There are now many different ways to record electric guitars, from miked-up valve amps to physical modelling preamps. So how do you decide which to use?
The theory of evolution says that the longer something has been evolving the more complex it tends to get, and this is certainly true of the electric guitar, which has been evolving for over half a century. Electric guitar sounds rely on the instrument itself, the amplifier through which it is played and also on the loudspeaker system used. Further variables are introduced when miking techniques are taken into consideration, though these days miking is only one of the ways of recording an electric guitar — we also have a number of effective DI techniques from which to choose.
Guitar amps tend to use 10-inch or 12-inch speakers without tweeters or crossovers, so they have a very limited upper-frequency response. These speakers may be used singly or in multiples, in either sealed or open-backed cabinets. The familiar overdrive sound was almost certainly discovered by accident when early amplifiers were driven beyond their design limits in an attempt to obtain more volume, but because of the restricted top end of the speaker systems employed at the time, the distortion was stripped of its more abrasive upper harmonics and actually sounded quite musical. So, what started out as a side effect of limited technology soon became adopted by blues players and turned into a distinctive style, which later evolved into rock, and then into heavy metal with all its spin-off genres.
As a rule, open-backed cabinets tend to have a different low-frequency characteristic to closed ones, partly because no air is trapped inside the box to act as a pneumatic spring. One characteristic is that low-frequency sounds, such as damped lower strings, cause the speaker cone to move a considerable distance, producing what is affectionately known as cabinet thump. In addition, there is interaction between the sound coming from the front and the back of the cabinet, which may cause some frequencies to cancel and others to be reinforced.
Then there's the issue of valves. Serious guitar players invariably prefer the sound of valve amplifiers to any form of solid-state circuitry, but the technical reasons are not as obvious as you might imagine. We all know that valves distort nicely when driven hard, but the use of output transformers also affects the sound. Then there's the choice of Class-A or Class-B (sometimes referred to as AB) power stages — Class-A stages clip asymmetrically whereas Class-B power stages tend to clip symmetrically. Even the make of valve has an effect, as do technical issues such as the choice of triode or pentode output valves, or even the types of capacitor used in the circuit. Because there are so many variables, not including the most important one (playing technique), the electric guitar is capable of a vast tonal range.
The guitar's contribution to the sound comes from its mechanical construction, including the type of wood used, and the pickup system fitted. The hugely popular Fender Stratocaster uses simple single-coil pickups, which tend to give it a bright, articulate sound that doesn't take up too much room in a mix. Guitars with humbucking pickups, on the other hand, tend to have a thicker, more solid sound that can overpower other guitar parts or other mid-range instrument sounds. Tonal qualities aside, humbuckers by their very nature are designed to reject electromagnetic interference, while single-coil pickups are very susceptible to it. Specialist stacked humbuckers, such as those made by Kinman, Dimarzio, Fender and others, are available for use where something close to the original tonality is desired but without the noise problems, and for serious studio work these are a good option. Note that CRT computer monitors emit a lot of electromagnetic radiation from their scan coils, so a flat-screen LCD display is always an advantage if you need to record guitar into a computer system.
Even when using noise-cancelling pickups, guitars tend to be relatively noisy sound sources, especially when used in conjunction with high levels of overdrive, either from a pedal or an amplifier. High levels of gain are needed to produce overdrive sounds, so any source noise will be increased accordingly. If an attempt is going to be made to remove some of the noise by electronic means (such as a gate, expander or single-ended noise-reduction unit), it may be best to leave this until the mixing stage, so that an incorrectly set gate or noise filter doesn't ruin a good take. However, gates should be used before delay or reverb effects to prevent the tail end of the sound being shut off abruptly when it falls below the gate threshold. If you simply have to record with delay effects, a dedicated guitar gate pedal placed before the delay unit will probably work best. If you are also using an overdrive pedal, then the gate should come after this so that it can deal with circuit hiss as well as hum from the guitar pickups. Note that all-in-one floor units tend to include gates at an appropriate point in the signal path, between the overdrive/amp section and the effects.
