Last month, we looked at which elements are most likely to go wrong in a mix. This month, Paul White takes a more positive line, and looks at how to record the right instrumental sound to tape in the first place.
At the risk of being boring, there's just no escaping the fact that a good mix starts by getting good sounds onto tape (or disk) in the first place. 'Fixing it in the mix' really is an uphill struggle, and rarely entirely successful. In fact, only a very sad person who has spent far too many hours in the studio would consider this kind of salvage operation to be either fun or artistically satisfactory. Having said that, next month's article is all about rescuing duff mixes!
First Bass
I'd like to start with the bass guitar, because on the face of it, all you need is a DI box and a compressor, and you're all set to get a great bass sound. If you're lucky enough to work with a good bass player, this may well be true, and I've done a good many sessions where the bass has been DI'd via an active DI box with a compressor patched into the channel insert point, just to keep the levels under control. You can get absolutely superb results this way that need virtually no further treatment or EQ, but pass the bass to a different player, and you may end up with a sound that's all squeaks and fumbles, with no power or depth. Yet this same player may sound fine live.
Once you've coached and cajoled the player to play the damned thing properly, it's up to you to make the best of what you have, and I've discovered a few tricks that might help. Firstly, if the player is really uneven, don't be afraid to pile on the compression to get the levels under control, and if the bass part is fairly simple, you might benefit from patching in a gate (usually before the compressor), to help keep the spaces clean.
Unless the player has a confident style, you might still have too much squeak or fret noise, especially if you've laid on the compression with a trowel, and under these circumstances it can help to use a DI preamp with a speaker simulator built in. The Award Matchbox, or a conventional DI box followed by a Hughes and Kettner Red Box speaker simulator works well. While we always appreciate the need for a speaker simulator when DI'ing guitar, it's often overlooked that a real life bass amp is usually loaded with 12‑ or 15‑inch speakers, which have a limited frequency range. Even a bass cab with 10‑inch speakers will sound warmer and less edgy than the DI'd equivalent, so taking steps to simulate the speaker coloration can help thicken the sound while making it seem cleaner. In some instances, you can't do better than to stick a mic in front of a bass amplifier — so don't rule this out.
While I seldom advocate the use of radical amounts of EQ, bass guitar is one area where you can really go to town, because electric guitars produce quite artificial sounds anyway. This means that you don't have to worry about making them sound natural. However, it's worse than useless working on a bass guitar sound in isolation, because it will sound totally different when you add the rest of the band. I've fallen for this one before, and what seems like a wonderfully rich bass sound when solo'd can easily dissolve into a glutinous mess when you bring up the rest of the mix. The only way to do it is to set up a nominal mix balance, and then EQ the bass in context, so that it cuts through properly. You'll find that the mid‑range is very important here, and you may have to apply quite a lot of boost somewhere between 2kHz and 4kHz to get the right edge to the sound. If you boost much higher up, you're into fret noise and string buzz territory, and if you've recorded without a speaker simulator, you might want to roll off some top at the same time.
At the bass end, you can add boost at between 70 and 120Hz to fatten up the sound, and a good sweep equaliser or, better still, a parametric, helps a lot. A simple low‑shelving EQ will emphasise the bass, but may leave you with a poorly defined, out‑of‑control bass end. Though most mixing consoles these days have pretty good EQ sections, a quality parametric equaliser is a lot more flexible, and every serious studio could use at least one.
Guitars
Like the bass, an electric guitar sound starts out with good playing technique, and you'll usually find that the better the player, the simpler the recording technique you can get away with. Clean rhythm parts need nothing more than an active DI box (and not even that, if you have an active guitar), with the possible addition of a compressor. Alternatively, nearly all the reputable guitar multi‑effects preamps produce wonderfully 'finished'‑sounding, clean tones, with or without effects — so guitar‑playing studio owners should look very seriously at these (check out our Effects Buyer's Guide starting on page 40 of this issue for models and prices).
Fixing it in the mix really is an uphill struggle and rarely entirely successful.
Lead guitar sound is a much more personal thing, but you can still get a good sound by DI'ing. In my experience, most of the all‑in‑one recording preamp/effects units give you a great range of heavy metal sounds, but all the ones I've tried fail to give you a natural, touch‑responsive amp sound — and that's what most guitarists seem to want. A good test is to set up an overdrive sound, and then see if you can bring it down to an almost clean sound, using just the volume control on the guitar. If the sound disintegrates into a fizzy mess, you know you're onto a loser.
