Recording Techniques (Part 15)
The Mix (1)
Part 15. David Mellor looks at how to create the perfect mix.
Is it true that you can 'Fix it in the mix'? David Mellor explores the most mysterious part of the recording process.
The phrase 'We'll fix it in the mix' has hopefully become part of recording history, now that everyone has well and truly learned that whatever is wrong on the multitrack is going to be wrong on the final stereo mixdown. And if there is a mistake to be covered up, it's going to take 10 times as long to fix it at the mixing stage than it would to correct during track laying.
Mixing a multitrack tape into stereo is definitely an art. There is very little 'technique' to it because every move you make is governed by what you hear coming out of the monitor speakers, and the rule is 'if it sounds good — do it'. Having said that, I should add that you really need to be sure that what sounds good to you is also going to sound good when played back from record or CD by all the potential buyers of your music, or when heard by listeners on their car stereos. In any case, the most important thing during any mixing session is to get the message of the music across to the listener. Whatever form that message takes, from 'eat your greens and save the planet' to simply 'get up and dance', this is the point where it could all come together or all fall apart.
The first thing you need for a successful mixing session is a well recorded multitrack tape. If this has been prepared properly, then mixing will be a smooth, painless process, but if there are problems to be ironed out, you'll spend much of your energy coping with these, leaving less to devote to creating a good mix. The worst situation you could find yourself in is being asked to mix a tape which has been badly recorded by someone else. If you are lucky enough to have some kind of automation to help you out then things might not be so bad, but if you have to fix problems and create a good mix at the same time by pure finger power, then jolly good luck to you. I've been there myself and I don't like it.
The one thing I can't do in an article like this is tell you exactly how to mix, because the only way to learn is by experience. However, I can pass on a few ideas which will help a budding recording engineer get a better sound, and get it more quickly. For the purposes of this article, I'm going to assume that there is no automation on the console, and everything is done by hand. Automation isn't by any means common, apart from at the high end of the business where it is almost ubiquitous. One day we'll all have it, but for now we need to have our fingers at the ready. Let's roll the tape...
There's no such thing as the perfect multitrack recording, all ready and waiting to be mixed down with no EQ, level changes, effects or jiggery-pokery of various kinds. But it would be nice if we could direct our efforts towards making the raw material of the mix as easy as possible to deal with. Let's see what the requirements are:
If the recording is of a band using 'real' instruments, then there is probably a lot of extraneous noise on the tape that you could really have done without. As I said earlier, the time to deal with problems is during the session. If, for instance, the guitarist hits a string accidentally at a time when he is not supposed to be playing, don't leave the noise in — erase it. The time to erase it is while the player is still strapped into his or her instrument and plugged into an amp, so if by some foul stroke of luck you erase something that should have been kept, you can record it again. This is sometimes a question of judgement because there may be a sound you don't want on the tape next to something that the player only got right by the most incredible stroke of good fortune. If this is the case then don't take chances; but always keep in mind that the cleaner you keep the tape during track laying, the easier the mix will be.
Most of the instruments you record, particularly distorted electric guitars, will produce a lot of noise as well as music. The cure for this is the noise gate. Sometimes it's best to leave gating until the mix, and sometimes it's better done during track laying. For instruments that aren't usually played with great subtlety, such as the aforementioned distorted electric guitar, it's fine to do the gating as you record the instrument, and then you have the gate free for something else during the mix. But whatever else you gate, it's hardly ever a good idea to gate vocals. Vocalists are so unpredictable, and their contribution to the song so important, that you can't risk losing any sound they make by using a slightly over-eager gate in the signal path. It's hard to gate out things like breaths anyway, and I've taken to copying the finished vocal to a spare track, for safety, and carefully spot erasing any noises I don't want to hear in the mix.
Of course, what you do with the vocal track depends on the type of sound you are aiming at. Sometimes a thick arrangement will cover odd little noises on the multitrack, and sometimes they may even add to the overall texture. When recording sequenced synths and samplers it's easy to get carried away with the idea that since they are pretty well a controlled sound source in the first place you can just 'bung them down' on to tape, but with some synths the small amount of noise they make quickly builds up if you lay down several tracks with the same instrument.
"There's no such thing as the perfect multitrack recording, all ready and waiting to be mixed down with no EQ, level changes, effects or jiggery-pokery of various kinds."