In the studio, a dynamic noise filter such as the Symetrix 511A, Drawmer DF320/330, Rocktron Hush or Dbx Silencer can be less obtrusive than a gate for cleaning up guitar parts to which delay/reverb has not yet been added. Very generally, such devices work by progressively reducing the audio bandwidth once the sound falls below an adjustable threshold. Transients pass through with very little change, while high frequencies are removed from the tail end of decaying sounds, which reduces the subjective hiss level. A conventional expander then mutes the signal entirely at very low levels.
Digital noise removal plug-ins (again best used before delay effects are added), produce even fewer side-effects and so may be the best option when recording into a computer-based system. However, you can often achieve a worthwhile improvement using simple low-pass filters — before I moved over to working almost exclusively on the computer, I often used the side-chain high-cut filter in my Drawmer DS201 gate to remove hiss from guitar tracks. The sharper the filter, the less the wanted sound will be affected, so a plug-in with an 18 or even 24dB/octave slope should be even more effective than the 12dB/octave filters the Drawmer uses. The trick when setting them up is to pick the lowest shelving frequency that doesn't materially change the original sound, other than to take the edge off the hiss. Using filters in this way also helps reduce finger noise and squeaking on acoustic guitar parts and can even help disguise moderate clipping distortion caused by recording at too high a level.
Having got the technicalities out of the way, it's time to look at recording methods. The traditional method, and still the most satisfactory in many cases, is to mic up a really good amplifier, but where this isn't appropriate, we have a choice of physical modelling guitar preamplifiers, complete guitar/pickup/amp modelling systems using Roland's VG series of products, or the slightly lower-tech approach of using analogue guitar recording preamps (solid state or valve). The latest option is to plug the guitar directly into the computer and use a software plug-in to handle the amp and speaker modelling, but I'll start at the beginning with the miking options.
When recording an electric guitar, the amp is the instrument as far as the mic is concerned, and mic position is important. While a lot of sound comes direct from the speakers as you'd expect, a significant level is also emitted from the back and sides of the box via panel vibrations. Also, an open-backed cabinet throws about as much sound out of the back of the box as it does out of the front. Choosing a mic for recording electric guitar isn't difficult, as virtually any decent mic of any type can be made to produce usable results. If I were to generalise, I'd say that British recording engineers tend to use cardioid, dynamic models while American engineers seem to prefer capacitor microphones. The dynamic mic produces a solid sound with a smooth high end, while the capacitor mic's increased definition produces a brighter, more open sound when used in the same way. However, the mic position has just as much bearing on the tone as the mic itself.
One way to capture the sound of a large stack, unsurprisingly, is to set up the stack in a large room, play loud and put the mic several feet away so that the combination of all the speakers can be recorded, along with any floor reflections and the subtle filtering effects these may cause. Using this method, the mic 'hears' the guitar sound much as an audience would. Nevertheless, the most common approach is to close-mike the amp with the mic set up very close to the speaker grille and pointed directly at the centre of the speaker cone — where the cabinet has multiple speakers, it's traditional to deliberate over which one sounds best and then mic that one!). Moving the mic to one side produces a less toppy sound, so before reaching for the EQ knob you should try moving the mic.
Where the cabinet is open backed, it's also worth experimenting with miking from the rear, as this produces yet another range of tonal flavours, usually warmer and less bright than miking from the front. It's also quite permissible to mic both the front and rear of the cabinet simultaneously, but experiment with phase inversion on one of the mics to see which setting gives the best subjective sound. Strictly speaking, one of the mics should be inverted with respect to the other, but that doesn't always produce the best result. If you really want to hedge your bets, use an ambience mic several feet from the cabinet and combine this with the close-miked sound, either summed to mono or with the two mics panned left and right. Using a capacitor mic as the distant mic often produces a more believable sense of space, but anything that sounds good goes with guitars.