If I must DI, my own preference is to use either an Award Sessionmaster recording preamp or a Sansamp unit. Both of these pass the above test, but both also have different characters, so you'll have to try them yourself to see what you want. On a recent demo, I had to recreate the Dave Gilmour 'Brick in the Wall' sound, the vintage Eric Clapton 'Hideaway' tone, the classic 'Money For Nothing' honk, and a generic raunchy 70s rock tone. Interestingly, using a Fender Strat Plus in all cases, I managed to come very close to the original sounds using a Sansamp XXL pedal plugged into the desk via an Award Matchbox, with a little compression (around 8dB of gain reduction), courtesy of a Drawmer 241 set on Auto. The main effect I used was a little ambient reverb, with the addition of a little subtle delay on the Floyd sound. The desk EQ was used to squeeze up the mid‑range on the 'Money for Nothing' sound, to produce the necessary nasal quality. The neck/middle pickup combination was used for the Floyd sound, and the bridge/middle combination used for all the others.
Interestingly, the most aggressive lead sounds come through turning the distortion amount down, not up. Use too much overdrive, and you'll end up with a wall of fizz that loses almost the whole of the character of the instrument you're playing. Again, I must emphasise that a lot of the sound starts with the player, and if you can't go some way towards emulating the technique of the artist whose sound you're trying to capture, you'll never pull it off. In particular, picking intensity and vibrato are very important, as is string‑bending.
Because of the physically aggressive nature of rock guitar playing, it is also vital to check the tuning before every take; instruments with tremolo systems are generally more susceptible to tuning problems than fixed bridge models. The temperature changes in studios also affect guitars and basses, and I find that my guitar tends to go slightly sharp after a while, even though it's been left to acclimatise to the room. Fitting locking machine heads such as those made by Sperzel or Schaller can make a big difference to tuning stability, and this is a simple DIY job. That way, there's no excuse for not renewing your strings the night before a recording session, on the grounds that the tuning might drift!
A convenient alternative to DI'ing an electric guitar is to use a combined speaker simulator and dummy load, of the sort that replaces the loudspeaker in a combo or guitar stack. This approach works particularly well with nice‑sounding valve amps, and avoids the noise problem very neatly. I use one of the old passive Palmer models, which turns in a very close approximation of what the amp would sound like miked up, though you have to add ambient reverb to recreate the room sound, just as you do with any other type of DI or very close mic recording.
Of course, all the previously discussed methods are designed to mimic the sound of a miked‑up amplifier, so if noise isn't a problem, why not mic up an amplifier? If you like to work and play in the control room, you can put the amplifier in the studio and run a lead through to it from the control room. That way, you get to hear the sound over the monitors, which tells you what the recorded sound is going to be like. It also helps prevent you from going deaf! Moving the mic a few inches to either side, or back from the speaker makes quite a difference to the sound, so if you're not happy with your first experiment, take a little time trying the mic in different positions.
If you mic very close, you'll still have to add reverb, though you may prefer the classic spring reverb in the amp itself, if you're a stickler for authentic sounds. It's also worth noting that the size of the sound on tape has little to do with the physical size or power of your setup, and many classic rock albums have been made using little Fender Champ 12‑watt valve amps or similar combos.
Drums
We've covered drum miking pretty extensively before (See SOS April '94, and September '95), so I won't dwell on the details too much, but once again, the sound comes from the player and the kit, not from the mics or the reverb unit. Good mics will make a better job than poor ones, but the way the kit is played still makes the most difference.
The drum machine is a useful alternative to a human drummer for pop applications, but only if it is programmed extremely well. A compromise approach is to use the drum sounds from a sampler or drum machine, but trigger them from drum pads such as the Trap Kat or Roland Octopad. On a recent project, the drummer used my DrumKat plus a bass drum pedal to put together a demo. We used a programmed drum part as a guide, so that we could work on our parts individually. This is more friendly than a click track, but for music with a freer feel, it's best to turn off the click altogether and just record the sequenced parts 'free'. This ignores the sequencer's own bar and beat reference, so you can't quantise the result, but it's the only way to work for music with real feel.