Another source of annoyance at the mixdown stage is instruments or vocals that vary considerably in level throughout the track. This means that it's hard to get a good balance without dodging the fader up and down simply to compensate for this defect, and harder still to enhance the mix. The tool for dealing with this is the compressor, either at the recording stage or during the mix. Compression during recording is theoretically better because it doesn't boost tape noise, but for important tracks like vocals I prefer to record clean and then use processing later.
Apart from the inevitable unpredictable problems that always seem to crop up, the other big problem with multitrack masters is that they will in all probability have more music recorded on them than will be used in the mix. Perhaps there will be two or three tracks containing attempts at the guitar part, all of which sound good, but you don't know which to use. Unfortunately, you will only know which is best when you hear them in context with all the tracks properly treated and balanced. There is then no problem, provided there is an entire track that can be dispensed with, but usually you want to keep part of one track, part of another, and so on. It's all part of the job.
There is no single right way to go about mixing a multitrack tape, but there are a number of techniques you can employ that may get you up and running with the mix, or which may come to your rescue if things don't seem to be going anywhere. My favourite way of working, perhaps because it suits the style of my music, is to concentrate on getting the best out of each track in turn, and then go about putting them together in a defined order following a master plan. Let me elaborate...
Individual tracks on the multitrack can usually be divided into two categories: thin and weedy, needing extra support; over thick and congested, needing 'sharpening up'. The third category of tracks that are in-between, and therefore just right, is so small that it can be neglected for all practical purposes. In fact, you can take a positive view of all these problems, as a necessary part of the creative process, because in solving them you'll give your music an individual style that simply wouldn't come about if everything went smoothly. As with everything in life, you have to take the rough with the smooth, but a bit of rough can sometimes make things much more interesting!
The first category, of 'thin' sounds, usually consists of clean samples from a drum machine or sampler, and unprocessed synth sounds. In the early days of drum machines, sequencers and DX-type synthesizers this thinness was a problem because processing equipment was still very expensive and multi-effects units unheard of. Now that we have a whole range of tools at our disposal, there is no excuse for a thin, weak mix which gets sand kicked in its face every time it goes down to the beach.
Thickening of weak sounds can be done at any stage from track laying to mixdown. One of the best thickeners is to simply play the sound through an amplifier and speaker and mic it up. This is good not only for guitars, but for synths and samplers too, and even vocals if you're after an interesting effect. Even in these days of super-electronic, hyper-computerised mega-effects, a small guitar practice amp, preferably with a few valves in it, can work wonders with a dull, lifeless sound. In normal use, small amps aren't much good for bass because of the size of the speaker, but for recording it doesn't matter because if you put the mic near the cone you'll find all the bass you need, and the miked up sound can always be mixed with the clean, dry, sound if necessary. If you're doing this in the mix, the only problem is that the sound from the amp will prevent proper monitoring, unless you can put it in a separate room. If you have a spare track, simply copy the beefed up sound to that, then mix as normal.
"Now that we have a whole range of effects at our disposal, there is no excuse for a thin, weak mix which gets sand kicked in its face every time it goes down to the beach."
One thickener that can easily be overused is reverb. You get a nice thick sound alright, but sometimes the music gets drowned in a swimming pool acoustic. When using reverb as a thickener, there are two techniques which you can mix to taste. The first is to use a very short reverb at a fairly high level. (A gated reverb is even better because you get the bulk of the effect very close in time to the original sound. It doesn't sound so much as though it has had artificial reverberation added, but it does have more body.) The other thing to do is to add reverb to a track in mono. This may sound odd in such a stereo-conscious age, but it does actually work, and it's very often preferable to stereo reverb because the mono version 'sticks' to the sound, whereas stereo reverb separates out into the left and right speakers. If you haven't tried it, then have a go right now, but remember to pan the reverb to exactly the same position as the signal. With stereo reverb, usually everything is sent to the same reverb unit via a single auxiliary send on the console, but the best way to add mono reverb, if your reverb unit has a reverb balance control, is to patch it into the channel insert point.
Still on thickening, but bypassing compression, chorusing and flanging because they are well known as thickeners, the console's EQ can also be used to beef up sounds. Many new engineers start EQing by boosting whatever frequencies they feel they need more of. But remember that EQ can be used to cut frequencies as well. Every instrument has a certain frequency range within which it is most powerful. If you boost this range, you make the instrument almost into a caricature of itself, just as a cartoonist will exaggerate a politician's most prominent feature. But if you cut this frequency range, then you will let the other frequencies the instrument produces come through more clearly. Once again, the only way to really understand this is to try it, but it certainly works and can be a powerful technique.
