Mixing: When Noise Is Your Friend

Yes, this sounds insane...but try it

By Craig Anderton

Do you want better mixes? Of course you do—the mix, along with mastering, is what makes or breaks your music. Even the best tracks won’t come across if they’re not mixed correctly.

Different people approach mixing differently, but I don’t think anyone has described something as whacked-out as what we’re going to cover in this article. Some people will read this and just shake their heads, but others will actually try the suggested technique, and craft tighter, punchier mixes without any kind of compression or other processing.

THE MIXING PROBLEM

What makes mixing so difficult is, unfortunately, a limitation of the human ear/brain combination. Our hearing can discern very small changes in pitch, but not level. You’ll easily hear a 3% pitch change as being distinctly out of tune, but a 3% level change is nowhere near as dramatic. Also, our ears have an incredibly wide dynamic range—much more than a CD, for example. So when we mix and use only the top 20-40 dB of average available dynamic range, even extreme musical dynamics don’t represent that much of a change for the ear’s total dynamic range.

Another problem with mixing is that the ear’s frequency response changes at different levels. This is why small changes in volume are often perceived as tonal differences, and why it is so important to balance levels exactly when doing A-B comparisons. Because our ears hear low and high end signals better at higher levels, just a slight volume boost might produce a subjective feeling of greater “warmth” (from the additional low end) and “sparkle” (from the increased perception of treble).

The reason why top mixing engineers are in such demand is because through years of practice, they’ve trained their ears to discriminate among tiny level and frequency response differences (and hopefully, taken care of their ears so they don’t suffer from their own frequency response problems). They are basically “juggling” the levels of multiple tracks, making sure that each one occupies its proper level with respect to the other tracks. Remember, a mix doesn’t compare levels to an absolute standard; all the tracks are interrelated. As an obvious example, the lead instruments usually have higher levels than the rhythm instruments. But there are much smaller hierarchies. Suppose you have a string pad part, and the same part delayed a bit to produce chorusing. To avoid having excessive peaking when the signals reach maximum amplitude at the same time, as well as better preserve any rhythmic “groove,” you’ll probably mix the delayed track around 6 dB behind the non-delayed track.

The more tracks, the more intricate this juggling act becomes. However, there are certain essential elements of any mix—some instruments that just have to be there, and mixed fairly closely in level to one another because of their importance. Ensuring that these elements are clearly audible and perfectly balanced is, I believe, one of the most important qualities in creating a “transportable” mix (i.e., one that sounds good over a variety of systems). Perhaps the lovely high end of some bell won’t translate on a $29.95 boombox, but if the average listener can make out the vocals, leads, beat, and bass, you have the high points covered.

Ironically, though, our ears are less sensitive to changes in relatively loud levels than to relatively soft ones. This is why some veteran mixers start work on a mix at low levels, not just to protect their hearing but because it makes it easier to tell if the important instruments are out of balance with respect to each other. At higher levels, differences in balance are harder to detect.

ANOTHER ONE OF THOSE ACCIDENTS

The following mixing technique is a way to check whether a song’s crucial elements are mixed with equal emphasis. Like many other techniques that ultimately turn out to be useful, this one was discovered by accident.

At one point I had a home studio in Florida that didn’t have central air conditioning, and the in-wall air conditioner made a fair amount of background noise. One day, I noticed that the mixes I did when the air conditioner was on often sounded better than the ones I did when it was off. This seemed odd at first, until I made the connection with how many musicians use the “play the music in the car” test as the final arbiter of whether a mix is going to work or not. In both cases the background noise masks low-level signals, making it easier to tell which signals make it above the noise.

Curious whether this phenomenon could be quantized further, I started injecting pink noise (Fig. 1) into the console while mixing.

Fig. 1: Sound Forge can generate a variety of noise types, including pink noise.

This just about forces you to listen at relatively low levels, because the noise is really obnoxious! But more importantly, the noise adds a sort of “cloud cover” over the music, and as mountain peaks poke out of a cloud cover, so do sonic peaks poke out of the noise.

APPLYING THE TECHNIQUE

You’ll want to add in the pink noise very sporadically during a mix, because the noise covers up high frequency sounds like hi-hat. You cannot get an accurate idea of the complete mix while you’re mixing with noise injected into the bus, but what you can do is make sure that all the important instruments are being heard properly. (Similarly, when listening in a car system, road noise will often mask lower frequencies.)

Typically, I’ll take the mix to the point where I’m fairly satisified with the sound. Then I’ll add in lots of noise—no less than 10 dB below 0 with dance mixes, for example, which typically have restricted dynamics anyway—and start analyzing.

While listening through the song, I pay special attention to vocals, snare, kick, bass, and leads (with this much noise, you’re not going to hear much else in the song anway). It’s very easy to adjust their relative levels, because there’s a limited range between overload on the high end, and dropping below the noise on the low end. If all the crucial sounds make it into that window and can be heard clearly above the noise without distorting, you have a head start toward an equal balance.

Also note that the “noise test” can uncover problems. If you can hear a hi-hat or other minor part fairly high above the noise, it’s probably too loud.

I’ll generally run through the song a few more times, carefully tweaking each track for the right relative balance. Then it’s time to take out the noise. First, it’s an incredible relief not to hear that annoying hiss! Second, you can now get to work balancing the supporting instruments so that they work well with the lead sounds you’ve tweaked.

Although so far I’ve only mentioned instruments being above the noise floor, there are actually three distinct zones created by the noise: totally masked by the noise (inaudible), above the noise (clearly audible), and “melded,” where an instrument isn’t loud enough to stand out or soft enough to be masked, so it blends in with the noise. I find that mixing rhythm parts so that they sound melded can work if the noise is adjusted to a level suitable for the rhythm parts.

FADING OUT

Overall, I estimate spending only about 3% of my mixing time using the injected noise, and I don't use it at all for some mixes. But sometimes, espeically with dense mixes, it’s the factor responsible for making the mix sound good over multiple systems. Mixing with noise may sound crazy, but give it a try. With a little practice, there are ways to make noise work for you.

   Craig Anderton is Editor Emeritus of Harmony Central. He has played on, mixed, or produced over 20 major label releases (as well as mastered over a hundred tracks for various musicians), and written over a thousand articles for magazines like Guitar Player, Keyboard, Sound on Sound (UK), and Sound + Recording (Germany). He has also lectured on technology and the arts in 38 states, 10 countries, and three languages.