by Paul White mixing methods & approaches
Although there are no hard and fast rules in mixing, PAUL WHITE provides some handy guidelines and ways for you to approach your own mix.
It is fascinating to see how music mixing has changed over the past two or three decades. In the Sixties, arguably the heyday of pop, the sound of a record came almost entirely from the performance of the musicians -- the mixing engineer's role was largely concerned with setting up an acceptable musical balance and adding a touch of plate reverb where needed. Today, music production is all too often less about musicians performing together and more about assembling a composition in pieces, rather like a jigsaw puzzle, where those involved are not entirely sure of what the picture is supposed to be until the puzzle is finished. Then, of course, there's the use of pre-recorded material to consider, whether it be snatches from a previous recording, samples of live musicians playing rhythm patterns or riffs, or 'found' sounds from TV and radio to add atmosphere and effects. Due to the creative nature of contemporary music production, today's mix engineer must be every bit as much an artiste as the musicians on the other side of the control room glass. So, is there any common ground with the mix engineer of the Sixties and Seventies?
Regardless of the style of music, or whether the sounds are samples or recorded performances, the mix engineer still has to balance the various parts, handle the stereo positioning of sounds within the mix, equalise and apply effects where necessary. In professional circles there's invariably a producer looking over the engineer's shoulder who calls the shots, but most SOS readers double up as engineer and producer, and quite often as the artiste too -- which means knowing something about music production as well as engineering. So how do you set about turning a good recording session into a good master tape? Here are a few mixing guidelines for you to follow...
APPROACHING THE MIX
Once the recording stops, and ideally after a suitable break, the mixing session starts. At this point you should have track sheets describing what's on the various tape tracks, any notes taken during the recording (MIDI instrument settings), and all the mixer EQ, level and routing controls set to 'neutral' ready for a clean start. The only exception to this latter condition is if you've been setting up your mix as you were recording, in which case it's simply a matter of checking that nothing is routed that shouldn't be.
It helps to separate logical groups of sounds into subgroups, so that the mix can be handled with fewer faders. This is particularly helpful if you don't have the benefit of mix automation. For example, if the recording includes real drums, these are likely to occupy several tracks so it makes sense to assign them to a single, stereo subgroup. Other candidates for subgrouping are backing vocals and multi-layered keyboard parts.
Unless the quality of the original recording leaves something to be desired, it should be possible to set up a workable balance fairly quickly, without resorting to EQ. Effects can be added later, but it helps to have the necessary effects units patched in and ready, especially reverb which will probably need to be added first.
In order to keep the sound quality as high as possible, de-route any mixer channels that aren't being used, turn down their aux sends, and mute any monitors that aren't being used during the mix. Also check the gain setting of any effects you have patched into the mix, so that their input level meters read a healthy signal level on signal peaks. The loudest channel aux sends should be set to between three-quarters and full up, and the input gain control of the effects unit adjusted accordingly.
If it is possible to route unused sends to an aux bus that isn't being used, this can reduce mix noise considerably, and any effect needed only on a single channel will be quietest if connected via the channel insert point rather than via the aux send system. If mixer noise is a problem, it can sometimes help to patch a gate between the console's effects send outputs and the effects unit inputs.
GETTING THE BALANCE
A good mix starts with a solid foundation, so it pays to sort out the rhythm section first, but don't spend too much time on things like the type and level of reverb on these instruments, because this will sound very different once all the other instruments are in the mix.
Once the rhythm section is working, then add in the lead vocal followed by the other parts of the mix, always being aware that you must leave space for the vocal. Up to this point, it can help if all your sounds are panned to the centre of the mix -- if you can get the mix sounding good in mono, it'll invariably sound even better in stereo. It's only when I've reached this stage that I start to worry about EQ, but I know other engineers who simply push all the faders up and try to sort everything out at once. There's no right or wrong way as long as it works for you, but if you don't have a lot of mixing experience, I think you'll find my approach less stressful.
Try not to monitor too loudly, as this will affect your judgement and may eventually damage your hearing. The most logical monitoring level to use would be at the same level as the end user is likely to listen at -- in other words, a sensible domestic listening volume.
Once the balance is working in mono, and maybe you've added a little vocal and drum reverb, you can concentrate on refining the effects, the EQ settings, and the stereo pan positions.
Traditionally, all the 'heavy guns' go in the middle of the stereo soundstage, by which I mean kick drums, bass guitars and bass synths. Aside from the fact that very low frequencies don't really benefit from panning, it helps if you distribute high energy bass and percussion sounds equally between the two speakers. Lead vocals also tend to be panned to the middle, not for any technical reason but purely because we expect a singer to be centre-stage. I invariably add some compression to vocals while mixing, simply to even out the level, and if noise is in any way a problem, then I'll also gate the vocals before the compressor. A useful compressor setting is to choose a fast attack combined with a release time of 0.5s or so and a ratio of between 4:1 and 12:1, depending on whether you want to hear the compressor working or not.
