Mixdown Lowdown (Part 1)
Paul White again, with the first in a new series on ensuring your immaculate recording isn't ruined by a mediocre mix. This month: what not to do with mixing desks and outboard processors.
You may have the finest piece of music this decade has seen, brilliantly played and beautifully recorded... it can all be ruined by indifferent mixing. In the first part of a new series, we guide you through some basic points about creative mixing.
CREATING MUSIC, particularly pop music, is not unlike painting a picture. A cliche, I know, but think about it for a while: artists' approaches can range from the emulation of real life (realism) to a completely unnatural style that appears to break all the rules (surrealism), and much the same can be said of composers, musicians and producers. And while painters rely on a palette of colours and painting techniques, producers use the mixing console and effects to shape and direct what has already been recorded onto tape.
This phase of recording is incredibly significant, as the producer brings together several independent musical elements that may originally have been separated by space, as well as time, and blends them into a cohesive whole.
Just how the sounds are mixed and treated depends on the result you want to achieve, but the processes can usually be divided into two groups: those that are necessary as a result of the physical limitations of recording equipment or performers, and those that are applied to create a specific, er, creative effect.
I'll call these two areas corrective and creative, though as we'll see later, they can overlap.
BY THIS, I mean the type of processing that has to be applied to minimise noise or crosstalk, control dynamic range, and to correct for tonal inadequacies created somewhere along the line.
The first line of defence here is the noise gate, but there are more sophisticated expanders and dynamic noise filters available, so there's no shortage of weaponry in the fight against noise.
Where possible, gating is best applied on a track-by-track basis rather than on a complete mix or submix, because a gate is only useful if there are gaps in the music where it can be allowed to close. When the signal is present, so is the noise, so there's no point in having the gate open at the start of a mix, only to stay open to the end: you gain nothing other than a quieter start.
But gating a single track, such as a vocal line, effectively mutes the tape noise between words, and as an added bonus, cuts out any crosstalk that might be on the track. However, you have to set the gate's controls carefully to avoid losing any wanted portion of the sound. The Threshold and Release controls are vital here, and you'll always have more success gating a dry sound and then adding delay effects to the gated signal, than you will gating a signal containing a significant amount of reverb or echo.
Apart from the obvious problem of making sure the reverb tail doesn't get chopped off, adding an effect such as echo or reverb to the gated sound may well hide any premature cutoffs caused by imperfect gating, or by awkward signal dynamics.
As a rule, expanders are a little more forgiving when it comes to setting-up, but are still best used on a track-by-track basis. But dynamic noise filtering can be applied to a complete mix with minimal side-effects, providing you're careful not to apply more than the bare minimum of processing needed to improve the noise performance to an acceptable level. Used on individual tracks, these devices work well as they're almost undetectable in operation - though it's rarely cost-effective for the small studio owner to own more than a single two-channel unit.
As always, prevention is better than cure: it's always worth trying to minimise noise when recording, by keeping an eye on levels and planning track bounces carefully.
THIS IS a complicated way of saying 'controlling levels'. Obviously, you have control over levels using the sliders on the mixing-desk. But what I'm referring to here is the use of compressor/limiters to tame those sounds which vary too much in level to be usable as they stand, and which do so in such an irregular way that correction using console faders becomes difficult, if not impossible.
A professional studio might well use a bank of compressors to control several tracks within a mix, but again, the small studio is likely to be limited by financial considerations. From experience, I've found that the singer whose voice doesn't need some amount of compression is a rarity, though you can get away without compressing most other parts of a mix, so long as they're carefully recorded and competently played.
It's possible to get around the problem of having only one compressor by applying some compression to the vocal track as it is recorded. This has the advantage of giving you a higher than average signal level on tape, and consequently a better noise performance. But you can also compress the signal coming off tape, just so long as you bear in mind that every dB of compression is a dB of degradation in the signal-to-noise ratio of that particular track. Again though, this isn't too much of a problem if you keep the amount of compression as low as possible, bearing in mind the job in hand. Many units now feature built-in expander gates, and these are a great help in cutting out any noise brought up by compressor action.