In our case, we settled for the convenience of playing to a quantised guide track, but once this was erased, it surprised me just how much real feel had been injected by the drummer. You can see this by looking at the piano roll edit screen, where it's quite obvious that the beats are consistently falling before or after the quantise tick that you expect them to fall on. Indeed, if you do quantise the parts, the feel may end up totally wrong.
If you record something dry, you can change the effects afterwards. If you record with effects, you're stuck with them.
If the playing is only a little sloppy, however, you can use quantisation, dialling in a percentage to bring the beats closer to their 'perfect' positions — and without losing all the feel. Between 50 and 70% is usually OK, but beware of quantising fills, because this can trash them altogether. In practice, it may be best to break the drum track down into sections, and then quantise only the parts that contain straight beats. If the fills sound natural, leave them unquantised. If some of the drum sounds have a flanging effect on them, you probably have a duplicated note caused by stick bounce, and if you've used quantisation, these notes are probably exactly on top of each other.
Some sequencers have a function for eliminating duplicated notes, but in the case of my own system, this function has a habit of removing the loud note and leaving the low‑velocity stick bounce note underneath. If this happens, you may have to go through the track manually, adjusting the velocity levels of rogue notes to make them match the rest of the track. Another method might be to use the 'delete notes below a certain velocity' function, to trim away unwanted low‑level stick bounces.
Good as sampled cymbals are, they are still too regular for my taste, so to add the real finishing touch to an electronically recorded drum part, use a real hi‑hat and real cymbals along with the drum pads. If you put the mics reasonably close, the thudding of sticks on rubber shouldn't pick up too seriously.
Keyboards
Keyboards are the easiest thing in the world to record, because the sounds are pre‑packaged, pre‑designed and pre‑digested. What's more important is choosing the right sounds, and if the instrument has built‑in effects, ensuring that the effect level isn't set too high, as you won't be able to take it off again afterwards. Most people now run their keyboards into the mix as virtual tracks, but there's a lot to be said for putting them onto multitrack as well, in case you want to do a remix without having to recreate your original MIDI setup. With analogue tape, you need to take care that high‑pitched synth sounds don't distort. Record them at a lower level if they give problems, but otherwise, record them flat, so you can EQ them later if there's a need.
On a general note, if you can avoid using gates during recording, so much the better, because you only get one chance to get the settings right. If you can arrange it so that you can gate during mixing, you'll be able to experiment with the settings for as long as you like, and there's the added bonus that the gates will take out any tape hiss as well. Likewise, unless you have a good artistic reason for doing so, avoid recording too many effects; if you record something dry, you can change the effects afterwards. If you record with effects, you're stuck with them. In the case of mono instruments with added stereo effects, it also saves tracks if you can record them dry in mono, and then add stereo effects during the mix.
On the other hand, some keyboard sounds rely heavily on the internal effects to make them work, so it would be wrong to switch these off without good reason. You can still keep the effect in stereo by running the instrument into the mix via a sync'ed‑up sequencer. Guitar players also need certain effects to be able to perform properly. Overdrive, wah wah and chorus are examples of commonly used guitar effects, but if you have a spare track, it might pay to make a duplicate DI recording straight from the guitar at the same time. That way, if the sound on your main track doesn't work out, you can take the clean DI'd sound from your extra track, and feed it back through an amp or effects unit, to create a new guitar sound without having to get the part played again.
Vox Pop
There's nothing difficult about getting a good vocal sound — so long as you have a good singer — and we've covered the mechanics of miking singers before, most recently in SOS January '95.
You must have noticed that you can just stick a mic in front of some people and whatever they do sounds great, while other, apparently musically proficient people can spend hours trying to get a sound. This just goes to show that most of what we consider to be a good sound comes from the voices and fingers of the musicians, not from a rack of esoteric equipment.
The main thing is to use the best mic you can get your hands on, use it with a separate pop shield and avoid having any reflective surfaces too close to the singer, especially directly behind them, as these will bounce sound back into the microphone. Record without EQ, and provide reverb in the cans to make the singer feel good.
Also, it's important to appear patient and supportive — artists don't work well if they feel hassled. You may need to add a little compression during recording to even out the level, but don't overdo it, as you can always add more when you mix.