A practical method endorsed by those engineers who don't like to leave their comfortable chairs too often is to combine the above techniques by using two close mics, one on-axis and one off-axis, plus one distant mic a few feet from the cabinet. If the close mics have very different characteristics, for example a capacitor mic on-axis and a dynamic mic off-axis, you'll get an even greater choice of tonality, as you can vary the mic balance being recorded. Switching the phase of individual mics can often yield interesting combinations and, if you really don't want to leave that chair, you can also delay the ambience to increase its effective distance when it is combined with the other mics. Each millisecond of delay is roughly equivalent to 12 inches of added distance.
With smaller combos, it is worthwhile experimenting with their position within the room, especially when a distance mic is being used. For example, raising the combo further from the ground will result in a different reflected signal path length for ground reflections. Placing a reflective material such as hardboard or linoleum on the floor between the amp and mic will emphasise any coloration this produces. Where a small combo or practice amp lacks bass end, you can try to exploit the boundary effect by placing a mic in the corner of the room, then facing the amplifier into the corner. If the added bass is too much, move the mic and amplifier away from the corner until the tonal balance seems right.
Even apparently crude solutions can produce useful results. For example, I was recording at a friend's flat many years ago and the amp I had only sounded any good when it was played flat out. The answer was to place the speaker cabinet on its back, place the mic right up against the grill, then cover the whole thing with blankets, sleeping bags and anything else that came to hand. It made the level in the room far more tolerable yet still produced the sound I wanted!
Although most of this article deals with miking regular guitar amps or using various DI options, there are also some great sounds to be had using very low-power guitar amps — even battery-powered ones. You can see these in music shops, often designed to look like miniature versions of Marshall, Fender or other top-name amplifiers. While some sound pretty grim, others sound surprisingly musical and raunchy. Obviously they don't have a lot of low end, but if you mic them closely to exploit whatever proximity effect your cardioid mic has to offer, and use EQ cut to tame the inevitable mid-range resonance, you can get some really great sounds. Best of all is that these things are cheap, so you can afford to experiment, and they also come in useful for their intended purpose, which is practising, something we all tend to do too little of when we get caught up in recording technology.
I commented earlier that the guitar amp is the instrument as far as the mic is concerned, but if the guitar is played in the control room with the amp itself in the studio (assuming decent separation between the rooms), a capacitor microphone close to the guitar neck may be used to capture the direct sound from the guitar strings. This sounds a bit thin and naff on its own, but sometimes works well in combination with the miked sound. A similar effect can be achieved by splitting the guitar output, feeding some of it direct to the console via a DI box and then either removing some low end or using an enhancer to exaggerate the brightness. DI'ing the piezo bridge pickups (where fitted) may also produce a similar result, and any of these techniques may also be used with guitar recording preamps as well as when miking.
Some studios are too small for regular amp miking, and even if they're not there are sometimes occasions on which you need better separation from the other instruments playing at the same time. A neat way around this is to use a soundproof box containing a guitar speaker and a microphone. Various commercial models are available where your guitar amp output feeds the speaker inside the box and the microphone feed comes out to the mixer in the normal way. The designer's challenge is to make this work without the resulting cabinet being too large or too boxy sounding. The DIY alternative to this is to place the combo or speaker cabinet in an adjoining room, or maybe even a cupboard or wardrobe (along with the mic, of course!).
If miking isn't practical, or you don't want to find space for a soundproof speaker/mic box, then perhaps one of the available DI options will give you the result you need. I was once reviewing an analogue guitar preamp that provided several programs, both clean and overdriven, and I noted that one of the clean settings sounded really superb. When I checked the manual, I found this was the bypass position! The reason it still sounded good was that the unit was matching the high impedance of the guitar to the medium impedance of the recording system and, for clean sounds, you can get this same effect by using any good-quality active DI box that offers a high-impedance (500kΩ or more) input. A little compression will add density and 'spring' to the sound (experiment with the attack time to get the best 'pluck' sound, as too fast an attack can squeeze all the life out), while reverb will put back the missing sense of space.