When you are recording live instruments, the problem is more likely to be that the instruments are not clear enough. In this case, EQ boost as described above can be applied. In extreme cases, a graphic equaliser can be used to filter out the rubbish surrounding the signal you want, but apart from this there are fewer options to make a sound more clear than there are to muddy it up a little.
Once I have all the individual sounds as I want them, without processing them to extremes, I start to build up the mix. I'm not using 'build' in a loose sense, but as in the erection of a skyscraper. I lay the foundations, construct the framework, top it out, and then attach the cladding. A pretty good analogy, even if I say so myself.
The foundations of the mix will probably, though not necessarily, be the drums and bass line. With a rhythmic piece it's most important to get the rhythm section to support the rest of the structure properly. I start by setting the bass drum so that the meters read about 10dB under maximum recording level, as a rule of thumb, and then balance the snare according to the type of drum pattern. After this, I bring in the bassline and the rest of the drums in whatever order I feel like, but I spend time making sure that they all pull together in the same direction. If I get it right at this stage the rest will be easy.
The framework consists of the 'pad' instruments — those which are there to fill in the harmonic structure. If the pad is synth strings from my Roland D50, I pan the outputs left and right and bring the channels up to a level where they fit snugly against the bass line — neither too quiet nor competing for attention with the bass. At this stage I can top the structure out by bringing up the vocal line, or melody synth line if it's an instrumental piece (which most of mine are). The rest of the mix is simply cladding — it's there to fill in the gaps and keep the chill wind of boredom out. A bit of decoration doesn't go amiss either, at the right level.
"The foundations of the mix will probably, though not necessarily, be the drums and bass line. With a rhythmic piece it's most important to get the rhythm section to support the rest of the structure properly."
Following this structural plan is very straightforward if all your sounds are good, or you have made them good. Hopefully, once you start bringing up the faders you will only need the odd tweak on the EQ to bring everything together. Reverb will also be necessary, mainly to fill in any structural gaps which remain.
Figure 1 illustrates an alternative way of creating a good mix. Instead of building a skyscraper, you draw a map. Imagine that the speakers are a window into another world — a world of sound if you like — and you are drawing a plan view of that world within which the sounds have their own positions. The first thing to do is to work out (from the track sheet) where you would like all the instruments to go. Important instruments will obviously come nearer the front, supporting instruments will be further back. Once you have a plan, involving specific instruments rather than my more generalised diagram, you can start putting it into action. The easy part is positioning the instruments between the left and right speakers with the pan control, but you also have to move them forwards and backwards. To send an instrument to a more remote point in the audio landscape you have to do three things: lower the level; cut high and low frequencies; add reverb. These three things are very easy to do, but it's not so easy to do them in the right way in order that it really does sound like the instrument is further away. Listen very carefully to what you are doing to each instrument as you move it, and consider whether it actually has moved to where you want it to be.
Whereas a mix performed according to the skyscraper method will have the principal characteristics of clarity and punchiness, a mix performed by the map method will sound 'natural', 'organic', 'orchestral'. I don't particularly like choosing adjectives to describe music, but I hope you now have an idea of what it's all about.
Spock: "Random factors seem to have operated in our favour, Captain."
Kirk: "You mean we've been lucky?"
Harnessing the random element in music can be a great aid to creativity. One way to do this in mixing is to throw the faders up on all tracks to random levels and then have a good hard listen to what you have. If these random levels turn out to give a good mix then you will indeed have been extraordinarily lucky, but conversely if there is no value whatsoever in the chance relationships between the the various instruments, then that in turn is very bad luck. There will almost certainly be something good there. The idea is to listen carefully, and then make adjustments to turn a chance mix, with some good points, into a wholly good mix with some features that you wouldn't have thought of yourself if you had worked at it all day.
Sometimes, when you have had a hard day at the console and things seem to be getting nowhere, the only answer is to turn to lady luck and see what she has to offer — but this isn't an act of desperation, it's another creative tool, and fortunately for us it costs nothing and is absolutely tax free. Also tax free, or at least zero VAT-rated, are still more ideas on mixing and on recording techniques in general which will appear shortly in a newsagent near you. See you next month.
Feature by David Mellor
Previous article in this issue:
Next article in this issue:
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!