Unlike the professional sound recording studio, which contains many devices to allow the engineer complete control over the sound at any point in the recording chain, the field or home recordist must exploit any 'trick of the trade' or Heath Robinson-type device to obtain the ultimate recording quality from the equipment, as signal processing and filtering equipment is usually costly and would probably be second choice to an echo or phasing device anyway if the money were available.

There is little comfort in the fact that wherever you position a microphone, be it indoors or out, there are going to be extraneous noises or undesirable room acoustics colouring the purity of the original sound and whereas it is possible to deaden down 'room boom' with a few carpets, upholstery, and mattresses the outdoor recordist's lot is either that of living with the recording eventually taken down or selecting carefully the type of microphone used, checking that it is more suited to the type of subject intended to be put in front of it. Even if the signal-to-noise ratio of your tape recorder were 80 dB it would mean nothing with a low 'signal-to-ambient-noise' ratio. Getting a good signal to ambient noise ratio is obviously achieved by getting the microphone closer to the subject but often noises appear which are completely out of your control at the time of recording. One can only hope and pray that it will be possible either to edit out the stupid people who continually walk up to the microphone and say 'one two one two testing' because they think it is funny, or that it will be possible at some time to clean up the low frequency rumble of distant traffic or aeroplanes using a graphic equaliser or filter. Even using filters, though, can partially destroy the complex pattern of frequencies, harmonics and waveforms which make every sound unique, but nothing will remove the transistor radio from the beautiful Welsh valley recording except, at the time of recording, walking up and smashing it to bits with a noise abatement flag marking the spot where it happened.

The most important feature of any sound, apart from its frequency content, is its dynamic range. The level at which someone sings, a guitarist plays, or a person speaks is a part of the communication process, an essential part, and the sound recordist must endeavour to retain as much of it as the equipment will allow. From the tape recording point of view the limits to this dynamic range off tape are hiss — tape hiss and system hiss — at the low end, and distortion, tape and system, at the other. Any recording equipment operates best at an input level which is known to be the best compromise between these two extremes — this level is referred to as 'operating level' and usually aligns with zero on the VU meter.

Any recordist knows, though, that keeping the input level of a signal within the range of a VU meter is already compressing it to about 20dB and the parts of the signal which were at the bottom end of the meter will be heard with hiss. Hiss is not as objections as peak distortion, though, where the input level has been so high that the pre-amps of the recorder have squared off the peaks of the incoming signal. When this happens the VU meter needle is on the endstop so eventually there is a tendency for the input level to be held well back and money spent on some type of noise reduction equipment to combat the resulting problem of hiss. Either that, or a tape recorder with some type of input gain control device is sought, as 'riding' the input level control is tricky and it's not always possible to know in advance what the sound level of your chosen subject will be — an atomic bomb explosion for instance (don't hang around or the hiss may be indistinguishable from the frizzle).

Noise reduction devices give an extra dB of dynamic range because they lower the hiss level by that amount. Unfortunately, noise reduction equipment is expensive and beyond the budget of the home recordist. However, modern recording tape can be 'pushed' slightly — operating level can become +4dB or even +7dB in some cases. Many recording studios are now operating at NAB +4dB on stereo reduction machines and at +6dB on some multitrack machines. When you consider that a sound-level change of 3dB is twice as loud then increasing input level by 3dB is making the noise floor twice as quiet, subjectively anyway. This is one way of combating tape hiss. Most 'noise' inherent in a tape recorder has its source as the erase head and the only way to overcome this is to disconnect it and work only on fresh tape stock. If the noise from the tape recorder is excessive it may be simply that the heads need demagging: in which case erasure of high frequencies will often be apparent.