The effect of amplifier coloration can be emulated using a parametric EQ, where you'll probably find you need to add some upper mid-range boost to get the same brightness as from an amp. Note that, if you're using a software amp modelling plug-in, you'll still get the best results if you feed your guitar via a high-impedance DI box — plugging it straight into a soundcard's line input is likely to result in a drop in level and may even affect the sustain and high end of the guitar sound due to the pickups being loaded by the impedance of the input circuitry. This does not apply to active pickup systems which, in effect, function as a combination of pickup and DI box.
While a straight DI approach can work for clean sounds, simply plugging a guitar into an overdrive pedal and then routing the output of the pedal into your recording system will sound unnaturally thin and raspy because there are no speakers to take away the unwanted high end, but that hasn't stopped some people using it as a 'trademark' sound! For a more conventional sound, a speaker simulator is needed to filter out the unwanted higher harmonics. A passive speaker simulator, such as the original Palmer Junction box can be placed directly after an overdrive pedal or a regular guitar amp preamp output, but it will need to be recorded using a mic input rather than a line input (passive filtering always results in a reduction in level). Active equivalents (usually battery or phantom powered) are also available, which may offer a greater range of tonalities.
The DI approach most likely to capture the spirit of the original sound is to take a DI feed from the speaker output of a conventional guitar amplifier and feed this into the type of speaker simulator that also includes a dummy load to keep the amplifier happy (valve amplifiers can be damaged by running them into an open circuit). Emulators with a built-in load are generally very much larger than the regular DI box-size emulator-only products. Some of the emulator-only devices can be run from a speaker output, but you still need to plug in either a separate dummy load (usually an 8Ω resistor rated at over 100 Watts on a heat sink) or the original speaker to protect the amplifier from possible damage. Apart from the dummy load, which is passive, the circuitry may either be passive or active.
Some emulator designs include switchable filters, enabling them to simulate open or closed-backed speaker cabinets, and can come very close to the sound of a close-miked amp, while ambience can be simulated using a reverb processor or plug-in. Even if the amp has a good spring reverb, a little additional digital ambience (mainly early reflections) will help create the illusion of the amplifier being recorded in a room.
The original guitar recording preamp was almost certainly the Scholtz Rockman, but within a few years we had several sophisticated competitors (from Sansamp, Groove Tubes and Mesa Boogie) using both solid-state analogue and tube circuitry. These all include speaker emulation of some kind, though usually offer few or no effects. On the whole they are easy to use and some produce excellent results, though they have less tonal flexibility than digital systems designed to model the characteristics of a range of specific commercial amplifier and speaker combinations.
By far the most convenient way to DI guitar is to use a digital preamplifier that combines amp modelling, speaker modelling and effects all in one unit. More flexible still is something like a Roland VG88. This requires the guitar to be fitted with a GK2 compatible hex pickup, but in return enables the system to model pickups and guitars as well as the rest of the signal chain. All these modelling systems sound different to each other and many guitarists say they don't 'feel' the same to play as going via a real amplifier, but, in many situations, the sound you record using them is very much like the sound you hear on records. As you hear exactly the sound you're recording over the monitors while you're playing, you know just how the end result will turn out, so there's little to be said about working this way other than to rely on your ears. Nevertheless, it is worth pointing out that some of the more unlikely combinations of amplifier models and speaker cabinet models can produce unexpectedly interesting results, so don't just rely on the factory patches, most of which I've found not to do justice to the products anyway. The morphing feature used in the Digitech Genesis 3 processor is also very creative, as it allows you to combine the characteristics of two completely different amp/speaker models, sometimes with surprising results.
Guitar players who haven't done much recording tend to comment that the sound coming back over the studio monitors isn't the same as what they hear when standing in front of their amp at a gig. This is hardly surprising, since studios seldom monitor at that kind of level, so the question they should be asking is whether the sound you hear over the monitors is comparable with the guitar sounds heard on similar records.
