Mixing Fundamentals: EQ, Compression, and Space

Here's the thing about mixing: if synthesis is about creating sounds from scratch and Foley is about performing them physically, then mixing is the craft of making them all coexist in the same imaginary space. You've got dozens of individual audio tracks — drums, bass, vocals, strings, effects — all fighting for attention, and your job is to balance them so each one stays audible and contributes to the whole without stepping on everyone else's toes.

There are three tools that, between them, account for nearly everything that happens in a professional mix: equalization (EQ), compression, and spatial effects like reverb and delay. I'm not talking about turning knobs until it sounds better — I mean understanding why these tools do what they do. Once that clicks, your mixing and your listening both transform completely.

graph TD
    A["The Three Pillars of Mixing"] --> B["EQ<br/>Sculpting Frequency"]
    A --> C["Compression<br/>Controlling Dynamics"]
    A --> D["Spatial Effects<br/>Placing in Space"]
    
    B --> B1["Removes masking"]
    B --> B2["Creates separation"]
    
    C --> C1["Evens out levels"]
    C --> C2["Shapes character"]
    
    D --> D1["Adds depth"]
    D --> D2["Creates realism"]

Equalization: Sculpting Frequency

Equalization (EQ)^[1] is, at its core, just a set of frequency-specific volume controls. You're adjusting how much of each frequency range is in your signal. The basic question it answers: do you have too much or too little of certain frequencies, and what are you going to do about it?

Why Frequency Matters in Mixing

Here's the problem: different instruments occupy different regions of the frequency spectrum. A kick drum lives in the bass. A snare thrives in the midrange. Cymbals dominate the highs. When you layer all of them together, they start crowding each other — a phenomenon called frequency masking. Your kick and your bass guitar both have energy around 80–100 Hz? They're fighting. Turn both of them up and you get mud instead of punch. The solution is usually to cut a little from one so the other has room to breathe — this is what mixing engineers mean by "carving out space" or "EQ for separation."

Think of the frequency spectrum as a physical stage. The kick drum gets center stage at low frequencies. The snare sits in the upper-midrange. Cymbals live in the highs. The vocal finds its sweet spot where our ears are most sensitive — around 2–4 kHz. If you try to cram everything into the same frequency zones, they obscure each other. Smart EQ moves each sound to its own "frequency real estate," and suddenly the mix sounds wider, clearer, more three-dimensional. Each element can breathe.

EQ Filter Types and Their Uses

There are several types of EQ filters, each with different character and purpose:

High-pass filter (HPF) — also called a low-cut filter: Allows high frequencies through and cuts everything below a certain point. (Yes, the naming is counterintuitive, but think of it this way: a high-pass filter passes highs; a low-cut cuts lows. Same thing, different way of describing it.) This one gets used constantly to remove rumble, low-frequency room noise, and unnecessary bass bloat from tracks that don't need it.

Example: A background ambient drone might get a high-pass at 80 Hz just to scrub away any low rumble that would muddy the whole mix. An acoustic guitar in a pop song gets high-passed around 60 Hz — the guitar barely produces anything useful down there anyway, so all you're removing is noise. Dialogue and piano recordings usually benefit from a gentle high-pass somewhere around 80–100 Hz.

Low-pass filter (LPF) — also called a high-cut filter: The flip side — lets the lows through, removes the highs above a cutoff point. Use this when you need something to sound warmer, duller, or when you're trying to kill high-frequency hiss and noise.

Example: A vintage electric bass gets a low-pass around 8 kHz to smooth out the harsh high-frequency artifacts and tape hiss, giving it that warmer, fuller character. A contact mic on a guitar string picks up finger squeak and handling noise — a low-pass at 10 kHz removes the distractions while keeping the tone intact.

Shelf filters: These boost or cut all frequencies above (high shelf) or below (low shelf) a specific point. They're gentler than other filters — they create a slope rather than a sharp cutoff. Use them for broad tonal shaping: adding "air" to a vocal with a high shelf boost around 10 kHz, or adding "weight" to bass with a low shelf around 60 Hz.

Example: A rock vocal needs presence — that crispy, in-your-face quality — so you boost a high shelf around 4–6 kHz to cut through a dense mix. A kick drum gets a low shelf boost at 100 Hz to add that visceral "thump" you feel in your chest more than hear.

Parametric EQ bands: The most surgical, powerful type. Each band gives you three controls:

• Frequency — which frequency you're targeting (measured in Hz)
• Gain — how much to boost (+ dB) or cut (− dB)
• Q (bandwidth) — how narrow or broad the affected range is. A high Q number (like 10) affects only a thin slice of spectrum; a low Q number (like 0.5) affects a wider range.