Backing vocals, on the other hand, can go wherever you want to place them, and you might want to compress the overall backing vocal mix. If this is in stereo, don't forget to use the compressor set to its stereo link mode. I like to hear different backing vocals coming in from different sides, but it's your song -- so put them where you like! If you're experiencing sibilance problems on your vocals, try a less bright reverb setting or try adding a little EQ cut at around 6kHz. If the problem persists (which it shouldn't do if you were paying attention at the recording stage!), you may need to bring in a de-esser. Having said that, I've never yet used one on a serious recording, as I feel they compromise the vocal sound to too great an extent.
When the mix finally starts to happen, it helps to take a break, have a cup of tea, listen to a few records, and then come back to it. I find it invaluable to listen to the mix from the next room, with the adjoining door left open, as even the slightest balance problems become very obvious. I mention this technique at regular intervals, because of all the tricks I've learned over the years, this is one of the most helpful.
On most mixes you'll need to do a little gain riding to sort out awkward vocal levels that the compressor can't handle, or to bring solos in and out, but again, listening from next door can help you identify the areas that need manual attention and those that can be left alone. One general rule is not to mess with the rhythm section level once it's set up, as this would run the risk of upsetting both the overall balance and the continuity of the song.
If there are several level changes to handle during a mix and you think you might run out of hands, then rope in the musicians to help, but always put wax pencil marks on the desk for them to follow, otherwise you might find the balance changing with every pass!
"A good mix starts with a solid foundation, so it pays to sort out the rhythm section first."
If a track requires a fade-out ending, make sure you start to fade at least 20 seconds before the recorded material runs out, and don't rush the tail end of the fade or it will sound unnatural. If the album is going to be compiled on a hard disk editing system, such as Sound Tools, don't bother with the fade when mixing but do it as part of the editing process; this will be smoother and will fade into true silence.
In a perfect world, every tape track would contain a perfect performance at exactly the right level with no noise or unwanted sound to be heard -- but life is rarely like that. Apart from some degree of tape and circuit hiss, and maybe the odd dB of hum, there are always extraneous sounds such as breath noise or digital synth grunge to consider. Gates are very effective in cleaning up noisy tracks, but care must be taken to match the release time of the gate to the natural decay envelope of the sound being treated. However, gates can only keep the noise down during pauses, they can do nothing when a signal is present. It stands to reason, therefore, that if you decide to gate a whole mix, the only real benefit will be a clean start and a clean end.
If your mixer has MIDI muting, this can be set up to kill any channels when they are not in use, thus reducing the level of cumulative noise build-up in the mix. It is necessary to go through each tape track and set up the mute points individually, and if you can arrange muting and unmuting to occur on a beat, it will help to disguise any sudden change in background noise level.
While gates can only clean up pauses, dynamic noise filters can actually remove noise in the presence of signal, though you have to take care that they don't introduce audible side-effects. Dynamic filter units simply filter out the higher frequencies when the signal level is low, and though they have no audible effects when the treated signal is strong, they do tend to affect the tail end of long reverbs. For this reason, it helps to route the reverb via one subgroup and the channels to be de-noised via another, so that the reverb escapes treatment.
FIXING THE MIX
Occasionally you'll end up with a mix which still needs that extra something, especially if you're working on a tape recorded by someone with different ideas to yourself when it comes to what things should sound like. Compressing a complete mix will reduce the dynamic range and increase the average energy of a mix, but as contrast is a necessary part of music, you might find the mix gains in one area and loses in another.
The attack time of the compressor may be increased to 20ms or so to allow transient sounds to cut through, though the type of compressor used can make a huge difference to the subjective outcome. Soft-knee compressors produce the most unobtrusive results, but the other side of the coin is that an obviously compressed mix can also sound quite exciting and vibrant, which is why certain vintage valve compressors are so popular.
If your mix is correct in the first place, why should it need any further EQ? I can't provide the complete answer, but I do know that some equalisers are capable of flattering even the very best mixes. Music can be made to sound 'louder' by gently cutting the mid-range slightly, and it's quite common to treat a whole mix with an exciter or a dynamic equaliser to add sparkle and detail. Since buying my SPL Vitalizer, I invariably use it when mixing and there's simply no way to simulate the effects using conventional EQ.
LISTEN WITHOUT PREJUDICE
Before finally approving your mix, make sure you listen to it on as many different stereo systems as possible, including the car, otherwise you run the risk of creating a mix which sounds good only in your control room. And if you have to ignore all my suggestions in order to get your mix sounding the way you want it, that's fine too -- there are no hard and fast rules and the end always justifies the means! Have fun.
Occasionally you get a mix that just won't sound right, often because the song hasn't been arranged well enough to leave space for all the important parts. If you come up against one of these, here are a few tips you can try. By working through the following points, you should at least end up with something usable.
