You may be a whizz with the little black boxes, but how are you at recording real sounds and real performances? Wilf Smarties serves up a selection of tasty studio tricks.
You may be a whizz with the little black boxes, but how are you at recording real sounds and real performances? Wilf Smarties serves up a selection of tasty studio tricks to spice up your production.
Many of you will have had some recording experience, even if this goes no further than using a Portastudio. However, as sequencers provide an ever increasing proportion of the music we listen to, a new breed of engineers is emerging who are more familiar with the properties of black boxes than those of microphones. In this article I will examine some of the tricks of a vanishing trade that might just give your production an edge.
Nowadays it is all too easy to record and produce music from a single workstation. Like Fat Freddy's Cat, who was convinced that fish came from tins, it must seem to many programmers that sound originates from somewhere under the bonnet of an M1 or similar. I would liken composing on one of these expanders or workstations to building a house out of Lego. There are an infinite number of ways to arrange the bricks, but a limited number of significantly different results. But never mind; by the time you have run out of ideas with one machine, another will be on the market.
The sampler offers hope: now you can choose the noises. But where do these sounds come from? Not from under the bonnet any more, but from floppy disks? Actually, there's a very noisy world out there, and it's all up for grabs.
Let us first consider the microphone. This is a device for transforming sound energy into electrical energy. The reverse, in fact, of the role required of the loudspeaker. Most high quality microphones are of the electrostatic type, which require DC power from a 'phantom' or battery power supply. Dynamic types can also perform well, but cannot quite match the bandwidth or transient response of the best of the former, due to the heavier mass of their moving coils. A large sensitivity range, and a wide and flat frequency response are nonetheless features of all good microphones. The most profound difference between microphones lies in their directional, or 'polar', response characteristics. Some, like the Neumann U87, can offer more than one. Others are highly specific, like the ultra-hypercardioid Sennheiser MKH816, which can pick out a single violin from an entire orchestra. Another consideration when choosing a microphone for a particular application is physical robustness. The Electrovoice RE20 is a pretty good dynamic mic with a respectable low end response, but what makes it special is its unerring ability to come out of a bass drum after a heavy recording session in one piece (it's also popular with saxophonists.)
So choosing the right microphone is important but more so is positioning it correctly. Consider this property of sound: its intensity falls off as a square of the distance from the source. A microphone positioned two feet away from a performer will pick up 16 times the relative room ambience of one at a distance of six inches. Digital reverbs can readily add ambience; no black box can remove it (yet).
Live recordings these days concentrate on the human voice, the most complex of all instruments. Here natural ambience is seldom desirable. Even a small amount of 'room' can seriously degrade the intelligibility of the program material. There are ways to mitigate against unwanted ambience at the recording stage. Not all are obvious.
The first thing you could try is getting close. As I said above, the closer a performer gets to the mic, the 'drier' will be the signal. Unfortunately, at very small distances between source and mic a phenomenon known as 'proximity effect' comes into play, resulting in bass enhancement, and some corrective equalisation might be necessary. Small movements on the part of the vocalist, hard to avoid if they are getting into a performance, can result in wild frequency response fluctuations, particularly when using unidirectional microphones. Percussive consonants, notably 'P', produce pronounced shock waves which can cause the mic diaphragm to overreach itself, resulting in 'popping'. Pop shields can help to reduce this effect, but I've yet to find one which doesn't knock out some of the high-end response. A more satisfactory result is obtained by putting a windbreak (traditionally constructed using a coathanger and a pair of tights, though some spoilsport manufacturer now provides these screens ready made), 'twixt phizzog and U87.
Both proximity effect and popping are much less severe if the mic is positioned looking down towards the mouth at an angle of 30 degrees, rather than straight on. The softer the voice, the closer the mic can get. For that Christians or Prince ultra-present breathy vocal sound you need a controlled dynamic vocal performance, a close mic position, and a dead acoustic environment.
A dead acoustic environment is an avenue worth pursuing in itself — obviously the more anechoic the room the better. I like the sort that make you think you have lost your voice until you put on the cans. Creepy. With this degree of acoustic control around, feel free to use an omnidirectional microphone. These give the performer more scope for movement, and in any case sound more natural.
In situations where the above is not possible — and not many of us can create an anechoic chamber at home — other tricks may be tried. Some vocalists perform best belting it out in front of a large pair of studio monitors playing back at a distinctly immodest volume. Not much chance of keeping the music off the mic here, you might think. Wrong. Record a take of singing/rapping/snoring or whatever. Nobody move. Now record a take out of phase by 180%, with no singing. Play the two takes back together and hey presto! The spill from the monitors magically vanishes. (If your console doesn't have phase reverse, you can easily make up a reversing mic lead — see Star Wires last month.)
