The Big Sweep
By Glen Stephan on July 23 2008 03:00 PM | Permalink | Author Info
The "parametric sweep" is one of the most basic and common EQ techniques amongst experienced audio engineers, yet not very well known to those new to the recording and mixing scene. Here's an explanation of the technique and why it should be one of the first techniques any rookie should learn.
This racket we call audio engineering is chock full of axioms, truisms, general rules. Many of them dance around an idea that I personally have never heard expressed directly as such a truism, but I think it ought to be: "half the job of making something sound good it getting rid of what sounds bad." This can apply on a track mix level, where we soon learn that sometimes "less is more" or "the silence is as important as the sound" and that just filling up a mix with a wall of sound from a million tracks is often not the best way to go (so let Phil Spector shoot me; it'll save me from having to look at his hair again.) Or it can apply on a single track level when we learn that we usually "use EQ boost to make things sound different, use EQ cut to make things sound better."
Take that last axiom about EQ boosts vs. EQ cuts, and combine it with another popular axiom about EQ that most of us have heard that says "boost wide, cut narrow", and you wind up with an idea that says "make things sound better by making narrow EQ cuts in it." This is quite the precise description of what the "parametric sweep" is all about. In my book, the parametric sweep is one of the most useful, most powerful, and easiest to perform techniques in the basic engineer's tool kit, and as such, is one go-to technique that should be gone-to quite often. While I would never advocate automatically EQing anything as a matter of routine - EQ should only be applied when our ear or our gear tells us it’s needed – the parametric sweep comes close. The number of tracks in one of my average mixes that winds up not getting a dose of EQ based upon the sweep is usually in the minority.
HOW IT WORKS
Performing a parametric sweep is quite easy. All it requires is your favorite parametric equalizer (hardware or plug-in) and a good set of ears.
Start with all of your parametric EQ bands set for flat or bypass, with the EQ set to work on a single track in your mix (e.g. guitar). Activate one band, setting the bandwidth (Q) fairly large for a fairly tight bandwidth, maybe somewhere in the 1/8th to 1/10th of an octave range. (The “Q” setting for this will vary from EQ to EQ, but will often be somewhere in the Q = 10 area, give or take a couple of Q points.)
Next, set the frequency for that band down at the bottom of the frequency range, probably at or near 20Hz. Then set the gain for that band for high boost, a good 10 – 12 dB. CAUTION: Because you’re about to run a process where there is a boost, however narrow in range, of that many dB, it’s a wise idea to keep the control room monitor volume in check so you don’t blow out any speakers, eardrums, or both when you do this.
Now, with that band set to narrow Q, high boost, slowly sweep through the frequency range; slow enough so that you can listen to the effect it has on the track you are sweeping. Note that if you are using a software plug EQ, there may be some lag or delay involved, making you want to sweep slower than you probably would need to on a hardware EQ where there is no lag between how you turn that button and what you hear.
Every once in a while, you will probably hit a frequency that seems to almost jump out at you as a particularly ugly-sounding, nasally, horn-honking-sounding frequency compared to the adjacent ones. Often this “honker” frequency will be accompanied by a noticeable jump in the meter level.
Staying on that frequency, reverse the gain to a fairly deep cut, maybe 6dB (more or less than that to taste.) We are now notching out that honker frequency and in doing so making that track or instrument sound cleaner, smoother, and overall just plain better. And as an added bonus, we are giving more room at that frequency to other instruments that are not “honking” there.
Repeat this process on the next parametric band, starting where we left off at that first honker frequency. Repeat as necessary for the number of bands of parametric EQ we have. It may not be necessary to use all the bands; a track may have only one or two noticeably offensive honkers. OTOH, you might find more honkers than you have EQ bands. Repeat only as you feel necessary.
BEYOND SINGLE TRACKS
While the parametric sweep may find it’s greatest use in sweetening individual tracks, it has good uses as part of improving the overall mix as well. Just a couple of examples:
We earlier mentioned that parametric notching in one track could make “room” for other tracks at that frequency. This easily leads to one style of “differential EQ”, where we can help two instrument tracks fit together like jigsaw pieces by cutting EQ on one track where we boost EQ on the other. By starting with a parametric sweep, we have already made the cut to one track. If the other track does not have any significant honkers at or near that frequency, a gentle wider-bandwidth boost of a couple of dB to the second track at that same frequency can help define that track as sonically different than the first without making it sound worse, and in fact, making it sound better in the mix.
This technique can be used with any two instruments that are otherwise battling for space in the mix, but is especially sweet when used on two different guitar parts (usually with different guitars or amps or settings, so that they are not honking in the same place to begin with.)
Another very useful example would be when mastering or re-mastering a stereo mix that is just way to shrill in the upper midrange (~2kHz – 6kHz); a fairly common problem with poorer-quality tracking or mixes, or with old mixes taken off of degraded tape, resulting in a somewhat flabby low end and a high end that has rusted away years ago.
Sometimes those upper-mids can be tough to tame without destroying them altogether. Wide-band EQing often scoops too much out, tossing the baby out with the bathwater, and multi-band compression just doesn’t quite scratch the itch because the problem is not do much runaway dynamics as it is sonic saturation.
In such cases, parametric sweeps can be a pretty good solution. By finding honkers in that frequency range (and they will surely be there in force) and surgically knocking them down, we are, in effect, punching holes in the midrange of the mix, rather like holes in Swiss cheese. This has the effect not only of getting rid of the major offenders, but reducing the overall sonic energy of the upper midrange. Yet we are not scooping out too much like with standard EQ, nor are we directly trying to target it by looking for dynamics to compress that are not necessarily there.
There are many more examples and uses for the parametric sweep. Just like voting in Chicago, use it early and use it often and you’ll find yourself with a mix that works like a well-oiled machine.