Finally, a few words about compressors' user-variable parameters and their implications. You'll probably be familiar with the term 'compression ratio', but in this instance it doesn't refer to a new Jaguar, or indeed any other form of motor transport; it's simply a way of expressing the change of output level in dBs for a given change in input level. So, a compression ratio of 2:1 means that a 2dB change in input only causes a 1dB change in the output level. The higher the ratio, the more severe the compression, and the less effect any change in input level has. A really high ratio may be called limiting, as the maximum output level becomes virtually independent of the input level, and you have, in effect, a fixed maximum output signal level.
For normal applications like levelling vocals or bass guitar, low ratios between 1.5:1 and 4:1 are adequate. And the lower the ratio, the less obtrusive the processing.
"Most singers' voices need compression, but you can get away without compressing most other parts of a mix, so long as they're well recorded and played."
Apart from the Ratio control, there are also Attack, Release and Threshold control settings to consider on compressors. But these are simple enough to choose, so long as you understand what they are actually doing.
The easiest way to explain the action of these controls is to look at what used to happen before compressors were invented - gain riding. This was the process whereby engineers had to change the gain manually during a mix, using only the mixer's faders. As soon as the engineers heard that a sound was getting too loud, they'd turn the gain down. They'd be too late to correct the first part of the sound because they couldn't respond instantly, so some of the loud sound would get through before it was brought under control.
This delay in responding is analogous to the Attack parameter on a compressor. After all, the compressor also takes time to respond to a rise in level, even though it's obviously much faster than any engineer. With a fast attack time set up, there's little overshoot because the compressor reacts very quickly, but there are occasions when it's desirable to have a slower attack time, to lend emphasis to the leading edge of a sound. If this is the case (typical examples of this kind of application are adding impact to bass-drum sounds or bass guitars), set up the Attack control by ear.
The release time is the time it takes for the compressor to stop applying gain reduction once the input signal level has fallen below the threshold. If the release is set too fast, the signal seems to pump. If it's set too long, the compressor hasn't time to recover and applies gain reduction to following low-level sounds that don't need it. If you're in doubt as to which setting to choose, set a fast attack time and a release time of half-a-second or so, and start from there.
On a gate, the threshold is the level above which no change to the signal takes place; for a compressor, it's just the opposite. A signal below the threshold passes through unchanged, but once it exceeds the threshold, the compressor starts to apply gain reduction.
When compressing vocals, I tend to set the threshold so that gain reduction is only just starting to occur on average-level signals; most compressors have a readout of some sort to show the gain reduction. With this method, low-level signals remain unchanged while the louder sections are controlled, which in turn allows you to use a higher average level in the mix so that quieter sections don't get lost.
PRETTY STRAIGHTFORWARD, this. You should be familiar with using the EQ controls on your own mixer, but what do you do if these aren't enough to cope? Well, you can patch in an external equaliser - either graphic or parametric. Either method will let you home in more accurately on the section of the audio band that needs treatment. The parametric offers more precise control, but isn't as easy to use as a graphic.
If you need to cut a frequency because of an unpleasant colouration in the sound, the best way to find it when using a parametric or sweep equaliser is often to set the equaliser to maximum boost, and then home in on the offending area using the frequency control before applying cut. This is a lot easier than setting the equaliser to cut, and then vainly hunting for the area that needs cutting.
Graphic equalisers are easier to use because the slider positions give a good indication of what is going on. But if you don't have a graphic with a lot of bands, the frequency you want to affect might fall between two bands where you can't tweak it without affecting frequencies to either side. However, the graphic does give you the chance to equalise different parts of the audio spectrum simultaneously.
But there are occasions where EQ alone can't help because it can only affect sound that already exists. So what do you do when faced with a muddy sound that refuses to be brightened? Well, you could turn to the trusty psychoacoustic enhancer. Such a machine generates additional upper harmonics to complement whatever input signal is fed into it, synthesising a new top end rather than trying to boost one that doesn't exist.
Whether you use enhancement or EQ, adding top invariably means emphasising any noise present in the recording, so you may need to use gates or a dynamic noise filter to keep things clean if the recording is especially noisy in the first place.
SO FAR so good. You've got a whole heap of outboard toys, and you've got a good idea of what they're used for and how to go about using them, at a theoretical level. But where do they fit into the scheme of things when it comes to connecting them up? Exactly what plugs into where?
As far as connecting up goes, there are two basic types of signal processor: those that treat and alter the whole signal, and those that create an effected version of the input intended to be mixed back with the dry signal.