You can do a similar trick to the above using two distant identical directional microphones, one pointing straight at the source, the other looking along a parallel path. Something like a 4:1 ratio for the mic to source distance versus the spacing between the two mics might be applied. The phase of the second microphone should be reversed, either by using a phase switch on the mixer, or a phase-reversing cable. The difference between the two signals reveals a much higher source to background noise ratio than could have been obtained by using a single direct microphone. You can play around with the in and out of phase mix and mic positioning in real time to optimise the room-cancelling effect.
Some FX processing can appear to dry up signals. The more observant among you will have noticed the overt use of dynamic expansion in OB news bulletins and interviews of late. Because background noise is suppressed between bits of dialogue, speech intelligibility is enhanced, although the accompanying noise pumping can be irritating. Gates are commonly used to keep inactive mic channels quiet, eg. on a drum kit, backing vocals etc. When applied to vocals (I recommend after, rather than during, the recording process) the unnaturally silent gaps, devoid of such signs of life as breathing etc., can throw the program material into high relief.
The importance of the mid-range of the frequency spectrum to speech intelligibility is well understood. Treble emphasis can also help project clarity. Choose a frequency well into double figure kHz to get above the sibilance zone, or use a phase-smearing harmonic-generating device like an Aphex Aural Exciter. A 'de-esser' (frequency selective compressor) might help to keep sibilance under control, but unless you have access to a BSS dynamics processor or something equally wonderful, you might find it tends to introduce 'lithips'. In this case I would apply overall compression, with fastish attack and release settings, to an already treble emphasised signal. HF overshoots can usually be reigned in fairly reliably this way. In addition I would probably use a high-pass filter (at say 100Hz, 12dB/octave) to avoid low frequency high energy transients such as pops suddenly 'ducking' the signal as they hit the compressor. (Some microphones have high-pass filters built in.) Lastly, some treble deemphasis might be applied to bring the frequency content of the processed signal back into line with reality. Incidentally, you can get a great treble-emphasised sibilance-controlled result by using Dolby A in record, and an expander/gate instead of Dolby decoding on replay.
"For that Christians or Prince ultra-present breathy vocal sound you need a controlled dynamic vocal performance, a close mic position, and a dead acoustic environment."
Now that you have successfully removed all traces of the real world from your vocal take, it's time to put some acoustic space back in.
Repeat echoes (typically between 150 and 300ms.) are often preferred over reverb, as the former can give a profound impression of space without destroying intelligibility. Feed these echoes into an early reflection or short decay room program of your reverb unit to give a more musical, 'diffuse', effect.
Close single or multiple delays (from 1 to around 80ms), sometimes with modulation, can be used to thicken and/or harden (or soften) the texture of a signal. Un-modulated close clusters give hard static flange effects ('I've Got the Power' by Snap). Longer delay clusters are used for 'multitracking' a la Kylie. Modulation softens; regeneration further hardens.
Where heavy reverberation (say a 4 second room) is required, try instead using a 1.5 second room, with a pre-delay of 2.5 seconds. Separating the dry signal from the ambience by such a large amount vastly improves clarity.
Using compression on a signal which has reverb or echo mixed into it is interesting and useful: whenever there is a gap in the program material the reverb or echo rushes in; when the program material restarts it immediately recedes. This can give a combination of signal clarity together with the illusion of constant heavy reverb/echo, and may be used in place of, or in conjunction with, the previous trick.
So that about wraps it up for the human voice. How about more complex set-ups? Recording string ensembles is a minority sport, but most studios will have to deal with drum kits on a regular basis. Looking at how a kit is miked up onstage gives a fair idea of the basics of multi-miking drums. Most principal drums have their own mic, as does the hi-hat. Cymbals can be miked individually, but a crossed stereo pair, 120 degrees apart, mounted centrally over the kit looking down, should cover them. In fact, a well positioned stereo pair can give a pleasing result on their own. For my money, if you are looking for a great kit sound, you should start by getting this right, then fill in where required with close mics, which brings us to another property of sound: its finite speed of 330 metres per second at sea level.
You may have heard (or heard of) 'time aligned' loudspeakers, where the distances between the listener and the various drive units in a loudspeaker enclosure are designed to be identical. This represents a correction of maybe a few inches over a normal flat-fronted enclosure. It stands to reason that in the case of a multi-miked drum kit, with distances of several feet between snare and overhead mic pair, there must be a huge potential for sharpening up the image by applying progressive delays to mics positioned further away from the stereo pair. I look forward to trying this out soon in a novel setup which I may describe in a separate article.
John Bonham made use of huge quantities of natural ambience to create a distinctive drum sound. Nowadays digital reverbs come prepared with complex algorithms that come close to sounding like a real room. But not close enough for some people. Instead of searching about in your Lexicon 480L (you mean you don't have one?) for the perfect room, why not use a real one instead? (Incidentally, this trick need not be confined to drums!) Take a kit which has been recorded dry, and feed the result through a pair of hi-fi speakers into a live room. You can then mic up the ambience pretty much as you would for a kit being played in that room.
