Many of us use synthesizers as the backbone of our tracks, but few of us take the same care when recording them to tape as we do with electric guitar or vocals. Dave Lockwood explains that there's a lot you could - and perhaps should - do to make your synth sounds more original and exciting.

The increasing use of sequenced 'virtual tracks' and the advent of affordable digital tape systems is undoubtedly changing the way in which synthesizers are recorded. By removing the need for 'treatment' solely to compensate for conventional tape limitations, more scope is created for processing purely for enhancement. Even where sequenced parts still have to be printed to multitrack tape, the fact that the performance is safely stored in the sequencer, and can therefore be repeated as many times as necessary, encourages much greater experimentation with the sound.

LEVEL MATCHING

Level matching is fundamental to recording any source. Too much level into a sensitive input will result in distortion, whilst too little signal, requiring excessive gain from the input stage, will result in noise. Curiously, many modern synths seem to have a typical output around -20dB, which is really neither mic nor line level.

If you find that your mixer's Gain controls have to be advanced to near maximum before you can get enough level, it is probably worth using the microphone input instead. This invariably means that you will need to interface a jack output with an XLR input. The best answer is usually to use a DI box as an adapter, as this will also give you an earth-lift facility (see the May issue of Recording Musician for more information on avoiding earth-loop problems).

If you don't have a spare DI, however, you can make up your own set of 'un-balancing, earthlifting', XLR-to-jack cables specifically for this purpose (see Figure 1). In general, no impedance mismatch will occur through omitting the DI box with keyboards. The one constraint to bear in mind is that you should ensure that any microphone 'phantom power' in the desk is disengaged on the channels you are using.

WARMTH AND COLOUR

Scope for creative equalisation of synthesized and sampled sounds is actually subject to much the same constraints as with 'real' sounds — too much HF in a good-quality string section sample will make it sound just as unpleasantly harsh as overdoing it with the real thing. The more interesting effects, however, are generally achieved by methods that alter tonality in more complex ways than just EQ. One simple, but very effective technique to add warmth and colour to a synth output is to feed it through an instrument amplifier, especially a valve amp, and mic up the speaker output (Figure 2). The amp must be in another room, however, to allow you to hear the effect on the monitors. If you run your synths 'live' via sequencer, you can continue to tweak amp settings and mic placement right up until your final mix; recording the signal to multitrack, on the other hand, has the benefit that you don't have to set up the amp and mic every time you want to work on the track. It's also worth experimenting with mixing the mic signal with the direct sound (also try phase-reverse on one of them).

Processing a bass guitar sample through a bass amplification rig can make the sound far more authentic and exciting by adding the missing elements of the 'thump' of speaker-cones moving lots of air, plus colouration from cabinet resonance and the microphone. This method can also give a different kind of depth to to that available from EQ alone to purely synthesized bass sounds, helping to avoid the common effect of all the bottom end disappearing from the mix whenever the bass moves out of its lowest octave.

SUBTLER SHADES

Another much-favoured technique is using guitar DI devices, such as the Rockman range, or units from Zoom, ART and many others, to process synthesizers. It is well worth experimenting with using these not just for all-out distortion. Set up for cleaner sounds, they can add subtler shades and useful filtering to change the character of a sound — don't restrict yourself to thinking solely in terms of wailing guitar simulations.

Dedicated speaker simulators, such as Hughes and Kettner's Red Box, or the active unit from the Palmer range, can also make a good job of rounding out a synth sound, without totally altering its character. They can even go some way towards giving a bright, modern digital synth the subjectively 'warmer' sound of an old analogue model. The big limitation of all these DI units, however, is that they tend to be mono.

When using either amps or DI devices, pay particular attention to level-matching once again — even a low-output synthesizer is capable of a significantly higher level than most guitars, and front-end clipping, or input overload (distorting the first amplification stage of a unit, usually before the first level control in the circuit and therefore unaffected by turning this down), is an audibly unpleasant form of distortion, even in a valve amp).

COMPRESS FOR EFFECT

When recording real instruments, compressors and limiters are used both to control extreme peaks in level, and to bring up the volume of low-level material. With synths however, particularly when sequenced, the degree of unpredictability encountered in real instruments is removed — you can optimise level to tape in a trial run. This removes any need to compress other than for a deliberate 'pumping' or 'squeezed' effect, or to bring up the level of subtle developments in the decay phase, which might otherwise be masked.

You might well expect samples of real instruments which are invariably compressed on recording, such as bass guitar or clean electric guitar, to benefit from being similarly processed. However, bass and guitar samples themselves will often be already compressed, both to aid in the sampling process and to make the sound a bit more 'polished', and more instantly like the sounds we are familiar with on well-produced records. Applying compression again usually results merely in noise and a dull, lifeless sound.

Ultimately, recording synthesizers differs from recording 'real' sounds primarily in that the normal psychoacoustic reference points to scale and tonal neutrality are usually absent. Apart from the case of high-quality versions of recognisably imitative sounds, there is nothing for the brain to compare to, making radical experimentation just as likely to yield worthwhile results as a more conventional approach. If it sounds good to you, then do it!