- Set up a rough mix, in mono, without using EQ or effects and then work from there. Also check that the mix sounds OK in mono.
- If the mix seems too busy, ask yourself if all the recorded parts are really necessary or can you lose something (try muting each part and see what impact that has). What matters most in the majority of pop songs is the rhythm and the vocals, the rest is decoration. If you can't lose something completely, try mixing it so low that you only notice it if you turn it off.
- If the mid-range sounds are fighting with the bass sounds, try using EQ to 'thin out' some of the sounds. Take some bottom end out of the pad synth, backing vocals or acoustic guitar parts.
- Still no joy? Then go back to the basic rhythm section plus vocals and see if that is working. If not, is it too late to try a different drum or bass sound? Similarly, if you're working with a sequencer, you could try picking thinner pad keyboard sounds or brighter bass sounds.
- Don't overdo effects -- reverb creates the illusion of distance and space in a mix, both of which are the opposite of 'up front' and 'in yer face'. As the eskimo said when burning his canoe to keep warm, "You can't have your kayak and heat it!"
- If things are looking up, try panning the instruments and effects to their desired positions -- this will help improve the separation between individual sounds and enhance overall clarity.
- Though exciters shouldn't be thought of as a means to salvage poor recordings, the extra separation they create can make the difference between a so-so mix and a good one. Use as a last resort only when you've got everything as good as it can be.
- Perhaps the mix sounds fine but just lacks cohesion or punch. In that case, try a little overall compression. A soft-knee compressor will usually provide the most transparent results, but try whatever you have and let your ears be the judge.
Digital reverbs create the illusion of stereo by synthesizing different sets of delay taps for the left and right channels, which makes the reverb patterns slightly different between the left and right outputs. This makes it possible for us to take a mono tape track and give it both a stereo identity and a sense of being somewhere, rather than existing in a void. Even if no obvious reverb is needed, a sound can still be given substance and width by adding a very short reverb, ambience or early reflections pattern to it.
For drums and vocals, where a longer reverb time is often chosen for artistic reasons, try to pick a setting that doesn't fill up all the space and stifle the mix; it may help to add a pre-delay of around 50ms or so. If the reverb makes the mix sound muddy, feed the reverb back through a channel that has EQ and roll off some of the bottom end. Alternatively, if the reverb is diluting the stereo image of a sound too much, try panning the instrument sound and its associated reverb to exactly the same point in the mix. This will kill the stereo width effect, but can be effective where a sound needs to emanate from a precise location.
Avoid putting more than the barest hint of reverb on bass drums or bass instruments (unless for deliberate effect) unless the mix has loads of empty space to allow the reverb to breath, without clouding the overall picture.
Try panning an instrument (eg. guitar) to one side of the mix with a delayed version (from 5 to 50ms) panned to the other side. The sound will appear to be coming from the speaker that's carrying the unprocessed (dry) sound, even if the delay is as loud as the original signal. The psychoacoustic reasons why this is so are rather too complex to go into here, but this does provide another way to add space to a sound.
If the delay is then modulated to produce a chorus sound, the result is to create the illusion of movement, and when listening in stereo you really can't tell that one channel is carrying a dry sound and the other a processed version -- the movement seems to occupy the whole of the space between the speakers.
Delay can be used to create more conventional echo and doubling effects, of course, and it has become fashionable to set up synchronised delay times that are a multiple of the tempo of the song. For example, if a song is running at 120bpm, each beat is 60 divided by two seconds long -- which is half a second. Therefore, a delay of 500ms (half a second), 250ms, or 125ms will always create echoes that are in sync with the music. You can also divide the beat time into threes to create echoes that occur in triplet time. Clever use of delays can help add drive and push to a song (check out The Edge's guitar playing on most U2 tracks).
EQ is complex enough to warrant a complete article in its own right, though I subscribe to the school of thought that recommends leaving it alone unless desperately needed, and even then using as little as possible. Using EQ to cut rather than boost, where possible, invariably results in a more natural sound. For general brightening of a track, try either a subtle amount of boost at 6kHz or a hint of high shelving boost (usually 10-12kHz on most desks).
EQ can be used to create separation in a crowded mix by using it to narrow the area of the spectrum occupied by a particular instrument or voice. Most sounds have the bulk of their energy in one section of the audio spectrum, but there will also be small amounts of high and low frequency energy outside this band. By using high and low EQ cut to 'trim' away these extremes, it may be possible to make a sound sit more comfortably in the mix. Even though such EQ'd tracks may sound a touch unnatural in isolation, they may still work well once in context. Electric guitars often benefit from this kind of 'spectral trimming' as do acoustic guitars (to take out some bass end), some drum sounds, and backing vocals. Acoustic instruments are best treated gently with maybe just a little LF cut. Lead vocals, or vocals that are very exposed in the mix, should be treated most cautiously of all. It's invariably better to get the right vocal sound at the outset, by choosing a sympathetic mic, rather than by using EQ later.