A narrow Q cut is perfect for removing a specific nasty resonance. A broad Q boost adds warmth or brightness across a whole range.

Example: A snare drum has a harsh, cheap-sounding resonance at 250 Hz. Instead of low-passing the entire snare (which kills the snap), you'd use a narrow parametric cut right at that 250 Hz peak. To add warmth to a harsh vocal, you might boost a broad band centered around 500 Hz — that's where the voice lives warmth-wise.

Practical EQ Strategy: The Workflow

A seasoned mix engineer doesn't randomly twist knobs. There's a system:

1. Identify the problem: Listen critically. Is it too bright (harsh sibilants), too dark (dull), too boomy (low mud)? Does it mask another track? Which frequency zone is the culprit?
2. Use narrow Q to pinpoint the frequency: Set a narrow Q and sweep the parametric band across the spectrum until you isolate the problem frequency. It's usually not where you'd guess.
3. Cut, don't boost: Most professionals default to cutting frequencies rather than boosting. A 3 dB cut at the problem frequency usually sounds better than a 6 dB boost somewhere else. Cutting is subtractive; boosting adds energy and raises the noise floor.
4. Broaden the Q: Once you've found the exact frequency, widen the Q a bit so the cut/boost sounds natural rather than clinical. A Q that's too narrow can introduce audible artifacts that sound artificial.
5. Use broad shelf filters for overall tone: After making surgical cuts, add a high-shelf boost at 8 kHz with a wide Q to add overall "air" or presence to the track.

EQ Matching and Continuity

Here's a technique used all the time in film mixing called EQ matching — processing one audio source to match the frequency character of another, or to sound like it was recorded in the same acoustic environment. This becomes critical for ADR (Additional Dialogue Recording), where an actor re-records their lines in a studio but needs them to sound like they were captured on the original film set.

Say you have a wide shot where dialogue was captured on location with all its ambient noise, and then a close-up where the actor re-recorded their lines in a clean studio booth. The studio dialogue sounds too clean, too bright, too obviously fake compared to the on-location recording, which has low-end rumble and a duller high-end. EQ matching uses a parametric EQ to adjust the re-recorded dialogue — reduce the high-end brightness, add a touch of low rumble (or at least don't cut the lows so aggressively) — so it feels like the same room. It's painstaking, meticulous work, but it's essential for invisible editing in film. The audience shouldn't know the dialogue was re-recorded.

Compression: Taming Dynamics

Dynamic range compression^[2] might be the most powerful and least understood tool in audio production. It's also one of the most misused — partly because it does wildly different things depending on how you set it up, and partly because its effect can be subtle yet absolutely transformative.

What Compression Does

At its simplest, compression reduces the dynamic range of an audio signal — the loud parts get quieter, so you can turn the overall level up and bring the quiet parts with it. The result is more consistent, controlled audio. But that's just the mechanical explanation. The real magic is that compression fundamentally changes how a sound feels.

Here's a metaphor that works: imagine having a conversation with someone whose volume is all over the place — sometimes barely audible, sometimes nearly shouting. A compressor is like a sound engineer with their hand on the fader, automatically turning it down when they shout and up when they whisper, so the overall level stays consistent and actually intelligible. The compressor does this automatically, continuously, with parameters you can fine-tune.

Understanding Compressor Parameters

Threshold: The level (measured in dB) where the compressor wakes up and starts working. Signals above the threshold get compressed; signals below it pass through untouched.

Example: Set threshold to −20 dB. A vocal peaks at −8 dB, so the compressor engages. But if that vocal never exceeds −20 dB, the compressor doesn't touch it. This gives you precision: compress only the loudest peaks and leave the quiet parts alone.

Ratio: How much the compressor turns down the signal above the threshold. A 4:1 ratio means for every 4 dB the input rises above threshold, the output only rises 1 dB. A 2:1 is gentler; 8:1 is aggressive. An ∞:1 ratio (infinite compression) means the output never exceeds the threshold — that's limiting, a safety tool.

Example: A vocal at −20 dB threshold with a 4:1 ratio: if the vocal peaks at −4 dB (16 dB above threshold), the compressor outputs −4 dB − (16 ÷ 4) = −10 dB. The peak is tamed.

Attack: How fast (measured in milliseconds) the compressor responds when the signal crosses the threshold. A fast attack (1 ms) clamps down almost instantly; a slow attack (50 ms) lets a brief transient slip through before engaging.

This is where compression becomes artful. A slow attack on drums lets the sharp initial snap of the snare through before the compressor engages — preserving the punchy impact while controlling the body and sustain. A fast attack on bass gives you even compression with no peaks, but sometimes it kills the punchiness.

