Without doubt the hardest part of recording is mixing, yet it is also the most enjoyable as this is when everything starts to come together and all the hard work justifies itself. A good mixer paints a picture in sound that attracts the listener and conveys the song clearly and simply. I could sum up a good recording as a series of priorities which are:
The song is set from the start and a good producer will have chosen a song that has 'something to say' and a good mixer will convey that something to the listener.
The singer is the next most important aspect and a good mixer will allow the singer to be heard and the lyrics to be conveyed clearly but with style. There is nothing more annoying than hearing a track and not being able to distinguish the lyric amongst a babble of instrumentation. Fortunately you don't hear recordings like that on commercial radio as they just don't get a look in. Engineers are often guilty of cluttering up tracks with all sorts of tricks and garbage that distract from the song and the singer because they know the song so well after days in the studio that they think everyone hears it like they do. If the track is to have a chance of commercial success it must be understandable from the first hearing. Always underestimate the ability of the listener as they are not professional listeners like you.
The feel or groove is what catches the listeners attention initially and sets up the mood and emotion of the track. This is created by careful balancing of the rhythmic aspects of the track be it drums, percussion or a great guitar groove.
Finally there is the fiddley bits as I call them, they are the musical phrases linking lyrics, joining verses to choruses and filling solo sections etc. that are created by the guitar licks, the piano fills, the answering vocal phases etc.
So where to start?
Monitoring speakers come in two types. Nearfield and Main. I like to use both. I work primarily on the nearfield to establish my balances etc. and then every now and then I will switch it up to the big speakers as they give a better idea of the low frequency balance, plus it sounds good eh! (I was a Yamaha NS10 freak for years but now I'm totally sold on the Event 20/20. Well done Event!) To me a big speaker system is like a magnifying glass, it blows the sound up and you can hear more but for a big system to be really good you have to flush mount them and have good speakers and a good amplifier system. Can I say here that I don't like equalised speaker systems. If they don't sound good flat, get another speaker!!
The first important procedure is to setup your console for mixing. The first requirement is to setup your levels to and from your master recorder, usually a DAT. If your console has an oscillator send tone to the DAT and balance left and right channels. Then check that the return to your console, which is what you'll monitor, is balanced correctly left and right. At this stage it is also recommended that you insert your master compressor either in the master stereo output inserts or inline between the console and the DAT and line up correct left/right balance. This procedure is very important as it effects your level structure from then on and if you don't do it now you can end up with your levels all over the shop later.
Next you must establish your auxiliary sends and returns.
One of the best ways to get perspective and separation within your mix is to what I refer to as "putting everyone in their own space". You can achieve this through the use of reverb and effects. I like to have one reverb unit dedicated to the drums. No other instruments are sent to this effect, only the drums which will put them in their space. The choice of reverb for drums depends entirely on the track but I start by putting reverb on the snare and going through the presets to find the one that works best for the track. I find it usually ends up with a bright reverb of shortish length around 1 - 1.2sec reverb time.
Note: A very fine producer in OZ was once quoted as saying "Give me a studio with 10 Midiverbs over a studio with one Lexicon 224XL" We all know what fantastic units the Lexicons are but if it's all you've got you are limited to only one perspective.
Next I'll dedicate a reverb unit to act as my overall reverb effect. I look for the best (not necessarily most expensive) unit in the studio for this will be my master reverb for vocals etc.
In the example above there are 6 sends with 5 & 6 being an option over 3 & 4. I therefore like to use 1 for my drums and 2 for my master verb. Then I can assign the others for effects. I do this so that I can always add master reverb as well as effects if necessary and if I had used say 3, I couldn't put master verb on channels where the effect was assigned to 5. Should I use a stereo or mono send to the effects?? To be perfectly honest I don't think it matters. Most of the stereo input reverb units I find have a mock stereo input, not a true stereo. If you use two sends it really doesn't make a difference unless you are working with the more expensive units like the aforementioned Lexicon, and even then I question the validity of two inputs especially if you are limited in the number of sends.
