with human intervention like comping, it’s possible to make waveforms that are not continuous, which will cause the speaker to jump as quickly as possible to keep up. This is what produces a click or pop sound. (This is why you often hear clicks when plugging in an audio source into your speaker. The speaker suddenly receives a signal other than zero, and it jumps to that position as quickly as it can.)
When comping vocals, it is common to accidentally join two vocals so that their waveforms don’t “line up” and that will give you a click!
Depending on your DAW, this may not be much of a problem. Many DAWs like Ableton apply a small envelope at the beginning and end of each audio clip by default to avoid clicks and pops. (To see this in Ableton, go to Create > Show Fades).
When both audio clips are faded like this, they are connected at 0 amplitude, and there’s no artifact.
This is handy, but it can cause problems. Take this example from Pro Tools.
This is an example of a bad place to comp because the speaker would have to suddenly jump positions creating a click/pop. This version of Pro Tools try to mediate the situation by connecting the lines. But what appears to be a “smooth” connection is still too quick and jagged for a speaker and you will likely still get an artifact.
Zero-Crossings
The solution to this problem is by using zero-crossings. Zero-crossings are literally just places where an audio clip crosses zero. By choosing these locations to be the point of comping for two audio files, we ensure that the speaker does not have to jump a large distance, even with rapid fading from a DAW.
Doing a little dragging of the audio clips on the timeline to accomplish this shouldn’t mess with the rhythm of the song or otherwise be noticeable. Since we are working on a such a small scale, we are dragging audio by amounts smaller than millisecond.
Advanced Tips
Now that you have the basic mechanics of comping down, you can learn some more advanced techniques to comp more efficiently and effectively.
1. Comp on sibilances
When possible, it’s always nice to comp together entire phrases using the silence that separates them. This is the most invisibile way to make a comp. But sometimes you really do need the first half of a phrase in one take and the second half of the phrase from another take.
In this case, sibilances are the way to go. Sibilances are bits of audio that sound like “s” or “sh.” Basically, sounds that contain a lot of noise.
Because of the noise in these sounds, it’s a great place to make seamless edits between two audio clips. When comping on sibilances, I mostly find that I don’t have to worry about zero-crossings and the default fading behavior of most DAWs is sufficient to mask the edit.
Other good sibilances to edit on are “ch,” “h,” “z,” “k,” “f,” and there’s a good number more that you’ll happen upon. Rather than memorizing these, just look for noisy sounding consonants in your vocal takes.
2. Hide comps behind downbeats
There are times when getting the parts of a take that you really want can mean you have to make a super audible edit. This is where masking comes into play.
There are two main types of masking: frequency masking, where sounds of similar frequencies tend to mask one another, and auditory masking, where louder sounds tend to mask quieter sounds. These are psychological phenomenon, meaning its in our heads. The frequencies are still there, and its our brains that ignore them.
(In fact, part of how mp3 files perform their compression is by removing audio information that you wouldn’t be able to hear because it was masked by other sounds.)
So how does this apply to comping? If you are able to place a nasty vocal edit behind a particularly loud sound and/or one of similar frequencies to the vocal, then the edit might be masked entirely. You’ll be surprised with the unforgivable comps you can get away with by hiding them behind a snare drum.
3. Use cross-fades sparingly
Cross fades are a pretty common way to avoid clicks and pops. Crossfades involve blending two audio clip by fading one out while simultaneously bringing the other one in. Film sound designers commonly use them to loop background ambiences without the listener noticing. DJs use them as a way to switch between songs in a mashup.
Comping is not my favorite use of crossfades, however. By design, crossfades are playing two similar sounding audio files at the same time. For a vocal, this can create an unwanted chorus effect that appears only for a split second, but that can be very disruptive to the listener.
The other approaches discussed should be tried first before trying a crossfade. In most cases using a zero crossing and/or finding a sibilance to comp on is possible and sufficient.
4. Keep track of good takes while recording
If possible, you (or a tracking engineer if you have one) should try to keep track of the good takes and bad takes while you record. When you start comping, you can go straight to the takes you know you love before having to circle back to the other files. This way you’ll be less frustrated sifting through takes and you’ll make sure you don’t miss out on your best deliveries. You really can get as specific as you like here, just figure out what works for you. Here’s a little example.
In this example, the ✓s and X’s help keep track of good and bad takes more generally so you know where to start when comping. There still might be useable stuff in the “bad” takes, but starting with the good ones will save a lot of time and limit the number of comps you need to do. (Comping is necessary, but if you can avoid doing it then you should!)
It also might be helpful to keep track on takes of any particularly good phrase that you want to keep, even in a bad take. You could do this with a lyric or even a timestamp. Comping is largely about finding all the little magic moments in all the individual takes and combining them into a super take that wows the listener.
It’s sometimes pretty common to redo just one or two sections of a song that you didn’t get good takes for rather than sing the whole thing through. So using some sort of neutral notation like a dash might help in that situation.
If you aren’t able to keep track of your takes while recording, it might be worth it to listen to all of your takes once through and take notes before you start your comping process. Also, be cautious about the number of takes you do. In the digital era where data storage is not an issue, it can be tempting to do dozens of takes to get a perfect vocal, but you might just end up overwhelming yourself. The law of diminishing returns definitely applies to the number of takes you have!
5. Listen with your ears, not your eyes.
Sometimes you’ll make a comp that you can just barely hear, and it can be frustrating. What you might forget is that if you are watching your playhead move across your audio file, you know exactly when to listen for the comp sound and you are going to hear it. The best comps are definitely the ones that you can’t hear even when looking at your DAW, but if you can’t tell without looking at your DAW then that’s totally fine!
Next time you’re unsure about a comp, close your eyes and then play it back. See if you can tell where is the comp is then. You might be surprised at how invisible it becomes.
Other Applications
Comping is not the only application of these techniques. Another big application is sound editing for films. Film sound designers often have to insert artificial ambience into a scene, and the ambience audio file might be shorter than the scene itself, and they have to loop the audio file without the viewer noticing. Or other times they are given a 4 minute chunk of music and asked to cut it to fit a 3 minute scene. This kind of task requires cutting out sections of music and seamlessly joining the remaining pieces by using zero-crossings in the music file.
That’s all for this post! Hopefully, you’ll be able to implement these techniques into your own music sometime soon.