Adding effects at the mixing stage gives the engineer greater creative flexibility, but if the guitarist needs to hear the effects to play, then you may get a better artistic performance by recording them with the take. All I'd say on this point is that editing is much more difficult if the sound is recorded with delay or reverb, so an alternative is not to record these effects initially, but still add them to the monitor mix for the player's benefit during performance. Effects like chorus and wah-wah can be recorded straight off, if required, as they don't affect the ease with which a part can be edited. Ultimately, the performance is what really counts, so compromise in favour of the player's artistic needs rather than your technical needs where a choice has to be made.
If you really like to cover all options, record using any of the above methods but also take a straight DI feed with no effects and record that onto a separate track so that you can process it later. Some engineers have been known to use a recorded DI guitar track to drive a guitar amplifier, which is then miked up and re-recorded, but you could take the easier route of using a hardware recording preamp or a guitar amp emulator plug-in to process the track.
There is no 'right' electric guitar sound because the instrument has no natural sound, so you can use as much EQ as you like without feeling guilty. Even so, getting something close to the desired sound at source is always the best way to work. Should you need to use EQ, here are a few comments concerning the frequency areas you might want to tweak. Boxiness tends to occur between 100Hz and 250Hz, so if this is a problem, use a parametric to tame it. Experiment with the Q setting as you may be able to notch out a narrower section than you think to get the desired result. Cabinet thump can be accentuated by boosting at around 80-100Hz, but take care not to boost anything much lower than this, as you'll just bring up unwanted resonances and hum.
Add bite or presence by boosting between 2kHz and 6kHz, depending on the tone you're after. Little over 4-5kHz is produced by a guitar speaker, though going for a brighter DI'd clean sound is quite legitimate for artistic reasons. Similar-sounding electric guitars that may be conflicting within a mix can be separated to a limited extent by adding bite at different frequencies, though choosing two different-sounding guitars and/or amp sounds and examining the arrangement carefully usually works better. As a rule, single-coil guitars are best for cutting through a busy mix without taking up too much space, while humbucking pickups create a thicker sound which may be beneficial in recordings where there is only one guitar part.
You can use compression not only to keep levels under control but also to increase sustain. Sometimes a guitar part will work better in a track if the sustain is created using a little less overdrive augmented by compression rather than relying on overdrive alone. Even a heavily overdriven guitar sound can be made to appear more powerful if compressed, as its average energy level is increased further, though any noise will also be rendered more obvious. Clean guitar parts can also benefit enormously from compression, as it creates a more dense, even sound and can emphasise the picking attack of the notes when a longer attack time is set. Using a faster release time, so as to cause audible level pumping, can make for exciting rock guitar sounds, but some compressors pump more gracefully than others — Dbx models are quite popular in this respect.
To create a fairly live, in-your-face sound, a short reverb or ambience program with a fairly bright character is ideal. A hint of slapback echo or a little pre-delay before the first reflection can enhance the sense of power and intimacy, as when a guitar is played in a small club. More generous reverb can be combined with repeat echo effects to create a lush, spacious sound, but you need to leave space in the instrumentation and arrangements for these sounds to work, otherwise they just sound confused and messy. Feeding the effects send through a chorus or flanger before the reverb gives a nicely complex twist to the sound without being too obvious.
A final tip: not a lot of people know this, but Antares Auto-Tune works rather well with electric guitar. Of course Auto-Tune is monophonic, so it will dutifully ignore chords and double notes, but it will come into play when it detects a single note. You might reasonably ask whether a guitar tuner might not be a wise investment, but the reason I suggest Auto-Tune is not to compensate for a badly tuned guitar, but rather to pull imprecisely bent guitar stings into pitch.
You can do this using a fairly slow tracking time in Auto-Tune so that the bend dynamics aren't changed in any obvious way — it's just that when you finish bending, the note will come to rest on a precise value. Not that I'm suggesting you need to do this, of course, but the day will come when a client plays a never to be repeated take that is perfect apart from a few bend intonation problems...