I then assign the sends 3 - 6 to additional effects like delay, pitch change etc. to act as perspective enhancers. When establishing delays I set them to the track tempo. See Tempo Chart. The idea is to add these perspective effects so you only just hear them when in solo and they appear to disappear when mixed into the track. Bob Clearmountain - the world famous mixer - always had two delays going, one on eighths and the other on 16ths. It puts an air around instruments and if mixed in correctly you won't actually hear them, just sense them. Pitch change is another effect to consider with say the left channel set to -.008 cents and the right to +.008 cents. This effect is great on harmony vocals and it puts them in a different space form the lead vocal. Finally a soft flange or chorus is another effect I'll have as an option for guitars etc. See Effects pages for settings.
Make sure that all your effects are returned through the effect returns and assigned to the master stereo output. If you are fortunate enough to have spare channels on your desk you can return your delay and chorus type effects back through a console channel as this gives you the option of adding master reverb to them and using the channel EQ. Delays can soften if master reverb is added to their returns plus you can attain your feedback from the console instead of using the control on the effect unit. Say you are using send 3 to a delay unit you can feed back to the delay by sending the send 3 on the return back into the unit. N.B. Incidentally, make sure that the dry/wet or mix controls on your effect units are set to wet as you are only wanting the effect from the units and you won't need any dry sound. (If you are using the Alesis Quadraverb check this as all the default settings have 50% dry and 50% wet.) The returns from effects are usually panned full stereo L/R, but you may wish to bring the drum reverb back half L/R to separate the two.
Your console should now be setup like this
Some mixers start with the drums, others start with the vocal. I must admit I start with the drums as they convey the dynamic of a song. Hopefully you will have automation on your console, if not, you must now start setting up a series of moves and remember where and how they occur because, let's face it, the balance within a mix is not static, it varies continuously throughout a song. For example lets say the drummer plays a rimshot snare through the verses and full snare in the chorus. The EQ required on the rimshot snare sound is probably different from the chorus snare sound so I often split the snare return from the recorder into two console channels so I can EQ and effect each separately and automate the switch between the two. For example, the snare in the chorus will probably require more reverb than the rimshot so having a separate channel allows for that. Automation also allows for the tom mikes to be muted when not needed thus reducing the spill of the rest of the kit and cutting out the constant ringing of the toms which occurs with undampened toms. The overhead mikes also will need to be ridden throughout the track, I tend to lower the overhead mikes when the rimshot is playing to achieve a tighter sound, then I lift them in the chorus when the full snare comes in. Reverb on the overheads gives reverb on the cymbals but it also adds reverb to the snare in the chorus and lifts the whole ambient sound of the kit. This has the effect of changing the perspective of the drums in a mix. You can also change the perspective by putting master reverb on the overheads which blends with the drum reverb.
Once we have achieved a reasonable balance of the kit and the dynamics are set in place we can add the bass. The bass and the kick drum will determine the bottom end of the track so the balance between the kick and bass is critical. The kick will give the bass punch and attack when they hit together.
Note: I must say a few words here about bottom end. The big mistake in mixing is to make the bottom end sound too big by adding lots of bottom end EQ to the kick and the bass. You must bear in mind how the track will be played back by the listener. Nowadays everyone has a stereo system with bass boost as an option either as a loudness switch or as a sub bass control. Everyone who has this option has it switched on!! If you get out a few of your favourite recordings and listen to them on your mixing speakers you will find that they are relatively shy in the bottom end and yet when played through your average boom box sound tight and fat. You have to start to understand what a flat response really means and learn to mix that way. If you put a bass on a VU meter you will notice how much energy there is in the bottom end. A bass peaking to zero will have the same apparent loudness as a highhat peaking to -30db. That's because a hithat has no real bottom end compared with a bass so be careful with your low end EQ on basses and kick drums. I like to solo the two together and EQ them so that they are tight but not boomy.
Add the vocal
OK, so the bass and drums are now at their first mix level so next I will add the vocal and mix it sitting just above the bass and drums. This might mean an EQ change so they all sit tightly together. The vocal might need to be ridden with the automation and I'll probably compress it again to keep the dynamic range within the boundaries of the whole track. I often find that the reverb on the vocal will need to be ridden so that the screaming high notes need more reverb that the quiet intimate sections in the verse. Here I take a feed from the direct out of the vocal channel and bring it up on another channel on the console. I then deselect this channel from the stereo mix output so it goes nowhere but the aux sends are still working. By adding reverb to this channel I can use this channel to ride the reverb on the vocal as an automated send.
