Home   Contents   About   Links

Saturday, December 13, 2008

Mojo Pie presents the secrets of audio mixing

Learn how to put together the perfect mix

There they are... the eight, 16, 24, 48, 96 tracks that you've painstakingly recorded, overdubbed, erased, and recorded some more. So, now what? You can't play with your mixer's knobs every time you want to hear your masterpiece! You need to blend all those tracks to (usually) a two-channel, stereo mix. In other words: "Mixdown your tracks."

Great -- so what does THAT mean? Well, first, you need something to mixdown onto: A computer, a stand-alone hard disk recorder (such as the ubiquitous Alesis Masterlink), a two-channel reel-to-reel, a Hi-Fi VHS machine, a MiniDisc recorder, or even a cassette deck. All are usable options. Some moreso than others. The pros and cons of each format are beyond the scope of this article, but suffice it to say, use the highest quality recorder your budget will allow.

I'm also going to have to presume you already have a decent monitoring system in place. You can't make good sonic decisions if you can't properly hear what your tracks sound like. And it is true: Headphones are not considered a decent monitoring system. They are horribly ineffective as a primary mixing tool.

You also want to pay attention to the volume of your monitors as you mix. There are many different preference engineers take to levels, but it is commonly accepted that 85 dB SPL is where human hearing frequency response is most flat, and this is typically where many engineers leave their levels at. It is good practice to vary listening levels while mixing to get a feel for the mix balance at different volumes. Mixing too loud almost always results in unbalanced mixes, mixing at softer levels usually produces more balanced results.

Get yourself mentally prepared. Rest your ears. Give your body some rest. It's virtually impossible to mix right after a tracking session, or after going to a concert. You need your ears to be in top form for analyzing and objective critical listening. What works for me is starting fresh in the morning when my ears are most rested.

Throw up the faders and see what you have to work with. Yes, ALL of them. What you want to listen for are the tracks that are working, and the tracks that aren't. Yeah, the guitars are fighting during the choruses. The background vocals aren't working midway through the verses. The lead solo starts too early. And, of course, you're taking notes for yourself as you do this.

The plan here is to take out the stuff that's not working together, and leave in the stuff that is working. There are a couple of ways to do this: 1) Use EQ to tailor the portion of sonic spectrum a track will fit in, and 2) Mute the track at the problem spots.

Mute means mute. That's right: You pull the track right out of there. If doesn't help the song it doesn't belong. Mixdown is the time to be blatantly critical of every track that's been put down. You don't reduce the volume, you don't bury it behind something else. Do it and it'll simply result in a messy mix. You can mute either by automation (via software or mixer hardware), or the old-fashioned follow-the-timelog-with-your-hands-on-the-mute-button technique. You don't have to mute the track completely either. You can add interest by pulling it in and out of the mix at key spots (obviously not at the places where it's causing mix difficulties!).

Let's be honest. Sometimes the track works, but is simply frequency-fighting with another track. In this case, you probably don't want to mute the track, but you need to shape it so that it doesn't interfere with another track's piece of the sonic pie. So, reach for the EQ knobs.

First, there are various ways to use EQ and EQ decisions you may have made during tracking will affect the mix process. The addition of EQ into the signal chain always results in some compromise of the waveform by introducing phase-shifts (time-based artifacts that can results in comb-filtering of the waveform) -- especially when boosting frequencies. Cutting EQ results in less of these artifacts, so it is preferable to apply EQ by cutting, rather than boosting, a practice known as subtractive EQ. The quality of the EQ itself also dictates the artifacts -- cheap EQ gear means more artifacts where mastering-grade EQ means significantly less artifacts but at very high prices.

Using EQ to shape sound is a bit of a compromise. It changes the signal, but it introduces a small signal degradation. It is a necessary evil. For line-level instruments such as synths, you certainly can use EQ to shape the sound. For mic'd sources, it's much better to use mic selection and mic placement to get the sound you're after, rather than reaching for the EQ knobs. For example, don't brighten an amp by boosting your high-shelf EQ. Instead, change amp settings, change guitar pickups, change amps, move that mic closer to the center of the cone. If you're not getting the sound you want, maybe you're using the wrong instrument/amp combination.

Another point to keep in mind: Try to get your tracks sounding the way you want during tracking. If the tracks are sounding the way you want them when going to tape, then selecting sounds during overdubs become much easier. And even better, during the mixing phase, you'll find your tracks will blend better (since you've already blended them correctly in the tracking process). Never tell yourself you'll fix it at mixdown. Get off your butt and move a mic, change the mic, change the source, move the source or switch rooms. If none of these work, then reach for the EQ.

So, back to mixdown... I strongly suggest you adopt the subtractive EQ approach - cut instead of boost. If there's too little highs, remove some mids or bass to shape it. This does two things - minimizes phase-related artifacts, and more importantly, reduces unnecessary signal level that will eat into your mixer's headroom since cutting will reduce the amount of frequency a waveform will take up.