Input gain control devices are often pretty useless for recordings of anything, other than speech and music programmes which are being bootlegged off the radio. This is because speech is of limited dynamic range and radio programmes are often heavily compressed and limited anyway in an attempt to protect the transmitter against overload, to squash the dynamic range into the limited bandwidth of the transmission medium, and as a safeguard against trainee studio managers who may not notice that the PPM is peaking at 8. The preset attack and release times of an automatic gain control device are often too slow, otherwise, and an acoustic guitar can end up sounding as if it was played backwards. Drum kits may sound pretty interesting, though, as the 'pumping' effect may help to tighten up an otherwise 'untight' track. To exploit this effect during, say, a 4- or 8-track or overdubbing session with Revoxes, simply route the drum kit, say snare or hi-hat, will have the 'bounced' through the front end of a cheap cassette recorder with a built in AGC device. Then take the monitor output of the cassette recorder back to the mixing desk or Revox overdub track. Whatever the drummer hits loudest or the engineer pushes up to the front on the drum kit, say snare or hi-hat, will have the effect of holding the rest of the kit down until the AGC device has recovered, and when it does, the rest of the kit will fade back in. Imagine that the drummer has just done a lick around the toms and then crashes the cymbals: because the AGC has been working overtime holding the loud smack of the toms down, when the cymbals are hit (which are not as loud), the AGC will already be recovering and this recovery will hold up the fade of cymbal crash. A similar effect is achieved in the recording studio by routing a submix of the drum kit through a compressor sidechain when the attack and release times of the compressor are adjusted for best effect. Even if the effect isn't used at the mixing stage it is interesting to add reverb to it and mix it behind the normal mix of the kit to obtain the open-mic 'large room' effect.

Some portable tape recorders have built-in limiters and these are preset to keep the peak level at a certain figure. The Nagra 4.2 and 4S models incorporate a 'record with or without' limiter facility, for example. And it is always handy to be able to work the input level of a recorder against the ceiling set by its limiters, depending upon how good the limiters are of course. Effectively one is then compressing the dynamic range of the subject and thus bringing the quiet sections above the hiss level, the loud sections being taken care of by the limiters (which, as I said earlier, are preset to a +VU figure anyway). It is often better to determine this input level by listening to the increase in ambient noise level off-tape so that you can determine the best compromise between hiss and ambience at the time, rather than trying to figure out where the -40 mark is on your VU meter. Also, monitoring in this way will help you ensure that the resulting input level in the loud passages isn't overloading the limiter and causing it to clip.

Limiters are sometimes confused with compressors which work in a slightly different way. A compressor is an amplifier with an input-to-output ratio which is constant until a threshold is reached. Above this threshold the ratio changes such that the output begins to fall off and so gain reduction is achieved — see fig 3. Compressors are often used in recording studios to smooth out erratic guitar levels, flutes that have been played with too much vibrato, and singers who have had too much brandy and are swaying around the microphone. Limiters are helpful in protecting the inputs of one piece of equipment from the outputs of another as, for example, when interfacing a Vocoder with a synthesiser. However, because compressors and limiters are often used in the final stages of the mixing process and compressors especially have a way of lifting ambient noise of hiss to an undesirable level, it is necessary to make sure that the recordings themselves are noise-free and undistorted.

A certain amount of any noise is masked by the signal content itself and the noise only becomes apparent when there is no signal present. Noise which is above or below the frequency bandwidth of the signal may be heard but it is possible to filter this off using equalisers and noise filters. But again, when there is no signal present the noise which was originally being masked will come to the fore. A device which goes a long way to overcoming this problem is called an expander and works in an opposite way to a compressor in that it is a linear amplifier until the input level falls to a certain — dB figure. Then the output falls steeply, by as much as 18dB per octave or, with a 'noise gate', the signal is attenuated completely and will only reappear when the input level rises again above the threshold — see fig 4.

Thus, using an expander it is possible to clean up a track by setting the expander threshold just below the level of the signal content. An apparent increase in dynamic range is the result, and during mixing of many tracks the necessity of fading the noisy track in and out is alleviated. Expanders were used originally by the film industry for transferring speech recordings to optical film soundtracks — this accounts for the 'choppy' nature of speech on old movies, where the expander/gate has removed the original attack and decay characteristics of the spoken word. Expanders are expensive but I'm sure it won't be long before I find one free with my corn flakes.

