Stereo Recording Techniques
A more advanced guide to various stereo microphone techniques.
When recording voices - either sung or spoken - or solo musical instruments, you will usually use a single microphone, recording in mono, and then pan it within the stereo soundstage, and add digital reverb and other stereo effects in post-production and mixing. In this way, multiple instruments can be recorded one at a time, and placed within the stereo (or surround) image.
There may, however, be occasions when you want to record in stereo to capture a larger group, or a nice acoustic. Stereo recording is a more advanced and demanding skill, and allows the recording engineer to make technical and creative choices, which will affect the final result.
There are several different techniques used for capturing a stereo audio image, and each has its relative merits, be they simplicity, flexibility, realism, or mono and surround compatibility. Here we cover the most widely used of these methods, with particular attention paid to the mid/side (m/s) configuration, as it offers the most variations in implementation and interpretation.
If you are not in a position to spend time manually configuring a stereo recording rig, then preconfigured stereo microphones offer ‘out of the box’ stereo recording capabilities, and are set up to give an optimised result for their chosen specifications and characteristics. All of them will, however, use one of the following techniques to obtain their stereo image, so even if you don’t intend to set them up yourself, then this section may provide useful background information to help in choosing the right stereo microphone for your application, by explaining what they are doing.
Stereo microphone configurations
1. Coincident X-Y pair
Two directional microphones are placed coincidentally (ie with their heads as close together as possible), and at 90° to each other. Each microphone represents one side of the stereo field.
Røde NT4 co-incident X-Y stereo microphone
As the microphone heads are very close to each other, any phase differences between the signals, caused by the minute differences in timing of the sound waves’ arrival at the two diaphragms, are minimised. While this minimises phase problems, it can be argued that these small timing differences are representative of those found naturally. This dichotomy leads to the alternative, but related technique of the spaced pair :
2. Spaced pair
In its simplest form, the spaced pair comprises two omnidirectional microphones placed a distance apart. Alternatively, directional or figure-of-eight mics may be used, at a variety of distances and orientations. Again, each mic represents one side of the stereo field.
In one particular implementation for example – the ORTF technique - two cardioid directional mics are placed 17cm apart and at an angle of 110° to each other.
Though this technique can yield natural and convincing results, the space between the pickups inevitably leads to phase cancellation at some frequencies and phase addition at others, leading to ‘comb filtering’ when the stereo signals are summed back into mono. If one of the final destinations of the recording is any kind off mono playback device, the spaced pair is not particularly suitable.
One microphone, of any pattern, is placed facing directly forwards, towards the centre of the soundstage. A second mic, with a figure-of-eight polar pattern, is placed coincidentally, as close as possible to the first, with the two 'lobes' of the 8-pattern facing left and right.
By combining the 'mid' signal with both in-phase and out-of-phase versions of the fig.8 'side' signal ("+side" and "-side"), the two sides of the stereo field are derived :
left = mid + side
right = mid - side
This technique has the major advantage of perfect mono compatibility, since when added together, the two 'Side' signals cancel out perfectly, leaving only the mono 'Mid' signal remaining. Indeed, by mixing in different levels of Side signal, the stereo width of the image can be altered after recording, simply by varying the ratio of Mid to Side signals.
There are various methods of summing the Mid and Side signals in the MS “matrix”. You can use a proprietory MS decoder, or, with a little ingenuity and a mixing desk, perform the phase inversion and channel summing yourself, and then experiment with balancing the signals to give different stereo images.
The MS technique offers infinitely variable implementations, thanks to the ability to use any mic for the Mid signal, and to vary the ratio between that and the Side signal as desired.
Below are a couple of representative examples for illustration; based on polar response patterns [see Microphone Guide]: the black lines represent the Mid mic’s polar pattern, and the blue lines the Side (Figure-8) pattern. The red and green shaded areas are the left and right composite outputs, after Mid and Side signals have been combined by the MS matrix decoder. You can see the difference in effect on the outputs, between using a cardioid and an omnidirectional as the Mid microphone:
4. Blumlein pair
Named after Alan Blumlein, an engineer at the forefront of developing stereophonic sound in the 1940s, the Blumlein pair is similar to an X-Y co-incident pair, but uses figure-of-eight mics instead of directional. Thought by many to give a more natural sound than the cardioid mics of the standard X-Y technique.
In an attempt to mimic the arrival of sound at the human ear, and the acoustic effects of having the skull in the path of the sound waves, binaural recording uses a dummy head, which is placed between two microphones. The mics are often recessed into the head at the ear positions.
In a simpler model - a variation on the spaced pair technique - the dummy head is replaced by an acoustic baffle, or 'Jecklin disc', which has a broadly analagous acoustic effect.
Neumann KU100 dummy head