Filter Audio Effects
Tweaking the Frequency Spectrum
Everything we hear is somewhere on the 20Hz-20kHz audio spectrum, and an equalizer allows us to boost or cut any frequencies in that spectrum. It’s not the ultimate magic solution to everything, but it’s one of the most basic and powerful tools in the mixer’s arsenal.
Google “how does an EQ work,” and you’ll get a lot of articles showing you how to use an EQ. There is one page I’ve found that properly describes what goes on under the hood, written by Ethan Winer (audio journalist, consultant, and founder of RealTraps). Basically, an analog EQ uses capacitors and inductors to delay a signal, then feed it back into the original at a slightly different time. This causes an intentional phase shift that creates comb filtering. Fortunately, this kind of comb filtering is controlled and useful. You never actually hear the phase shift from an EQ, just a boost or cut somewhere in the spectrum. Digital EQs arrive at the same end via a different method (described in the linked article), but the end result is the same.
Equalization was created with spaces in mind (and telephones, but we’ll ignore that). It was a way of overcoming poorly-designed performance venues. If a space naturally accentuates certain frequencies, they are given a corresponding cut; those that aren’t heard as well are given a boost. In this manner, each frequency band is “equalized” with the rest. Today, we use EQs to alter any sound in a DAW or playback space, whether or not the final sound is more “equal” than it started out.
Traditionally, venues use a 1/3-octave, graphic EQ. Every musician knows what an octave is, so it’s a great unit to use for EQs. “1/3-octave” means that each slider on the graphic EQ is that far from one another in the spectrum; it takes 31 bands to cover what we hear. A great sound man will go through each band very quickly, pushing it to one extreme, then adjusting it to where it sounds the best. This can help a little, but the limitations of both the room’s effects on the sound and the fixed nature of each band make this type of EQ far from ideal.
The modern alternative is the parametric EQ. These let you center their effect right where it’s needed, and have an adjustable width (known as Q) as well. Most DAWs come with a 5 to 9-band parametric EQ; here are a few ways to use it.
Take out the Undesirable
Most freshly-recorded sounds have some frequencies that aren’t flattering, at the very least. A lot of these can be taken care of before recording via a great room, nice instruments, a good performance, and mics that are well-matched to the instrument/voice. However, pretty much everything gets EQ’d at some point.
The most common approach to “fixing” audio with EQ is called subtractive EQ. This method makes one or more cuts to parts of the spectrum, without boosting anything. The main reason for this is that boosting any frequency puts the channel in danger of clipping. The predominant technique is to grab a band, make it narrow, boost it about 10 dB, and slowly swipe it across the frequency spectrum. When you hear something ugly, dip the gain into the negative and adjust the width until it’s as good at it gets (it’ll rarely be perfect). You’ll usually want to be pretty surgical about EQing since any extra width is going to affect tone’s you didn’t mean to alter.
Some key places to watch out for (according to Bobby Owsinski) are:
General Use: 200 Hz mud; 300-500Hz boxiness; 800Hz cheap speaker effect; 1-1.5kHz nasality.
Kick drum: Hollowness at 400Hz.
Horns: Piercing at 5kHz.
Vocals: Boomy around 240Hz. Honky at 500Hz. Sibilance somewhere in the 4-7kHz range. Note: these frequencies vary by the singer—e.g. They will usually be higher for a soprano and lower for a bass.
Strings: Scratchy from 7-10kHz.
Accentuate the Positive
Generally, engineers make narrow cuts to clean audio and wide boosts to accent frequencies. This is usually done to make a sound more clear and defined or simply bigger. Wide boosts are also commonly applied to entire instrument groups or mixes to alter the overall sound of a song (brighten it up, thicken the low end, etc . . .). There are many common butter zones for individual instruments (again, we refer to Owsinski’s text).
General Use: 4-6kHz presence and 10-20k air.
Bass Guitar: Attack at 700Hz, snap at 2.5kHz.
Kick Drum: Point at 3-5kHz.
Snare: Fatness at 240-500Hz, attack at 5-7kHz.
Hi-hat/Cymbals: Clang at 200Hz, sparkle at 8-10kHz.
Electric Guitar: Fullness at 240-500Hz, presence at 1.5-2.5kHz.
Acoustic Guitar: Fullness at 80Hz, body at 240Hz, presence at 2-5kHz.
Piano: Fullness at 80Hz, presence at 3-5kHz.
Horns: Fullness at 120Hz.
Voice: Fullness at 120Hz, presence at 5kHz, air at 10-15kHz.
Strings: Fullness at 240Hz.
A Place for Everything, and Everything in its Place
Mixers use EQ on each track in the context of the entire mix. As stated earlier, an effect may sound terrible when a track is heard on its own, but sound great with rest of the song. A couple bands are particularly important.
The low end: It can be easy to have too much low end, resulting in a big, muddy mess. Anything below 60 Hz should be felt more than heard; too much and you get mud. Up to 250Hz is what makes a song thin, fat, or boomy. A bass guitar should go down to 50-80Hz, with a kick drum bottoming out at 80-100 Hz. Most instruments sound better with a low cut of at least 70 Hz (usually higher).
Lead vocals: Wherever the vocals sound the best, everything else should stay away from. Vocal clarity is usually somewhere in the 6-8kHz range. A dip of a few dB’s is usually enough (with a corresponding boost in the vocals and proper panning) to get the voice its own space.
Besides similar instruments covering each other up, similar sounds can also cause phase interference. For example, if an electric bass and a few uprights in an orchestra are in the same song, they may cancel out parts of each other’s sound. In cases like these, it’s best to give each a different spot in the spectrum (or change the arrangement so they play at different times or unrelated figures).