Adding the rest
Now we can start to add the fiddley bits like the rhythm guitar and keyboard pads etc. adjusting their balance to fit tightly but not overpowering the vocal. (Please understand I am not defaming guitars etc. by calling them fiddley bits, they are just as important as every other part) The track should now be starting to take shape. If the dynamics of the drums and vocal have been set correctly the placement of the additional instruments will fall into place easily. The vocal harmonies, and solo instruments can now be mixed into the track and we are nearing the completion of the first mixdown.
Note: It is important to keep checking your mix in mono. Unfortunately stereo and mono are not compatible. When you switch to mono, instruments that are panned centre are 3db higher than in stereo so your vocal, kick and snare, for example, will come up in the mix. Some engineers actually make two mixes of a track: One that is full wide stereo with full dynamic range for home listening and one where all the hard left and right signals are panned to the centre or half centre and compressed for radio. It's really hard because if you make a mix sound great on a good home hi-fi it won't have the tightness and punch a mix made for commercial radio will have where the dynamic range is low. It's common practice to make separate mixes of the singles from an album for radio whereas the remaining tracks are mixed totally for home hi-fi. I think you will find that most commercial records are mixed to sound great on FM Radio.
Rest and Recreation
It is important that you constantly give your ears a break during the mixing process as your ears have little compressors in them that will progressively shut your ears down. Have you noticed that when you've been in a loud club with a loud band when you go outside you can't hear as well. It's part of your ears protection system and a cup of coffee in another room watching TV or something will allow them to start opening back up. I like to "mix from the kitchen" as I call it. This means playing the automated mix and listening to it from an adjacent room with the control room door open, you'd be surprised how clearly you can hear the balance between instruments when you get away from the direct sound from your speakers. The relationship between the bass and kick, the balance within the harmonies, the clarity of the vocals etc. all become clearer when you relax and listen from another room.
Unfortunately the human ear is not flat at all levels. Some guys called Fletcher and Munson worked out what the response curve of the ear was and found that at low levels the ear missed out on the low frequencies and the high frequencies, whereas at loud levels it was the opposite.
From the above chart you can see that around 80 - 90db the ear is the flattest. The fact that we don't hear low frequencies and high frequencies at low levels created the Loudness switch on stereo systems which boosts the low and high frequencies to compensate for the ear. Unfortunately, Joe Public doesn't know this but knows that when it is switched in things sound fatter and brighter so they leave it in all the time. It is generally recognised that a level of 85db is where the ear is at it's flattest so don't mix too loud if you want a flat response.
The important thing about mixing is apparent loudness, or relative loudness. If I whisper into a mike and then I shout into a mike the shout will appear louder because I know that shouting is loud. It's the same with mixing. You create an illusion of loudness, everything is relative. You can't get bigger if you are already at your maximum. If I mix a soft acoustic guitar and vocal and peak to zero then bring in a full kit and grunge guitar also peaking to zero it will apparently get louder because I know that drums and guitar are loud. Mixing is the art of making signals that all peak to zero sound as if there is a dynamic range. Nowadays with the excellent compression systems we have most recordings are heavily compressed. I was told of a producer who hired a mixing engineer to mix an album. The guy turned up with racks and racks of compressors and set about compressing every track. He had one compressor for this and another for that etc. In the end the whole mix was pumping away and almost mixed itself. That album went on to sell millions of copies world wide. Those of you who have played with Waves Ultramaximiser will know what compression can do for a mix. If you watch most modern pop recordings on a VU meter the needle is almost static varying only a few db yet the tracks go from quiet intros to full on chorus and solo sections yet still there is only a small variation in level. So setting compression (and limiting) levels is important. I will always have a compressor across the output of my mixes as it helps control the peaks and brings up the loudness of the track but I may use individual compressors on separate channels.
Finally - do take the time to get a good mix. If you don't you have not given justice to all the effort you put into recording it in the first place. It may take a few remixes, so what - it's the final product that counts.