Noise gates are the natural first choice of uninspired studio engineers when faced with the problems of separation between drums, but when armed these can be extremely dangerous, causing less than-full blooded tom hits, for example, to become lost forever. They are best applied post-recording, if at all. A little-used tom group can then be bounced into a stereo pair, taking care not to miss any hits, or let any false triggers (from whacked snares, etc.) getting through.
Separation between individual mics on a kit can be a nightmare, and attempts to achieve this have frequently led to serious over-gating. There are, however, some simple methods for achieving better separation at source. Obviously, the closer the mic to the source drum, the better the signal to background noise ratio will be. You can improve this by the judicious placement of small, acoustically absorbent screens, for example between the line of sight of the snare mic and an adjacent tom or hi-hat. In order to minimise sympathetic resonances between them, no two drums should be tuned to the same note. Damping toms is not generally recommended, but if you must, use a large, loosely taped piece of foam, rather than the internal dampers within the drum. The latter tend to stress the skin and give rise to unwanted resonances.
Tightening the snares can reduce sympathetic sizzling, but also tends to make the snare sound more like a timbale. However, those missing snares can be put back by the following method. Place a snare drum with nicely loosened snares on top of a loudspeaker. Now fire the signal from the recorded snare track through it, and mic up the result, close or distant, according to taste. The rattle of the snares thus activated will follow the dynamic of the original performance, and should sound perfectly natural. This can be done during or post-recording.
"The only time I ever heard a combo amp realistically miked up, a PZM-type microphone was lying on the floor about two feet in front of a Fender Twin Reverb."
It is common practice for drummers to remove the bottom tom-tom rim and skin: similarly for the front kick drum. This more often than not means that the now redundant hardware (retaining lugs) on the shell of the drum(s) in question is prone to rattling. Plasticine or Blu-Tack forced into the lugs solves this particular problem. And talking of hardware: kits are prone to physical movement (thence the anvil in the bass drum). Care should be taken to prevent hard surfaces (eg. cymbal, mic stand or drum shell) coming into contact with one another during a performance. Improvised foam 'spacers', Sellotaped strategically at potential points of contact, should eliminate this risk. (Reads a bit like the Viz readers letters page, doesn't it?)
Hearing mic channels opening and closing as noise gates do their dirty work around a drum kit can be very annoying, but if digital reverb is to be applied to snare, kick and/or toms, an ungated signal generally produces a messy result. The solution here is to apply the gate to the effect send only, rather than the complete signal. Voices to be treated with this degree of respect will usually require two desk channels each, the second (gated) channel being sent to the reverb unit while being removed from the stereo bus.
By which I mean electronic instruments. We will get on to spicing up DId Roland D50s later, but we will start with rock'n'roll's favourite son, the electric guitar. Miking up a guitar amp surely needs no justification. Yes, well... how many times have you heard (or made) the cry "but it doesn't sound like that next door", referring to the fact that the guitar sound in front of the amp is just wonderful, whilst out of the studio monitors comes a dull and unpleasant grinding noise. How can such a large and powerful guitar stack sound so tiny? Consider again the microphone, invariably pointing at one loudspeaker. What the studio monitors will deliver here will be the sound of one loudspeaker, not that of the stack of combo as a whole. The only time I ever heard a combo amp realistically miked up, a PZM-type microphone was lying on the floor about two feet in front of a Fender Twin Reverb.
Of course, in some cases it's the sound of the loudspeaker you are looking for, so always make sure that you have at least one decent speaker in your 4x12. If you do fit a single high quality driver, make sure it matches the others with respect to impedance and phase.
Here's a strange fact: the smaller the guitar amplifier, the bigger it sounds through a microphone. I have no sure answer as to why this should be, but I suspect it has something to do with matching the scale of the source to that of the microphone, both physically (size), and acoustically (sensitivity). Many a heavy metal album has been recorded satisfactorily using tiny Pignose practice amps.
Putting a DId keyboard sound (eg. organ or electric piano, or even synth) through an amplifier, either during or post-performance, can add depth and 'realism'. A favourite of mine here is a Roland JC120 miked up with a pair of Shure SM58s. (The JC120 has a stereo chorus/tremolo which pans across its two loudspeakers in a reasonable simulation of a Leslie.)
Vocals, too, can benefit from being fed through an instrument amplifier, then re-miked. You can then subsequently mix the dry and treated signals.
Don't forget that you can apply the 're-mixing' techniques described above post-record, via an aux send, as well as live. And while we are on the subject of aux sends, it pays to be imaginative with what you feed them into. Has anyone tried wiring one up to an electronic tuner?