To use EQ to shape the audio picture, think of 1) the various frequencies of the tracks, and where they sit; and 2) the placement of the tracks in the soundstage in front of you (between the left and right speakers). The charts below will give you some indication of the frequency ranges for various sound sources that will help guide your use of EQ.

1/3 Octave Frequency Charts *

Audio Octave Ranges
Key Frequencies For Instruments
These charts, however, don't tell you the whole story. Think of a mix as a three-dimensional space in front of you. You have control over the left/right, the high/low, and the front/back of the sound stage. The tools that let you manipulate this area are panning (for left/right positioning), EQ (for high/low positioning), and fader level (for front/back positioning).

Homing in on EQ for the moment, keep in mind that as you shape your tracks, higher-frequency tracks will appear to come from higher up in the monitors than lower-frequency material. This can be useful in positioning guitar tracks. If a guitar track is fighting with something else in the mix, you can move it away from the offending track by removing some bass content in conjunction with panning.

As for panning, it alone can also be used to separate mix elements into distinct positions in the mix. For example, panning a keyboard rhythm part off to one side while panning a complimentary rhythm guitar part to the other will result in a pleasant, wide and full rhythm section whose elements don't interfere with each other. When using panning, it is often helpful to envision a music stage in front of you, and place the tracks within that space as you would normally hear at a concert. You may not keep the tracks in this position as you build-up and further define your mix, but it does make a useful starting point.

Faders allow you to control the level (and thus how close or how far away the source is) of the track in the mix. No tricks here except that you really shouldn't use level to hide a track -- if the track doesn't work, simply mute it. If you want a track in your face, then bring up the level. Lower the level to move it away from you.

An extremely important point to remember is to maintain your relative levels so that you don't eat away at your mixer's headroom -- if you've set all the levels of the tracks and find you have to push your solo track level very high for it to cut through, then you've got all your other track's faders set far too high. Most mixers will start sounding pretty harsh if pushed too hard.

What about reverbs, delays, and all those cool other effects? Effects play an important role in mixing - much like spices do in fine cuisine. It's all in the way you use them. The topic merits its own full-blown article (maybe in the future), but for the moment, here's a brief overview.

My own approach is that unless a particular effect is an integral part of a track's sound (such as a chorus, or wah on a guitar part), I leave all effects until the mix stage. The tracks really should stand on their own merit without any effects added-in, and then you make the mix bigger, fatter, wider, with careful and judicious use of things like reverb, delay, and chorus. Little touches such as timing delay and chorus settings to the tempo of a track really do a lot to make the tracks shine. Overuse of effects usually results in muddy, poorly-defined mixes. Like with EQ, less is more with effects most times.

Pay particular attention to reverb used for vocals. For most listeners, vocals are the component of the song that reach the people first, and poor effects, or bad EQ is immediately noticeable -- it's not uncommon spending hours on reverb selection for the lead vocal track.

One effect that is very misunderstood (and often poorly used) is compression. This effect can be an important part of a polished sound both during the tracking AND mixing stages of the production process if used properly. A colleague of mine wrote an excellent article (with examples) on the use of compression. Check out Moshe Wohl's article on compression for some great notes on the subject.

Now, it's time to print your mix. You might be working with software. The software will render a data file of your mix. Or, you might be working the old fashioned way with a console. If so, you might think of writing a script telling you when to do what. You might even need help from a friend because an extra pair of hands can be of use.

After you have printed your mix, then the time comes to reference your mix on different playback systems. Do you have a boombox? Do you have a car stereo? Do you have a hi-fi? You'll have to learn by experience how your mixes translate to different listening systems. The goal is for your mix to sound reasonably good on most systems. The only way to learn how to do this is by trial, error and experience.

Take heart. As you learn mix translation with your system, it does become easier. With enough practice, you'll develop a feeling for what works and what doesn't.

The art of mixing is very broad... There are various mixing guidelines for all the many styles of music (you don't mix a big band song the way you would a pop song). As well, within each mix style there are commonly accepted practices. What I've tried to describe in this article was to provide an overview of some basic mixing practices that are common to all mixing styles, as well as some functional tips regarding one of the most abused mixing practices. The best way to approach mixing is to read as much as you can about common approaches, while at the same time, practicing the techniques you've read about and adapting them to suit your own needs.

Always remember: Music is art. Art is subjective. And, your music is your art. Make it sound good to your ears first, then worry about everyone else's.

I want to recommend the following reading:
"Modern Recording Techniques" by Huber and Runstein;
"Art of Digital Audio, Third Edition" by John Watkinson;
"Project Studios: A More Professional Approach" by Phillip Newell
"The Mixing Engineer's Handbook" by Bobby Owsinski
"Behind the Glass" by Howard Massey

* Frequency Charts taken from Modern Recording Techniques, by Huber & Runstein, copyright 1995 Sams Publishing.

--Bruce Valeriani, Blue Bear Sound

No comments: