We've all done it -- spent hours working on a mix, only to feel like tearing it all down and starting again. The more work you put in, the worse it seems to get. Yet when you listen back to your early work, often done on a 4-track with virtually no outboard equipment, the results can be surprisingly good. This, at least, should give you some clues as to what's gone wrong with your later mixes. If your equipment is simple, you're forced to work simply -- recording as few parts as you need to make the song work.
The only effect you have, if you have any at all, is probably reverb, and your mixing facilities may well only have treble and bass EQ. The outcome is that you record the individual sounds more or less as they are, balance the result, and that's about it. In my case, my first reverb was a spring unit that was so twangy, I had to use as little of it in the mix as possible. Any guitar effects were added during recording, and were usually confined to chorus or echo. Because there were no DI solutions to guitar recording around at that time, I miked up my amplifier.
So, what does this reveal? Basically, it tells me that if you record the various elements of a good performance without messing about with them, and then adjust the balance, the result will be at least acceptable, and at best, excellent. It's true that some things can be improved by using EQ and other forms of processing, but a good rule is to at least record everything with no EQ, because then, if the mix isn't working out, you can easily get back to a clean slate by re-setting the channel EQs to flat.
Some engineers insist on adding EQ when recording, but I don't subscribe to this way of working myself. You don't know what EQ a sound finally needs until you hear it in context with the rest of the mix, and if you're building up a mix in layers by overdubbing, how can you possibly know what EQ you'll eventually need? What's worse is that if you do make a mistake, it's virtually impossible to un-EQ a sound, so that it sounds as it did originally.
A good mix starts with a good performance of a suitable musical arrangement, and unless everything is in tune and in time, there's virtually nothing you can do to fix it, short of sampling the few bits that work, and building a new arrangement out of the rubble. That kind of work is outside the scope of this article, though mix salvaging is something we'll be looking at in the near future. However, one very common problem you can do something about is arrangements that get too busy, resulting in all the sense of space and separation being lost.
A good arrangement will also take account of what instruments are playing at the same time, so that you don't end up with three things playing at once, all occupying the same space in the musical spectrum. While instruments such as brass and strings blend nicely, the whole point there is to create an ensemble effect which can really be treated like a single sound. If, on the other hand, you have a middly bass guitar, a low, thick synth pad and a woolly-sounding rhythm guitar, then the chances are that the mix is going to sound confused. Again, some engineers tackle this problem by trying to force similar sounds into being different by using heavy EQ, but the only really satisfactory approach is to choose the right sounds at the outset. EQ should be regarded as a tool for fine polishing, not as a large hammer and chisel. Invariably, there are exceptions to this rule once you start getting creative, but when it comes to making instruments and voices sound good, I've always found that the less EQ you can get away with, the more natural the results.
If you're creating sequenced music, MIDI gives you the choice of changing sounds before you get as far as the final mix. Even so, you still need to be careful, because many modern synths sound so rich and impressive that they can take up all the space in your mix. This is particularly true of factory presets that are designed to sound impressive when played on their own in the shop. Creating thinned-down versions of some of the factory presets can be a good idea, but it's always better to create your own sounds if you can.
Continuing on the subject of sounds, you may find that once the recording has been made, you have far less control over things like electric guitars and basses than you think. For example, it's virtually impossible, using a conventional desk EQ, to thin out a humbucking guitar sound to produce a single-coil sound, or vice versa. However, I've already said that hearing sounds in isolation is of little help, because they sound so different when everything else is up and running. So if you're recording a band with two guitar players, you should really get them to play together in the studio, so you can see how the sounds work together. This is even more important if you are planning to record the guitars independently. A common mistake made by bands with little studio experience is that they use far too much guitar overdrive. While this may be fine live and for studio solos, two guitars playing wall-to-wall fuzz power chords all the way through a song is a sure recipe for migraine at mix time.
What goes for the guitar goes for the bass too, and you often find that a bass that sounds rich and punchy on its own, just sounds dumpy or soggy when everything else is going. What's not often appreciated is how important the lower mid frequencies are in making a bass sound full, and if you don't believe this, listen to a good record on a small radio or mini hi-fi. You'll get little real bass from such a system, but what's going on in the 80Hz to 150Hz region still provides the illusion of depth. If you can get the bass player to play along with the guitarists or whatever else is in the mix, while you listen to the result from the control room (or from a test recording), you'll get a better idea of how things are going to sound. If you're DI'ing the bass via a DI box, adding compression can significantly improve the sound, but there are occasions on which miking the amp is the only way to get a good sound.
One very important point here is that the sound of an instrument changes very significantly, depending on how it is played. Put two drummers on the same kit and they'll sound quite different. If you hit drums hard, then it sounds as though you're hitting them hard, but you can't make a quiet drummer sound loud, simply by turning him up in the mix -- he'll just sound like he's thwacking a hot water bottle with a wet haddock! If this looks like being a problem, don't be afraid to talk it over with the musicians. Bass guitars also suffer dreadfully if they aren't played with confidence and power. No amount of EQ, compression or anything else will help, so again, have a word with the player before recording.
