email the curator

Visit our light-hearted music gift shop!

This article is intended as a brief guide for the amateur audio engineer who wishes to have a go at recording the "King of Instruments". I am going to concentrate on the most important aspects, which are microphone choice, configuration and placement, and take it for granted that a good quality recording medium is available (DAT, CDR, hard disk etc., or open-reel if you're an analogue fan!)

Listen to a sample of Paul Arden-Taylor's recording here

The Room
This equipment should be set up in a room such as a church vestry or town hall dressing room, sufficiently isolated from the main church or hall to allow you to monitor the sound exclusively via the recording system. It is useful to have a pair of loudspeakers, for playbacks and for judgement of the sound as most listeners will eventually hear it, but I recommend the use of headphones during recording to check for extraneous noises which could be missed on speakers.

Communication
Unless the recording is of a "live" recital, you will also need some means of communicating with the organist, such as a talkback unit or a pair of walkie-talkies. Because the microphone cables will be fairly long, they should be "balanced" (i.e. 3 connections per channel: positive, negative, and screen) and of high quality. Unbalanced systems, in which the screens double as the negatives, are to be avoided, as they are susceptible to hum and interference when long cable runs are used.

The microphone
If your aim is to make impressive-sounding recordings, it is essential to budget generously for microphones, which are the first and most critical link in the recording chain. Cheap battery-operated mics of the type intended for use with equipment such as portable recorders and camcorders, or "vocal" mics primarily intended for use with public address systems, will not exhibit acceptable signal-to-noise ratios or sufficiently extended frequency response.

Professional capacitor "studio" microphones are called for, and if selected and set up appropriately, a simple "stereo pair" is normally all that is required to produce a faithful, natural-sounding recording. Mics of this type generally require a few milliamps of "phantom power", an elegant system by which power is supplied to the mics via the same balanced cables which carry the audio signal from them. This power (usually 48 volts) is supplied by the mic preamp, which should be a top quality 2-channel unit (I invariably use an Amek 9098, which was designed by the legendary Rupert Neve).

Alternatively, two channels of a mixer may be used. Mackie, for example, manufacture small portable mixers which incorporate mic preamps of remarkable quality, together with phantom power.

Multi or omni? Which shall it be?
Microphones come in all shapes and sizes, but may be divided into two broad categories: directional and omnidirectional. As far as organ recording is concerned, most directional mics have two disadvantages: (a) their low frequency response tends to fall off when placed at some distance from the sound source, as is the case in a typical organ recording setup, resulting in a disappointingly bass-light recording, and (b) they reject some of the ambient sound, resulting in a somewhat dry recording.

Most engineers, therefore, will choose omnidirectional mics for their consistent response throughout the frequency spectrum and greater ambient sound pickup. Note that omnidirectional mics actually become slightly directional at high frequencies and should, therefore, always be "pointed" at the sound source.

Hearing double and how to configure it
As far as the matter of configuring two microphones for stereo recording is concerned, there are two main camps: spaced and coincident. Learned and lengthy scientific papers have been written on this subject, but in practice it all boils down to a few salient facts.

A spaced pair consists of two identical mics (which may be of any polar pattern, but omnidirectionals are normally chosen for the aforementioned reasons) mounted apart by 30cm or more, directed at the sound source. With omnidirectionals, this configuration captures the full frequency range and gives a pleasing sense of "air" around the sound, but the stereo image can be somewhat amorphous and unstable, and there is a risk of phasiness, comb-filtering effects and a "hole in the middle".

A coincident pair consists of two mics of identical directional polar pattern (cardioid, hypercardioid or figure-8), mounted with their capsules as close as possible to each other, forming an angle of 90 degrees or greater. An imaginary line bisecting the angle between the capsules is directed towards the centre of the sound source.

