Stereo recording microphone techniques
In stereo recording, it is important to have a plan for the mic technique. There are several main stereo recording techniques to use; each have variations to try. They differ in the way that they separate the stereo information for the brain to decode. There are 4 main ways:
- Near coincident
- Baffled or binaural
Coincident mic technique
The word coincident means occupying the same space or time. That is a very good description of the coincident mic technique, since it uses two microphones at the same place, angled differently, to pick up the sound.
The most popular coincident technique is the XY method. To make an XY setup, place the mics so one mic's diaphragm is directly above the other. Angle them 90 degrees apart, and point the middle of the two toward the middle of the sound source.
This requires directional mics, like cardioids. The stereo image given by this technique is based on level differences, or the volume of the one signal compared to the other. When a sound from the far right side of the stage comes to the microphones, the mic pointed in that direction picks it up louder than the other one. The second mic still gets sound, but it's not aimed the same way, and the picked up sound is quieter. When the sound is played back on stereo speakers, the right channel will have that sound louder than the left, so it is perceived as coming from the right.
XY is the most popular coincident miking setup used, but it has its drawbacks. For example, while the XY pattern has excellent localization (the ability to determine where a sound is coming from in the soundstage), the stereo image tends to be narrow. You can help by moving the microphones closer to the sound source and surrounding them with the source, but it just won't sound as spacious as some other stereo recording setups. Another type of coincidental mic setup is called the mid-side technique. It also involves two mics, but placed in a different pattern. A single mic of any polar pattern (omni, cardioid, hypercardioid, ect.) is set up pointing straight on to the sound source. This forms the mid part of the setup. A figure-8 mic is placed in the same area with the null axis towards the middle of the sound source. (The null axis is 90 degrees away from the front of the mic.) This will pick up sound from either side of the sound source, but not in the middle, so it is called the side.
In order to play it in stereo, a decoder is needed. The decoder combines the side signal with the mid signal and separates the left from right so the recording can be heard as stereo. The side mic is the one translating to stereo. The main principle here is polarity -- anything coming from the left side will make a negative signal, while the right side produces a positive signal (providing the mic is pointed to the right). The mid mic is used to sort out the two, and you have a stereo signal. The mid-side technique also provides good localization, but requires the use of a decoder, or some smarts and a few extra console tracks.
The Blumlein pair is yet another coincidental mic technique. It uses a pair of figure 8 microphones, one above the other. They are angled 90 degrees from each other so their figure 8 patters combine to form a cloverleaf pattern. Aim the mics at the sides of the sound source (like an XY technique), and pan them opposite each other. This technique provides excellent localization, but also picks up sound from the rear.
Spaced pair stereo recording technique
Spaced pair microphones are the most basic setup for stereo recording. Two microphones are set up apart from each other to record separate signals for stereo recording.
In an AB setup, you take two mics, usually omni but they can be directional, and set them up side by side, however far apart you want them. The gap is arbitrary, from really close (about 2-12 inches) to far apart, (the width of an orchestra maybe).
The spaced pair uses a different principle of sound to relay a stereo image – time differences. When the microphones are spaced apart, sounds coming from the right side of the soundstage will reach the right microphone first. There will be a slight delay until it reaches the left, but your brain can hear that distance. When spaced pair stereo sound is played back on speakers, the right signals play from the right speaker slightly before the left. Your ears hear this time delay, and can instantly place it in the soundstage.
When you are setting up a spaced pair microphone set, you may find yourself placing the mics further apart to get a better mix of the whole ensemble. This is good, but after a while you start having a "hole" in the middle. Putting a third mic in the middle is a quick and dependable solution to this.
A variation of this setup is called the Decca Tree, named for the tree-like placement of 3 microphones. There is one for each side and one for the middle, but usually the middle one is front further than the side ones. The usual distance is half as far front as the side mics are apart. This option offers good localization with excellent stereo spread.
When you compare coincident to spaced pair stereo recording, you have two different styles. Coincident is sharp on localization, but slightly narrow in stereo image. Spaced pairs have a much nicer sound when it comes to width and ambience, but the image isn't quite as sharp. The wider you place the mics apart, the wider the image becomes, but the localization is a bit fuzzy.
Another drawback is phase problems. If your target media will always be stereo, like a CD, you don't need to worry about it. But if you expect that it will be listened to in mono, you might want to listen to how it sounds. The microphones might be out of phase with each other, causing phase cancellation when combined to mono.