In the first category we have effects like gates, compressors, equalisers, enhancers and dynamic noise filters. Whatever is fed into them is processed, and the output then replaces the original signal.
"If you use an, input channel as an effects return, make, sure the auxiliary sends are turned right down, or you'll create a feedback loop and the whole system could start shrieking..."
The second category comprises effects like echo units, reverbs, multi-purpose DDLs and so on, which generate a treatment that needs to be added to the dry sound before it sounds acceptable. Some of these units contain a Balance control so that some dry sound may be mixed into the output but, when used in conjunction with a mixer, it is normal to set these so that only the effected sound appears at the output. This way, the right amount of effected sound can be added to the dry sound using the console controls.
Any piece of equipment from the first category should be connected via the mixer's insert points, and should not be connected to the auxiliary or effects send circuit. If you have no insert points, you can unplug a tape return from the mixer and feed it into the processor; the output of the processor will then feed the mixer input.
If you want to effect the entire stereo output, you have to connect a stereo or two-channel processor to the master insert points, or between the mixer and your master tape recorder if insert points aren't available.
When using a two-channel unit to process a stereo signal, use the stereo link switch (if one is fitted) so that both channels track properly. This is especially important with compressors, gates or dynamic noise filters, as the control circuitry needs to be fed from a mix of the two input channels if serious image shifts aren't to appear.
If you do have to patch in between the mixer and the mastering machine, remember to monitor the two-track output rather than the mixer's left/right outputs, otherwise you won't know what effect your processing is having.
If you're using anything other than EQ, this last arrangement can make fading difficult. For example, if you're using a level-sensitive device such as a compressor on the whole mix, it'll try to turn up the gain as you're turning it down. You will win in the end, but the fade won't sound right. If you're forced into this situation, you can avoid the problem by doing the fade using the level controls on your mastering machine, rather than the faders on your desk.
The second category of processors is generally connected via the auxiliary (or echo) send controls on your desk, because in this way, several channels can share the same effects unit without being restricted to the same amount of effect. But if you've already used all your effects send facilities, you can patch in, say, a single DDL to a specific channel via the insert points and use the DDL's own mix control to set the dry/effect balance. This limits you to using that effect on only one channel (or subgroup), but that may be all you need.
Echo send controls are known as postfade auxiliaries, because they're connected internally at some point after the channel fader in the signal path. When the channel gain is reduced, the send to the effect unit is also reduced. So, turning a channel down will also vary the level of the added effect in exact proportion. And if you turn the channel right off, the effect will be turned off too.
The other type of auxiliary control that you'll almost certainly have on your desk is the foldback or pre-fade auxiliary. This is independent of the channel fader, and it's important to bear this in mind if it's used to drive an effects unit. If you have a reverb connected here, for example, you can set a perfect balance between the dry and the effected signal. But if you then fade the dry signal using the channel fader, the effect level will remain the same as it was before. So, this method of connection should only be used if you're not going to change the levels much within a mix, or if you actually want to fade the dry sound and leave only the effect present as a specific production trick.
Take care with effects when subgrouping, too. Imagine a situation where you have several channels, each with effects added via the auxiliaries, all mixed down to one or two subgroups. You'll find that unless your effects returns are routed to the same subgroup or subgroups, you'll get the same problem: turning down the subgroup level using the subgroup faders will not affect the level of any added effects.
On the other hand, you may want to use a specific effects unit on some channels that are subgrouped and some that aren't, so you can't simply route the effects returns to those subgroup faders. You have a problem, and one for which there is no easy answer. Don't give up, though: if you can anticipate these difficulties before they arise, you can often plan around them.
Now let's look at effects returns, or auxiliary returns as they are sometimes known. These are often considered to be something separate from the rest of the mixing console, but in reality they are just simplified input channels; you can plug line inputs into them, or you can plug effects into channels and use them as returns.
The only real difference is that the auxiliary returns only work at line level, and their facilities are much simpler than those found on a main input channel. But if you do use an input channel as an effects return, make sure the auxiliary sends on that channel are turned right down, or you'll create a feedback loop and the whole system could start shrieking at you.
And the last thing you want, when you're in the middle of an important mixdown, is something, somewhere, starting to shriek at you.
Feature by Paul White
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!