This configuration affords a sharp stereo image, but as directional mics are involved, the recording may be somewhat dry and bass-light, and there may be inconsistencies in frequency response due to the mics being orientated off-axis to the direct sound source. An interesting variant of the coincident pair configuration, known as "MS" (mid/side), overcomes these problems to a certain extent by having one mic (the "mid", which may be of any polar pattern) pointed directly at the sound source, and the other mic (the "side", which must have a figure-8 pattern) orientated left-to-right across the sound stage.

A special MS decoding circuit is required to derive a standard left/right stereo signal from this pair, but the advantages are that the mid mic is directly on-axis to the sound source and may be an omnidirectional, the stereo image is extremely precise, and the stereo width may be controlled remotely by adjusting the level of the side mic relative to the mid mic, rather than by physically moving the array.

There are various "semi-spaced" configurations which, as the name implies, are compromises between the spaced and coincident approaches (e.g. "ORTF"), and still others which involve the use of some sort of acoustic obstruction between the microphones to improve channel separation (e.g. the "binaural" or "dummy head" configuration).

Having experimented with most of the established methods of configuring two microphones for stereo recording, my own preferred solution is to use a pair of industry-standard omnidirectional mics (Bruel & Kjaer 4006), forward-facing and slightly divergent, placed either side of a slim, acoustically-opaque, non-reflective baffle ("Jecklin disk"). This arrangement delivers the warm and spacious "omni sound" thanks to the smooth, extended frequency response and generous ambient sound pickup of these mics, combined with a focussed, stable stereo image due to the presence of the baffle. There are no apparent drawbacks to this setup, and I am pleased to report that a number of CD reviewers have approved of the results!

Getting into position
Having selected and configured the microphone pair, the remaining crucial consideration is the positioning of the pair within the recording space. Like all "classical" recordings, a good organ recording should not only capture the sound of the instrument as faithfully as possible, but also convey a sense of the acoustic in which the instrument is placed (it has often been said that the most important stop on the organ is the building in which it is installed). Place the mics too close to the organ and the sound will be uncomfortably dry and "in your face", with individual pipes being highlighted rather than blending into a coherent whole, and there will be little sense of the acoustic surroundings.

Also, any problems such as wind leaks and mechanical clatter will be horribly exaggerated! Too far away and the recording will be muddy and over-reverberant, dull-sounding (due to atmospheric attenuation of high frequencies), and generally distant and uninvolving, with a very narrow stereo sound-stage.

Finding the "sweet spot"
The trick is to find the "sweet spot", where the organ pipes blend into a homogeneous, transparent ensemble with a pleasing full-width stereo spread, and where there is an appropriate ratio of direct to reverberant sound, combining warmth with detail. Finding this sometimes elusive location will obviously require some initial experimentation in each venue, and it should be borne in mind that the optimum mic position may well be somewhat higher and closer to the organ than might be suggested by live listening tests; human ears are able to "zoom in" on a sound source, a feat which microphones cannot emulate.

As a general rule, where the organ is speaking along the main axis of a building from one end (the ideal situation), start with the mics about one third of the way down the building from the organ. In a very reverberant acoustic it may be necessary to move the mics closer to the organ, or further away in a drier acoustic. Try to keep the mics well away from walls, which can produce undesirable reflection effects. In less satisfactory situations, such as where the organ is speaking in two directions from a corner of a church (often encountered in England), try to find a sweet spot somewhere along the centre line of the building, from where the mics can "see" both case fronts.

Where the organ is in a transept, try placing the mics in the opposite transept. If this results in a "boxy" sound due to the proximity of the walls, move towards the centre of the crossing. In all situations, use a tall stand so that the mics are at least 8ft above floor level; ideally, they should be midway between floor and ceiling if possible! If the floor is wooden, place the stand's feet on pieces of carpet or stiff foam to isolate it from any vibrations.

Proof of the pudding
Finally, ask the organist to give you a test blast on full organ! During this, set your recording levels to read almost maximum level on each channel, leaving a small safety margin (digital recordings must never be allowed to go "over"), and you should not have to make any further adjustments throughout the entire session. Good luck!