Audio foundations
EQ vs compression
Both make a mic sound better, and people mix them up constantly, because they are the same operation pointed at different axes. EQ decides what your voice sounds like; compression decides how consistently it sounds like that. This page draws the line, then makes it mathematically exact.
If you want either tool explained from zero, start with what an EQ is or what a compressor is. This page is the comparison: which problem each one actually solves, why running them in a different order changes the sound, and the one idea that makes the whole distinction click.
Two different axes
An EQ reshapes your frequency balance: which pitches in your voice are loud and which are quiet. A compressor reshapes your dynamics: how your loudness changes over time. They do not overlap. EQ cannot fix an uneven performance, and compression cannot fix a boomy tone, so the fastest way to pick the right tool is to name the complaint.
- Reach for EQ when the complaint is a tone word: boomy, muddy, thin, dull, harsh. Those are all statements about frequency. The standard voice moves are a high-pass around 80 to 120 Hz to remove rumble, a gentle cut around 250 to 500 Hz to tame boxiness, a presence lift around 3 to 6 kHz for clarity, and a high shelf around 10 kHz for air.
- Reach for compression when the complaint is a consistency word: quiet words vanish, loud moments spike, the voice gets lost under game audio, the level wanders as you turn your head. Those are all statements about time. A compressor turns down your loud moments so the gap between loud and quiet shrinks, and makeup gain brings the evened-out result back up.
If the recording sounds wrong when you say one sustained "ahh," it is an EQ problem. If each moment sounds fine but the moments do not match each other, it is a compression problem. Most voices eventually want a little of both, which is why the two effects are neighbors in almost every chain.
Why the order changes the sound
Effects run in series, and a compressor does not react to your voice in the abstract. It reacts to whatever signal arrives at its input. That means any EQ placed before the compressor changes the energy the compressor detects, so EQ-then-compressor and compressor-then-EQ sound different even with identical settings on both nodes.
The clearest case is rumble. A 40 Hz air-conditioner hum or desk-bump thump is barely audible as tone, but it is real energy, and the detector counts it. Leave it in and the compressor clamps your voice every time a truck drives past, for no reason your ears can name. A high-pass filter in front of the compressor removes that energy before anything measures it, which is why the standard advice is cleanup EQ first. The same logic applies to a boxiness cut around 250 to 500 Hz: take that energy out first and the detector sees a leaner signal, so the compressor works less and pumps less.
Flip the order and you get the opposite problem. Any EQ boost placed after the compressor is unregulated: a +4 dB presence lift applied downstream can push the very peaks the compressor just tamed right back up. That is not always wrong, some people like a polish EQ after compression precisely because it is untouched, but you should choose it on purpose rather than discover it.
There is also a third option that gets you the detection benefit without reordering anything: the compressor's sidechain filters. The sidechain high-pass (20 Hz to 2 kHz, default 20 Hz, which is bypass) and low-pass (1 kHz to 20 kHz, default 20 kHz, also bypass) filter only what the detector hears, never your actual audio. Raising the sidechain high-pass to 100 Hz makes the compressor deaf to rumble while your full-range voice passes straight through. The broader question of where everything sits is covered in what order DSP effects should go in.
The same operation, aimed differently
Here is the idea that makes the comparison exact. Each EQ band is a biquad filter built from the standard RBJ cookbook formulas, and a filter is nothing more than a gain that depends on frequency. Its magnitude response |H(f)| says, for every frequency, how much gets through:
gain_dB(f) = 10 * log10(|H(f)|^2)Evaluate |H(f)| across the spectrum and you get the familiar EQ curve: +4 dB at the band center, sliding back to 0 dB a couple of octaves away. That curve never moves while you talk. An equalizer is a volume knob that plays favorites.
A compressor is also a gain, but its input is time, not frequency. A detector tracks your level with an envelope follower smoothed by the attack and release times (defaults 10 ms and 100 ms), and the gain is computed from how far that envelope sits above the threshold:
gain_dB(t) = (1/R - 1) * (level_dB(t) - T)At the defaults (threshold -18 dB, ratio 3:1), a peak at -6 dB is 12 dB over the line, so the gain is (1/3 - 1) * 12 = -8 dB and the peak comes out at -14 dB. Twelve dB over becomes four dB over. It is a fader ride performed by something with a ten-millisecond reaction time.
Put the two side by side and the whole EQ-versus-compression question collapses into one line. Both effects do exactly one thing to your signal: multiply it.
y = g * xEQ chooses g by frequency and holds it constant in time; compression chooses g by time and applies it (almost) equally across frequency. Same knob, different axis.
The two gains are perpendicular. An EQ curve is frozen in time and varies across the spectrum; a compressor's gain varies moment to moment and, unless you filter its sidechain, treats the whole spectrum the same. That is why neither can do the other's job, and why stacking more of one never substitutes for the other: no EQ setting makes a mumbled word louder than a shouted one, and no ratio turns a boomy voice bright.
So which one do you need?
For a voice chain, usually both, in a specific order: cleanup EQ first, compressor after the gate, per the canonical chain in the chain-order guide. If you only have energy for one, fix tone before consistency; listeners forgive a slightly uneven level faster than they forgive mud. The step-by-step setups live at how to EQ your microphone and how to set up a compressor on your voice.
In patchd, EQ and compressor are both free nodes in the effects rack on every channel, so you can drag them into either order and hear the difference this page describes. The compressor panel draws its transfer curve from the same formula the engine applies, so the shape on screen is the shape on your voice. patchd is in development now. Join the waitlist and we will tell you the moment it is ready to install.