EQ is usually the first tool people learn and the last one they really understand. On screen it looks simple: a line with a few points you drag up and down. In practice, a lot of mixes that sound "almost there" are being held back by EQ moves that made sense in solo and make no sense in the song.
This is the approach I come back to in almost every session. It isn't the only way to work, but it's dependable, and it keeps you away from the mistakes that cost the most time.
What an EQ actually does
An equalizer turns specific frequency ranges up or down. Nothing more. It can't create content that was never recorded. If a vocal was captured with a dull microphone and has almost nothing above 10 kHz, a big boost up there mostly brings up hiss and room noise. When I catch myself reaching for +8 dB, it's usually a sign the problem is somewhere else: the recording, the arrangement or the sound I picked.
Almost every parametric EQ, from the one that came with your DAW to the expensive ones, gives you the same band types:
- High-pass filter (low cut): removes everything below a chosen frequency.
- Low-pass filter (high cut): removes everything above a chosen frequency.
- Bell (peak): boosts or cuts around a center frequency. The Q control sets how wide that area is.
- Shelf: raises or lowers everything above a point (high shelf) or below it (low shelf), like a tilted step.
Filter slopes: what 12, 24 and 48 dB/octave mean
The slope tells you how fast a filter attenuates once you pass the cutoff frequency. Each filter "order" adds 6 dB per octave. A 12 dB/oct high-pass at 100 Hz is about 12 dB down at 50 Hz and 24 dB down at 25 Hz. A 48 dB/oct filter at the same frequency is already 48 dB down at 50 Hz, close to a wall.
| Slope | Filter order | Character | Where I use it |
|---|---|---|---|
| 6 dB/oct | 1st order | Very gentle, barely noticeable | Subtle tone shaping, taking the edge off a source |
| 12 dB/oct | 2nd order | Smooth and musical | Vocals, guitars, most everyday cleanup |
| 18–24 dB/oct | 3rd–4th order | Firm, clearly defined | Busy mixes, keeping pads and keys out of the low end |
| 36–48 dB/oct | 6th–8th order | Brickwall-like, can ring near the cutoff | Removing sub rumble on the master, surgical separation |
Steeper isn't automatically better. In a minimum phase EQ, steeper slopes create more phase shift and a small resonant bump right at the cutoff. On a kick or bass, that can change the punch more than you expect.
Start with the high-pass filter, but don't overdo it
Most recorded sounds carry energy below the range where they're musically useful: mic stand rumble, air conditioning, someone tapping a foot, low room tone. On one track it's barely noticeable. Across twenty tracks it piles up into a cloudy low end that eats headroom and hides the kick and bass.
Reasonable starting points:
- Lead vocal: 70–100 Hz, 12 dB/oct (lower for deep voices)
- Acoustic and electric guitars: 80–120 Hz
- Pads and keys in a dense mix: 100–200 Hz
- Hi-hats and cymbals: 200–400 Hz, sometimes higher
- Master bus (sub rumble only): 20–30 Hz, steep slope
The trap is high-passing everything at 200 Hz because a tutorial said it "cleans the mix". Do it on every track and the song loses its body. A better habit: with the whole mix playing, sweep the filter up until you hear the sound get thinner, then back it off a little.
Cut first, boost second
Cuts are more forgiving than boosts. A narrow boost sounds unnatural very quickly, while a narrow cut on a problem frequency often goes unnoticed, except that the sound gets cleaner. So when something sounds wrong, I first ask what is too loud before asking what's missing.
The classic way to find a problem frequency is boost-and-sweep: create a narrow bell (Q around 6–8), boost it by 8–10 dB, and slowly sweep through the range. When a ring or honk jumps out, turn the boost into a cut of 2–5 dB.
Studio Tip
Everything sounds bad when you boost a narrow band by 10 dB. If you cut every frequency that sounds ugly during the sweep, you end up with a thin track full of notches. Only cut what bothered you before you started sweeping.
The frequency map I keep in my head
These ranges overlap and every source is different, but a mental map speeds things up a lot.
| Range | What lives there | Too much sounds | Too little sounds |
|---|---|---|---|
| Sub bass, 20–60 Hz | 808 and kick weight, lowest bass notes | Rumbling, wasted headroom | Weak on club and car systems |
| Bass, 60–250 Hz | Kick body, bass, low piano, vocal chest | Boomy, bloated | Thin, small |
| Low mids, 200–400 Hz | Warmth, but also mud | Muddy, cloudy | Cold, hollow |
| Mids, 400 Hz–1.5 kHz | Core of vocals, guitars, snare body | Boxy, honky, nasal | Scooped, distant |
| Upper mids, 2.5–4 kHz | Presence, attack, speech clarity | Harsh, tiring | Dull, buried |
| Highs, 5–10 kHz | Sibilance, cymbals, definition | Sharp, sibilant | Dark, muffled |
| Air, 10–20 kHz | Breath, shimmer, space | Brittle, fizzy | Closed-in |
Mud removal: the 200–400 Hz zone
Close-miked vocals, acoustic guitars, kick drums and most synth patches stack energy between 200 and 400 Hz. Each one sounds warm alone. Together they make the mix sound like it's playing through a blanket.