I was once faced with the problem of recording a guitarist whose playing technique was next to non-existent. Asked why he was in the band, I was duly informed that he had the right "attitude". For once I had to admit defeat: I could not record his attitude, only his performance.
Most musicians and singers, however, are capable of turning in a usable performance. As the recording engineer your job is not just to make sure that input gains are set optimally and that the mic is pointing in the right direction, but also to try to get the best out of the artist. If the artist feels uneasy, uninspired, or inhibited, you all might as well retire to the bar over the road. So how do you get around the nervous singer?
It helps if you are totally confident in your abilities as a recording engineer, and can convince the artist that all they have to worry about is their end of the business. Frankness and honesty about your own and your artist's abilities is a must. A fair amount of good natured abuse, oddly enough, seems to help people relax in the studio, probably because it reflects real life. Be prepared to take it back.
"Doubling (or tripling, or whatever) a voice, by recording two takes on parallel tracks, is a good way of thickening up a modest vocal."
To look specifically at recording vocalists, there are basically three ways to achieve a final vocal in a multitrack environment: firstly, the singer can do it in one take; secondly, the engineer can drop the singer in whenever a mistake or below par note is struck; or thirdly the engineer can record successive takes on successive tracks, and then bounce these down onto one, taking only the best bits from each. The first needs no further comment: you are probably by now in the bar across the road celebrating a probably unique event. The second is an acceptable way to build up a vocal take if recording tracks are scarce.
The third is the professionals' method. This method has several advantages over the second: the singer can get into his/her stride rather than being continually pulled up to repeat a line; there is no chance of a dodgy drop-in; and the engineer can bounce down the composite track away from the heat of the moment.
• Tip 1: Some of the best vocal takes are the very first. Don't miss them 'getting a level' — you have been warned.
• Tip 2: Honesty is not always the best policy: some artists have a recording phobia which causes them to fret whenever that red light is on. Tell them it's just a run through, and go for it.
• Tip 3: The higher the tape speed, the better the quality of the recorded signal. Also, very importantly when dropping in, the less time it takes for the tape to travel between the erase and record head. At high speed (30ips) you can pretty much punch in on the beat or syllable you want. If the tape speed is any slower you will have to be quicker!
• Tip 4: When bouncing four vocal tracks down to one, make up a score sheet referenced to the lyrics, and audition each track ticking off all the good lines. Then do your bounce. If it doesn't work, try again. Using mute switching to select between tracks during a bounce-down is usually OK, but for a softer merge try punching into record at the edit point.
Recording instrumentalists requires an approach broadly similar to that for vocalists, though you can probably get away with two tracks instead of the usual four, and dropping in is usually acceptable.
Doubling (or tripling, or whatever) a voice, by recording two takes on parallel tracks, is a good way of thickening up a modest vocal. The chorus effect this generates can be readily simulated by a harmoniser, but the resultant signal will still track the pitch and timing errors of the original. By repeating the performance these errors are (hopefully) randomised. Often when doubling a line I will use two tracks, and keep switching takes between them, alternating the record track, until the vocalist settles down and starts to reproduce nuances and so on in exactly the same way each time. Don't expect take 2 to track take 1 exactly.
Doubling is especially effective on backing vocals and harmonies. Work up your chorus arrangement to perfection, then sample the blighter. (You don't want to go through all that again, for every chorus, do you?) 'Flying in' (especially with backing vocals) has been standard practice in recording studios for decades, except nowadays you don't have to keep two tape machines in sync to do it.
Some performers play or sing consistently ahead of or behind the beat. Retarding an early tape track is easy (put it through a delay line). But what about the bassist who perennially lags behind? If the tracks on your tape machine can be individually monitored from either the sync or play head, you could bring that bass player bang up to date by monitoring the bulk of the music from the sync heads, and he or she of the lazy plectrum from play. Applying a robust delay, corresponding to the distance between the heads divided by the tape speed, should bring the bass right back to its original timing, and a further minor downward adjustment of the delay time will bring the miscreant to heel.
Unlike my previous studio articles, Star Wires and Sound Spaces, where I could present the subject matter in a logical and evolutionary way, Radical Recording is somewhat fragmented, being a collection of tricks gleaned piecemeal during years of open warfare with the customers in my erstwhile studio. Many of the best came to light as a result of solving intractable problems, or from gratuitous experimenting, 'pushing back the envelope' as they say at Mach 1. In order to gain maximum freedom of expression in a recording situation, it pays to start by being well organised. Struggling to find a stubborn fault or a missing sound file in the middle of a session does nothing for musical creativity. So name those sequencer tracks, backup those disks, fix that dodgy lead, and take chances with your music instead.
Feature by Wilf Smarties
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!