Release: How fast (after the signal drops below threshold) the compressor lets go and returns to no compression. A fast release (10 ms) springs back immediately; a slow release (500 ms or longer) gradually eases off, creating a "pumping" or "breathing" effect.

Example: Compressed drums with fast attack and slow release create that classic "squeeze" sound — the transient punches through, the compressor clamps down on the sustain, and the room ambience swells as it releases. That sound defined countless classic rock and pop records.

Knee: How smoothly the transition happens between uncompressed and compressed. A "soft knee" compressor starts compressing gradually as you approach the threshold, creating a smooth, natural transition. A "hard knee" flips the switch abruptly right at the threshold — more noticeable, sometimes more aggressive.

Makeup gain: Because compression makes loud parts quieter, the overall level drops. Makeup gain turns the whole signal back up afterward — now the quiet parts are louder than before, while the peaks stay controlled. Good compressors show you input and output levels so you can dial in makeup gain and keep things balanced.

How Compression Shapes Character

Beyond just keeping levels consistent, compression shapes the personality of a sound — this is where it stops being functional and becomes creative:

• Heavy compression on drums creates that explosive, larger-than-life sound you hear on classic rock records. The compressor's release causes the room ambience to "pump" after each hit, adding energy and gluing the kit together.
• Parallel compression (blending heavily compressed audio with the uncompressed original, sometimes called "New York compression") adds thickness and sustain without losing transient sharpness. The uncompressed version keeps the punch; the compressed version adds density. Essential in modern pop and hip-hop.
• Light compression on vocals (2:1 ratio, medium attack) smooths out the inconsistencies in a vocal performance — a singer who leans away from the mic or leans in — while keeping the vocal sounding dynamic and emotional.
• Serial compression (chaining multiple compressors with different settings) uses each stage for a different job: the first controls peaks, the second adds glue and character.

Compression in Different Contexts

In film, compression on dialogue is essential — it ensures intelligibility whether the audience is listening in a massive cinema with a $300,000 speaker system or on a phone speaker. A compressed vocal stays balanced and clear across all those very different playback systems.

In music production, compression is structural. The bass and kick drum need to lock together tightly. Compression on both ensures they don't pull apart dynamically, creating a solid, reliable rhythmic foundation.

In game audio, compression on footsteps and impacts makes sure they're always audible and punchy, whether the player is in a quiet library or in the middle of an explosion. Without it, subtle foley cues disappear into the chaos.

Reverb and Delay: Placing Sounds in Space

The Natural Acoustic Space

Every sound you hear in the real world comes wrapped in information about the space it's in — reflections bouncing off walls, floors, ceilings that tell your brain about room size, the materials it's made of, how far away you are from the source. These reflections are reverberation^[3]. Our brains have evolved to decode reverb instantly, subconsciously. A voice in a small bathroom sounds fundamentally different from the same voice in a cathedral — the voice itself is identical, but the reverb is completely different.

Professional recording studios are designed to be acoustically "dead" — walls covered with absorptive material (foam, bass traps, diffusers) that kill reflections. The result is a squeaky-clean recording of just the voice itself, with almost no room information. This is entirely intentional: you want to add artificial reverb in post-production because that gives you complete control over the apparent acoustic environment.

Imagine you recorded dialogue in a small office and it already has that room's character baked in. Later, you need that same dialogue to sound like it was recorded in a courthouse, a subway station, a cathedral. Too bad — you're stuck with the office. So professionals record completely dry, then add whatever space they want afterward. Same with music: most professional vocals are recorded in treated booths with minimal reflections, then artificial reverb is added to taste.

Types of Reverb Processing

Reverb plug-ins work in different ways, each with different sonic strengths:

Algorithmic reverb: Uses mathematical algorithms to simulate how reflections behave in a space. Computationally light compared to other options, and very flexible — you can dial in any room size, decay time, and character you want. Older algorithmic reverbs sometimes betray themselves with a telltale "fingerprint" that sounds a bit artificial. Modern ones (from companies like Fabfilter or Native Instruments) are genuinely convincing.