Once you have the required tracks on tape, you have to set a balance, and the more tracks you have to play with, the more there is to go wrong. If the vocal keeps drifting in level, you should consider compressing it, and in a typical pop mix, I like to hear a good balance between the bass, drums and vocals before I start to add the other instruments.
It's here the musicians' egos can become a problem, because you'll sometimes find there are parts that can be best used very low down in the mix. If the player in question is feeling insecure, however, they may be wondering why they aren't as loud as the others in the band. That's one advantage of working with a sequencer, I suppose -- computers don't have attitude problems. Also, don't rely on the pan controls to keep your sounds separate. Get the mix working in mono first, and only then pan out the various elements. This will also help ensure that your mix is mono-compatible, though you must check this again, after adding any effects.
The other reason for not getting too hung up on EQ'ing sounds until everything is playing, is that not everything can be at the front of the mix -- some things must take a back seat, which may mean they sound slightly dull or insipid when heard on their own. That doesn't matter, though, because they're not going to be heard on their own. Assuming it's an upbeat song or dance track, the vocals will need to be somewhere near the front of the mix, and the rhythm section right behind them, with rhythm guitars and keyboards playing a supporting role. Thinner guitar sounds such as you get from a single-coil pickup leave plenty of space in a mix, and cut through effectively, even when they're quite low in volume -- whereas thicker, humbucking sounds can be overpowering, while still having a tendency to merge in with other sounds.
Before you switch on an effect, ask yourself "Why am I doing this?". Once you've paid out for a good multi-effects unit, there's a temptation to use it wherever possible, but you already know this is wrong. Effects are there to create illusions, and before you can start, you have to decide what illusion you wish to create and why. Most Western music is designed to be heard indoors, so the first illusion to construct is that of an indoor environment, and for this you only need a little reverb. Other than concert halls and bad venues, few listening rooms have huge amounts of reverb: you just need enough to create the effect of your music existing in three-dimensional space. Add too much reverb, and your sound will recede into the distance.
Another consequence of adding reverb to instruments indiscriminately, is that the space in your mix starts to disappear, and as has often been quoted, the spaces in music are as important as the notes. If you don't have a reason to use it, turn it off, and if you do need to use it, use it sparingly. Similarly, if the lower mid section of your mix seems awash with confusion, switch off the reverb returns for a moment, and see how much your use of reverb is contributing to the problem. The same is true of echo effects, and while chorus is less likely to steal your space, it can have the effect of putting the treated instrument 'out of focus'. Chorus creates the illusion of movement, which can smear your stereo imaging, and I find that repetitious modulation can get annoying. By all means use it to round out a pad sound, but if you're about to add it to a guitar, ask yourself if there's a good reason for doing it.
So far, I've covered just a few of the major things that go wrong with mixes. Ultimately, a mix is just a balance of instruments and voices that provide a pleasing result, and the more you process and manipulate the various elements without reason, the less natural the result is likely to sound. It's rather like photography -- put a good subject in front of a camera, make sure the lighting is okay, and you have a respectable picture. If you then start to touch up that picture using something like a computer graphics program, you may be able to change the colour balance, move trees out of the background and give the donkey on the left two heads, but unless you're a real expert (and have the very best tools), it's never going to look as natural as the original. The first rule should always be 'if it ain't broke, don't fix it'.
Perhaps the second rule should relate to the ingredients of the mix, and if an instrument or voice can't justify being there, then why have it? It's rather like cooking a curry: if you throw in the entire contents of your spice rack every time you cook, you'll never get any variety, and the flavours will be confused and overbearing. And the third rule? The third rule is that if you can get a good artistic result by breaking the first two rules, then that's fine too. This talk of curry has made me hungry, so I'm off to phone the take-away now...
If you find that your mixes always sound great in the studio but awful on everybody else's music system, there may be something wrong with your monitoring. However, before jumping to that conclusion, have you made a point of playing other CDs through your monitors, to get an idea of what you should be aiming for? It's a great temptation to monitor everything loudly in the studio, and when you combine this with the excitement of mixing (and possibly with a temporary hearing shift, caused by listening to loud music all day), you may produce a bad mix, even if the monitors are okay. If possible, mix on a different day to recording, or at least take a break.
If things seem to be going badly wrong, first check your speakers are wired in phase, and that they are not mounted too close to corners. If the room is fairly small, you should also be using small-to-medium sized monitors. Anything bigger may just emphasise the room problems. The next step is to play a few CDs over the system to see if they sound okay. If they do, then your monitoring system is at least usable, but if they don't, then look again at your choice of speakers and room. If you have a graphic EQ connected between your mixer and power amplifier, take it out of the circuit -- they rarely help.
Next time you set up what you think is a good mix, take the time to play a similar-style CD, and compare the CD with your own work. Listen particularly for the way the bass end sounds. Many mixes go wrong because of the temptation to add excessive bass to compensate for using small monitors. Also, the nearer you can work to your monitors, the less the room character will interfere with what you hear.
Finally, don't forget to check the mix from the next room with the door left open -- I keep repeating this, but it's the most valuable mixing tip I've learned to date. If there are balance problems, the next room trick will show them up. Just don't ask me why...