Recording Vocals without Headphones
I’ve always disliked tracking anything with headphones, especially vocals. Some vocalists have pitch problems on headphones. Some are just uninspired. If you’re recording yourself, it can be a pain in the head to keep switching your set-up.
The simple solution is to record without using headphones and instead monitor with speakers. There are a few ways to do this. I’ll go over all of them briefly, then show you a way to do it that works amazingly well.
First things first: remember that most of the leakage a directional microphone picks up is reflected sound coming in the FRONT of the mic. Remember that the back, or sides, of a directional mic are designed to reject sound and overall that works pretty well. I’m assuming you’re not cutting a vocal with a omnidirectional, especially if your goal is to cut down on leakage.
A moving coil mic is generally going to pick up less of the stuff you don’t want, which is the room reflections coming over the shoulder of the person singing. Why? Because moving coil mics are less sensitive overall—a big heavy diaphragm attached to a big heavy coil of wire has more inertia. Condenser mics will generally pick up more of everything, but I’ve cut tons of vocals with speakers for monitoring using condensers and it usually works out fine.
Just Record and the Hell with Leakage
You can set up a mic in front of the speakers and cut the vocals and just ignore the leakage. Depending on your room and the volume you’re working at, leakage might not be an issue. Obviously, use a directional mic with a cardioid or hypercardioid pattern, experiment with where you place it—you might get less leakage if you put it right in front of one speaker rather than in between the two (the polar pattern will affect this a lot).
In the mix, you’ll have to gate things or edit out the leakage. However, there can be problems because the leakage on the vocal track, when the gate is open or when the vocal is playing, can mix with the music on your tracks and you might hear a change to the snare or the low end whenever the vocal comes in and out of the mix. This is evident on the Chris Isaak song Wicked Game. My quick fix is to ride (automated) the vocal level rather than gate or edit it, so I can control how much of the overall sound of the track changes.
Using a moving coil rather than a condenser is recommended if you’re doing things this way.
If you’re sitting down, think about throwing a piece of acoustic foam over the keyboard/work surface and on the monitor. Close reflections suck for a vocal.
If the leakage is a problem, then you’ll probably have to start playing around with phase.
Put the Speakers Out of Phase
I think this is an awful solution to a problem, but I’ll explain it anyway, provided you promise not to use it. I tried it once and it was a waste of my time.
Set up the vocal mic so that it is in the exact sweet spot of the speakers, then reverse the phase of one of the speakers and THEORETICALLY the resulting phase cancellation will result in far less leakage, and because the singer is not exactly in the sweet spot — the spot of maximum cancellation — they’ll still be able to hear the music well enough. Engineers also have done things like putting out-of-phase speakers to either side of the singer, equidistant, pointing at the mic.
Why this sucks #1: It sounds awful
This sounds awful. Out-of-phase speakers sound awful. Mostly you’ll kill a ton of bass, so the music won’t be exciting—there's nothing quite like an unenergized and uninspired singer, and the net sound will be phasey and plain old weird. If the singer shifts or moves, they’ll hear all sorts of swooshing and if the phase issue is bad enough, they might get nauseated. Did you know huge weird phase shifts plays ear games and causes something akin to motion sickness? Ever cut tracks with a singer who wants to vomit?
Why this sucks #1: It works like ass
Because most of the leakage that comes in a mic is coming in the front, and is predominantly indirect sound, chances are the speakers out of phase trick isn’t going to buy you much. The out-of-phase sound that goes bouncing around the room comes back as in-phase leakage.
If you want to cancel using phase, you have to flip the phase AT THE MIC, not in the air.
Use Two Mics
Get two identical mics. Flip the phase on one of them. Put them very close together, displaced vertically rather than horizontally (one over the other rather than side by side). Have the singer sing into the in-phase mic, combine them into one track or do it in the mix.
This basically sucks too. In a live situation, this might be workable, but in a studio situation, unless the singer is working really close to the in-phase mic, this is going to be all over the place. Little movements will change the frequency response; the out-of-phase mic is picking up the singer's chest so things could wind up overly warm or thinned out, depending on how the phase cancellation affects frequency response. The singer has to put a lot of effort into staying still and in one place, and that usually results in a stiff, bad vocal.
Here’s the best way to do it.
Record an Out-of-Phase Leakage Track
The first time I tried this was tracking a jazz choir and I didn’t have enough headphones. I put them out into a room and set up a pair of big, loud monitors, and put two Crown PZMs mic out, each taped to a music stand. The speakers blasted into the choir, the mics were spread about 8’ roughly 6’ from the first rank of the choir. We cut a good take. I had the choir shuffle their positions around and recorded another take. I played back the four tracks and it was a big leaky mess until I reversed the phase of the second take. Leakage GONE. Vocals untouched.
This technique works jaw-droppingly well. Here’s how you do it:
First of all, set up the mic so the singer is really comfortable and loves how the speaker monitor mix sounds. I usually did this right in the control room in front of the console dead center, but it can be anywhere. Once you get the singer happy, tape the mic stand down, tape the mic to the stand — whatever you have to do to make sure that microphone doesn’t move at all.
You also can’t move the speakers, and you have to do all the recordings using the same speakers. And you should tape down the monitor level: it has to be the same for every take. Ideally, you do all the vocal tracks in one day.
So, set up your vocal chain (mine was usually a Quad Eight mic pre followed by a Summit TLA-100 and then an Aphex 551). Get it to sound dandy good. Put the vocalist in position, then have them sing. As many takes as you need. Punch in, etc. Don’t do any comping yet!
Once you get a good take, have the singer stand there in front of the mic silently and record just the leakage — exactly what the singer was hearing when they did their vocals, but no vocals. Now the magic: play back the vocal track, reverse the phase of the silent track, and bring it up in the mix until the leakage disappears. And disappear it will. This works like magic. Comp the vocal til the cows come home this will still work. You can even bounce the vocal track and the leakage track together — just make sure the leakage track is phase reversed.
In the old days before unlimited tracks, if I was going to double the vocal, I would cut the double and reverse the phase of that — I have yet to have a singer cut a double so close that any of the vocals canceled, but I suppose it’s possible. I also don’t see the point of cutting a double so tight that you can’t tell it’s a double — just bring up the original track by 3dB and get it over with.
This technique works amazingly well. It’s how Chris Cornell cut a lot of his vocals. I’ve used it on hundreds of sessions. This method will also work with an omnidirectional or a bidirectional mic, and it works like a charm with condensers as well.