I rarely fix mud on the master. I find the two or three tracks contributing most and take 2–4 dB out of each with a medium Q (around 1.5–2). Spread across several tracks, small cuts clear the mix without making any single sound thin.
Presence and harshness: 2.5–4 kHz
Our ears are most sensitive in this range, which is why it controls how "forward" a sound feels. A little boost around 3 kHz brings a vocal toward you. Too much and the whole mix becomes tiring at high volume.
Harshness here is often program-dependent: a singer is fine in the verse and piercing when they push in the chorus. A static cut fixes the chorus and dulls the verse. That's a job for a dynamic EQ band, which only cuts when the range crosses a threshold, like a compressor working on one frequency area.
The air band
A high shelf from around 10–12 kHz, boosting 1–3 dB, adds openness to vocals, acoustic guitars and full mixes. Shelves sound smoother than bells up here. Keep an eye on sibilance and noise: an air boost raises "s" sounds and hiss as well as shimmer, so de-ess first, then add air.
Q factor math, without the headache
Q describes how wide a bell is relative to its center frequency:
Q = center frequency ÷ bandwidth (Hz)
A bell at 1,000 Hz with a bandwidth of 500 Hz has a Q of 2. Because hearing works on a logarithmic scale, it's often easier to think in octaves. The conversion is Q = √(2N) ÷ (2N − 1), where N is the bandwidth in octaves:
| Bandwidth | Approximate Q | Typical use |
|---|---|---|
| 2 octaves | 0.67 | Broad tonal moves, mastering |
| 1 octave | 1.41 | General shaping, mud cuts |
| 1/2 octave | 2.87 | Focused cuts on a problem area |
| 1/3 octave | 4.32 | Resonances, boxiness |
| 1/6 octave | 8.65 | Surgical notches, ringing drums |
For example, a cut at 3 kHz with Q 4.32 affects roughly 2.67 to 3.37 kHz at its −3 dB points. My rule of thumb: wide boosts, narrow cuts. Broad boosts sound like a change in tone; narrow boosts sound like a whistle.
Note that not every EQ defines Q the same way, and some change the effective width as you add gain (so-called proportional Q). Trust your ears over the number on the screen.
Minimum phase vs linear phase EQ
A normal EQ is minimum phase. When it changes the level of a frequency range, it also shifts the timing (phase) of frequencies around that band. Usually that's inaudible, and it's part of why analog-style EQs sound natural. The catch: if you EQ one copy of a signal and blend it with an un-EQ'd copy, those phase shifts can cause cancellation.
A linear phase EQ delays all frequencies by the same amount, so phase relationships stay intact. The costs:
- Latency: often tens of milliseconds or more, which makes it unusable while recording.
- Pre-ringing: the filter's energy is spread symmetrically in time, so a faint "swoosh" can appear before a sharp transient. It's most audible with steep or narrow moves in the low frequencies, where the filter kernel gets long.
- CPU: noticeably heavier than a regular EQ.
I use minimum phase for normal mixing, and linear phase in mastering or on parallel and multi-mic setups where the copies must stay aligned.
Decide in context, not in solo
Solo is good for finding problems and bad for judging whether a sound is right. An acoustic guitar with its low end rolled off sounds thin alone, but in a dense arrangement it may be exactly what lets the vocal and bass breathe. My routine: solo briefly to locate the issue, make the move, then un-solo and judge with everything playing. If the change only sounds better in solo, I undo it.
Signal Chain Example: EQ on a busy pop mix
- Vocal: HPF 80 Hz 12 dB/oct → −3 dB at 300 Hz, Q 1.5 → dynamic cut at 3.2 kHz, Q 3 → high shelf +2 dB at 11 kHz
- Acoustic guitar: HPF 120 Hz → −3 dB at 250 Hz → −2 dB at 3 kHz to make room for the vocal
- Pad: HPF 180 Hz 24 dB/oct → −2 dB at 2.5–4 kHz, wide Q
- Master: HPF 25 Hz 48 dB/oct, nothing else
Mistakes I see (and made) all the time
- Judging boosts without matching levels. Louder sounds better for a few seconds. If a move raises the level by 2 dB, you'll like it even if the tone got worse. Match the output gain, then bypass and compare.
- Using too many bands. If a track needs nine bands, the source is probably wrong. Try another sample, take or mic position.
- Fixing arrangement problems with EQ. Two instruments playing the same rhythm in the same octave will always fight. Sometimes the best EQ is muting one or moving it up an octave.
- Chasing the analyzer. A spectrum analyzer is a good second opinion, but a mix isn't supposed to look like a smooth line.
- Not trusting, or trusting too much, your room. If your room has a bass bump at 80 Hz, you'll cut low end that isn't really there. The acoustics and monitoring guide explains why.
A simple EQ workflow
- Set a rough balance with faders and panning before touching EQ.
- Add high-pass filters where they help, checking in context.
- Remove obvious problems: resonances, mud, harshness. Small, targeted cuts.
- Shape the tone with broad, gentle moves.
- Level-match and bypass. Keep the change only if it's actually better.
- Revisit after compression, since compressors change how a sound's tone comes across.
If you take one thing from this page, make it this: EQ is for helping sounds fit together, not for making each one impressive on its own. A mix full of modest, well-placed cuts usually beats a mix full of big boosts.