The third type of stereo recording mic setup we'll look at is called near coincident. While the first two used level or time differences to make the stereo image, near coincident offers both. Let's look at how it's done.
Two directional mics are placed in an xy pattern, but instead of putting the diaphragms at the same spot horizontally, you space them apart about 7 inches. Often they are angled out more than 90 degrees.
This stereo recording technique offers the sharpness found in coincident patterns combined with the width and spaciousness of a spaced pair. It builds a more accurate stereo image using time and level differences, one that is understood by the brain better.
By far the most common near coincident technique is the ORTF. (It's an acronym in French for the French broadcasting system.) It says that the diaphragms must be 17cm apart, (about 7 inches), and angled 110 degrees from each other. You can either use two mic stands for this or get a stereo bar and mount both mics on one stand.
Some other near coincident techniques are NOS and DIN coming from Holland and Germany respectively. The NOS technique specifies the spacing of the mics to be 30cm with an angle of 90 degrees, while DIN specifies a spacing of 20cm and an angle of 90 degrees.
Which of the three stereo recording setups is best? I hope this doesn't surprise you, but there is no best one. What! Yep, if you pressed the question "Which is better, ORTF, NOS, or DIN?" to me, I might respond with "yes!" They each have their own sound. I tend to like the ORTF technique, but don't take my word for it. Go try yourself! But do it in some spare time so you'll know before the recording session what might work better.
All of the stereo recording mic setups we looked at so far use either time and/or level differences to create a stereo image. We can get a fairly good stereo image from this, quite impressive actually. But let's think for a bit how we as humans hear.
We have two ears separated by a dense object in the middle. (Hmm, who's is denser, mine or yours? ;-) ) Our ears work as omni mics, although separated by the head. The pinnae around our ears reflect the sound into the ear drum, separating the frequencies by where they come from. The [physical] head also does this, filtering out higher frequencies in the ear opposite the sound source.
When our brain hears all these patterns (level, time, and frequency differences) it decodes the information to localize exactly where the sound came from. The time and level differences are close to what is created with a near coincident mic technique. But our natural stereo hearing system uses three principles to form a stereo soundstage -- time, level, and spectral differences. How can we get something that mimics as close as possible the natural stereo hearing we use all day long?
This study is called binaural recording, trying to get a recording that sounds like you were there. There are two main ways of doing this, a baffled omni setup, or actually using a dummy head with microphones where the ears would be.
The baffled omni is easier to set up. Basically it's two omni microphones with a baffle in the middle to separate the stereo image. A Jecklin disk is a common way to set this up – take a 12inch round piece of hard material (plywood or plexiglass) and put the mics on either side.
You can buy some disks that look professional or you can make one that serves the same purpose. When you have this ready, you mount the mics on each side with the diaphragms about at the center of the disk, separate them about 7 inches, and angle them out slightly. This mimics the position of a human head and ears, giving a stereo image based on level, time, and spectral differences.
The other way to get a binaural stereo recording is to use a dummy head and put microphones in the spots where the ears should be. There are dummy heads that closely resemble a human head -- complete with nasal passages -- providing very accurate filtering resulting in superior imaging.
Summary of stereo recording techniques
Phew! That was a long way, so let's tie up the loose ends. Here's a brief summary of each stereo recording technique discussed.
Coincident stereo recording
- Uses level differences to produce stereo image
- Provides a sharp image, excellent localization
- Soundstage is a bit narrow
- Signals are mono-compatable
Spaced pair stereo recording
- Uses time differences to produce stereo image
- Provides more diffused stereo image, localization is a little fuzzy
- Soundstage or stereo spread can be exaggerated unless a 3rd mic is used
- Provides a nice sense of ambience
- Signal may have phase problems when combined to mono
- Can use omni mics, so low frequency sound can be better
Near coincident stereo recording
- Uses level and time differences to produce the stereo image
- Provides a fairly sharp stereo image with good localization
- Stereo spread or soundstage is accurate
- Gives more air and depth than coincident methods
Baffled omni stereo recording
- Uses level, time, and spectral differences to produce the stereo image
- Stereo image is sharp
- The stereo spread tends to be very accurate
- Realistic, more what your ears hear
- Low frequency sound is excellent
You can use these different techniques for many effects in stereo recording, and I encourage you to try them out. What happens with each one? What are the characteristics of each? Let me know; I'd love to hear from you!
Return to the main stereo recording page.