In the home studio 'cleaning up' usually means roping in some old bass and treble controls from some smashed-up transistor radio, but if you can't even find one of those I've drawn the circuit diagrams for simple high- and low-pass filters which can be built by anyone with the ability to twist two bits of wire together and save 5Op out of their beer money to buy the squibs with. They should be connected between the inputs of the copy recorder and the outputs of the playback machine. The values of the resistors and capacitors were formulated for a voltage level of 0.775 volts which is 'line' or 'zero level' from most tape recorders. Resistor/capacitor networks form the basis of many equaliser modules and experimenting with different values of R and C will show that these circuits can be built to work at a range of different frequencies. However, simple equalisers may remove some of the harmonic spectrum of a sound, so a new breed of equaliser has come to prominence — the 'parametric' equaliser. Unlike conventional low, mid, and high, pass, peak, or cut filters, parametric equalisers usually have low to low-mid, low-mid to high-mid, high-mid to unbearable bands which make it possible to remove a band of frequencies in one section, say 200Hz while peaking 250Hz in another. This is possible because not only do the frequency bands overlap, but the 'Q', effectively the bandwidth of each section, is tunable from quite wide to very, very narrow and it is possible to not only remove unwanted sibilances but to bring out voices etc which may have got lost when Concorde went over.

To avoid the necessity of equalisation is the best remedy, however, and achieving a good signal-to-ambient-noise-ratio by close miking is often the best policy. If it is not possible to get a microphone close to the subject then a directional mic is a good standby; a device which is often overlooked by the amateur recordist is the Parabolic Reflector. This is a saucer-shaped disc which focuses the sound waves of whatever it is pointed at on to the diaphragm of a microphone which is mounted slightly in front of its focal point. Any microphone may be used — this is one advantage, but another is that the parabolic reflector works only on frequencies whose wavelengths are shorter than the diameter of the reflector. Thus, a 24in dia reflector will work only above 500 Hz. So the recordist is not only gaining an extra, say, 14dB of gain for the subject but also a built-in high-pass filter too. It is possible to mount a stereo pair of microphones into a parabolic reflector but I have found that the stereo image is slightly compressed. Parabolic reflectors are suited to long distance work but were often used in film studios to overcome the problem of not being able to get a microphone too near to the subject and overcome the other problem, that of boomy studios which, being large, had low frequency room resonances.

Exact details of where to buy or how to make parabolic reflectors (I once made one out of papier mache and silver foil and it were champion) are available in books on wildlife sound recording by Eric Simms and Richard Margoschis, the former from most bookshops and the latter from Print and Press Services in Barnet, Herts.

The Parabolic Reflector has now been replaced in film recording circles by the 'shotgun' type directional mic like those featured in the photograph in the previous article (The Compleat Sound Effects Recordist, Jan 79) which were Sennheiser 816s — no prizes for guessing — and no I don't get a hole in the middle, and yes the bird was a Red Kite but it wasn't real, and the location was Snowdonia and thanks for all the letters, when am I going to get the tapes..., mainly because they are light and not cumbersome although small 12in parabolic reflectors are now available.

Many microphones are corrected in some way: either their response is tailored to a certain type of programme content or they incorporate small switches which select either a low frequency rolloff or presence boost, both of which are extremely handy for speech, vocals or a nice clean acoustic guitar sound. Outdoors the Beyer M201 mics are useful because they are very robust and have a good back-to-front ratio. The small capsule size makes them fairly free of wind noise if the windshield provided is used. Dynamic microphones are better suited to outdoor work because excessive humidity can cause the capsules of condenser microphones to clap out and make stomach-like gurgles, although I have been very fortunate in being able to use a Schoeps CMTS 501 stereo mic outdoors with a special windshield around it.

It is really a studio microphone but as soon as I tried it with the Nagra 4S I was hooked. Recording carnivals and brass bands on a Blumlein figure-of-eight pair with a mic of that sort of quality is like getting photographs on 10x8 transparencies. The quality of the recording is unbelievable. The frustration comes when you've recorded just about everything within the immediate 50-mile radius that's of any worth and then you go to an incredible football match where the atmosphere is electric and the fans are singing and you want to record it in Binaural with the mics positioned right in the middle of the centre line. I came pretty close to that sort of experience when I recorded stock-car racing at Hednesford Raceway in Staffordshire and positioned the microphone about six feet away from the figure-of-eight race crossover point. On the resulting recording the cars are just everywhere, left/right/back/front/smash/bang/wallop. I tried to get a recording of a worm going by with two ECM 50s mounted either side of the garden gnome's head for binaural effect, but the worm went the other way.