Recording an instrument and a voice together

You set up two microphones for a lesson. One on the instrument, one on you. Sensible, and the right instinct.

Then you sit down to edit and find that turning the instrument down turns you down with it, and bringing your voice up drags the instrument along. The two are locked together, and it makes no sense, because they are on separate tracks.

They are locked because a microphone is not a camera. It does not point at one thing and exclude everything else. The microphone on your instrument also hears you talking, and the microphone on you also hears the instrument. That is bleed, and it is the most common reason a two microphone lesson cannot be balanced afterwards.

Why does turning one down not work?

Because you are not turning down one sound. You are turning down two copies of two sounds.

The instrument exists on both tracks: a close, strong version on its own microphone and a weaker, more distant version on yours. Pull the instrument track down and the weak copy on the voice track stays exactly where it was, so the instrument does not get quieter so much as it changes character. It loses the close, clear version and keeps the roomy one. The balance moves and the tone moves with it, in a direction nobody asked for.

There is a second problem underneath. Two copies of one sound arriving at slightly different moments partly cancel each other, which makes the result thinner as well as murkier. That fault has its own post; bleed is what puts you in a position to suffer from it.

How do you reduce bleed without a bigger room?

With two microphones recording simultaneously in the same room, expect some bleed. Placement can reduce it enough to make the tracks more useful.

Point them away from each other

A cardioid microphone hears least from directly behind. If the microphone on your instrument has its back turned towards you, and the microphone on you has its back turned towards the instrument, each one is rejecting the other's source with the quietest part of its own hearing.

Check where each microphone is least sensitive, then aim that part towards the unwanted source. The best arrangement depends on the polar pattern and the room.

Use distance and angle, not volume

Bleed is a ratio between what a microphone is meant to hear and what it is not. Moving a microphone closer to its own source improves that ratio without touching anything else, because the wanted sound gets louder while the unwanted one stays exactly where it was.

Turning up the recording gain lifts wanted sound and bleed together. Adjust placement first, then set gain with enough headroom for the loudest playing.

Record the empty room

Record a few seconds of the room without playing or talking. This gives the editor a reference for steady background noise; it cannot remove instrument bleed or guarantee clean separation.

How one setup was redesigned

On one music education partnership, the audio problem was diagnosed from the delivered files alone. The placement was audible in the recordings: what each microphone was hearing that it should not have been, and what that was doing to the balance between the instrument and the speech.

That reading was confirmed with the partner before the placement changed. A diagnosis made from files can be wrong about a room, and the person standing in it knows things the files do not. The revised positions gave us more control over the balance.

Two things followed. The room tone became cleanable, because there was finally a version of the noise not welded to a voice. And the ceiling on the series went up, because every later decision about balance was now available rather than already settled.

None of that involved better microphones. It was the same microphones, in different places.

Does this mean one microphone is better?

Sometimes, yes, and it is worth saying before you buy a second.

If your instrument and your voice stay in roughly the same proportion, and you never have to raise your voice over the playing, one well placed microphone between the two sources will give you a recording that sounds like a person and an instrument in a room. That is a legitimate result, and it is one decision instead of five.

Two microphones earn their trouble when the balance has to change: a lesson where the talking must stay clear over the playing, or a series that will be recut at different lengths later by somebody who is not you. If neither of those is true, keep the setup time and spend it on the playing.

Before the next lesson

Turn both microphones so that neither is looking at the other's source. Then move each one closer to what it is supposed to be hearing, until it sits nearer than feels polite.

Then record two minutes of playing and talking exactly as you normally would, and listen to each track completely alone. If you can hear the instrument clearly on the voice track, the placement is still doing the blending for you, and it will keep doing it until it is moved.

That listen is a useful first check and costs one take. Listen to both tracks together as well: a cleaner solo track is only useful if the combined recording works.

Sources and evidence

What here is established acoustics and what is our own method.

What is established. Two microphones on one source at different distances interfere and thin the sound through comb filtering, documented by DPA Microphones and analysed for multiple sources by Clifford and Reiss (2011); the fix is spacing and alignment, not a better microphone. How close a microphone sits also changes its tone through the proximity effect (DPA), which depends on the polar pattern.

Our method. Capturing instrument and voice on their own channels, spaced to avoid interference and aligned before anything is balanced, is the practice here.

References.

• DPA Microphones. The basics about comb filtering and how to avoid it. dpamicrophones.com. Multiple delayed versions of a signal can create comb filtering. Manufacturer education; diagnosis still requires listening and measurement.
• Clifford and Reiss. Calculating time delays of multiple active sources in live sound. eecs.qmul.ac.uk. Time offsets and multi-source interference. Specific technical context.
• DPA Microphones. Proximity effect in microphones explained. dpamicrophones.com. Distance response depends on microphone design and polar pattern. Manufacturer education.