Convolution reverb: Uses a recorded "impulse response" (IR) — a captured snapshot of how a real room sounds. You play a very brief, loud sound (like a click or gunshot) and record how the room responds to it, then use that measurement to process dry audio. The result is extraordinarily realistic because you're literally using actual measurements of actual spaces. You can take a dry vocal and place it in the Sydney Opera House, a Viennese cathedral, a parking garage, and it sounds like the voice was genuinely recorded there. The tradeoff: convolution reverbs demand serious CPU power and they're less flexible — you can adjust decay time but you can't fundamentally reshape the character the way you can with algorithmic reverb.

graph LR
    A["Reverb Type"] --> B["Algorithmic"]
    A --> C["Convolution"]
    A --> D["Hardware Emulation"]
    
    B --> B1["Flexible<br/>Computationally light"]
    B --> B2["Can sound artificial"]
    
    C --> C1["Extremely realistic<br/>Real space captured"]
    C --> C2["CPU-intensive<br/>Less flexible"]
    
    D --> D1["Distinctive character<br/>Musical coloration"]
    D --> D2["Vintage warmth"]

Plate and spring reverb emulations: These simulate old hardware reverb units that used physical plates or springs vibrated by transducers. A plate reverb uses a large thin metal sheet suspended in a frame; a transducer vibrates it, the vibrations spread across the plate, and a microphone picks them up. Spring reverbs use coiled springs with that characteristic metallic "boing." Both have very distinctive tonal colors — plates are smooth and bright with a sheen, springs have a shimmer — and they're loved for being musical rather than accurate. Classic records from 1960s rock to modern pop used plate and spring reverbs, so their sound has become iconic. Most DAWs include emulations because they add such distinctive character.

Key Reverb Parameters

These parameters give you control over how the space feels:

• Pre-delay: A short gap (in milliseconds) before the reverb actually begins. In a real room, you hear the direct sound first, then reflections a few milliseconds later. Pre-delay recreates this critical spacing, and it's crucial for clarity — zero pre-delay smears the direct sound and its reflections together, muddying the source and making dialogue hard to understand. A pre-delay of 20–40 ms often makes reverb sound much more intelligible. In a real cathedral, pre-delay might be 80–100 ms because of the actual physical distance before reflections bounce back.

• Decay time (RT60): How long the reverb takes to fade completely. A cathedral: several seconds (4–6 or more). A small room: under 0.5 seconds. A shoebox recording booth: 0.2 seconds or less. This single parameter communicates the perceived size of the space better than anything else.

• Wet/Dry ratio (or just "mix"): How much of the signal is processed (wet) versus original (dry). 100% wet is pure reverb with no original signal. 100% dry is no reverb. 50% wet is half and half. In a send-and-return setup, you often use very high wet ratios because you're blending the return with the original dry track. On an insert plug-in, you might use 15–30% wet to add space without drowning the source.

• Diffusion (if available): Controls how quickly early reflections scatter. High diffusion creates a smooth, diffuse reverb; low diffusion creates discrete echoes within the tail.

• Color/EQ: Most good reverbs include built-in EQ. In a real room, high frequencies get absorbed faster than lows, so the reflections naturally get duller. A reverb that's too bright sounds unnatural. Darkening the tail slightly with the built-in EQ usually improves realism.

Reverb in Context: Storytelling Through Space

In film, reverb is an invisible storytelling tool. A close-up in a tight, dead room cuts to a wide shot with ambient reverb — your brain registers "they moved to a larger space." In horror, sudden over-reverberation disorients the viewer. A living room scene shouldn't sound like a cathedral — it would feel wrong and pull you out of the story. But a dramatic monologue in a stone church needs that long, mournful reverb tail. It's emotionally appropriate and enhances the mood.

In music, reverb creates cohesion. A completely dry mix sounds like isolated samples. Add subtle reverb to all tracks (usually via a single reverb send that multiple tracks feed into) and suddenly every instrument feels like it's in the same acoustic space, playing together. Without it, even subtle stereo separation can feel artificial and wrong.

Delay: The Discrete Echo

Delay is related to reverb but fundamentally different. Reverb is many reflections blending together; delay is one or more distinct repetitions of the sound. A delay repeats the input after a specified amount of time.

Tempo-synced delay (or rhythmic delay) locks to the song's tempo. A quarter-note delay syncs the echoes to the quarter-note beat of the music. If the song is 120 BPM, a quarter-note delay is 500 milliseconds. An eighth-note delay is 250 ms. This is why delay is so essential in music production — the echoes feel rhythmically natural, not like a mistake.

Example: A vocal gets a quarter-note delay set to 40% wet on the right channel. Every time the vocalist sings a line, an echo repeats a quarter-note later on the opposite side of the stereo image. It's musical, not distracting, and adds spaciousness and movement.

Fixed-time delay isn't synced to tempo — it repeats after a fixed number of milliseconds. Use it for slapback echo effects (a single repetition 80–150 ms later, classic rock and roll) or to recreate acoustic reflections from a real room.

In film, delay gets used more subtly — maybe just 1–2 repetitions at low level to add space or distance, or to create special effects. A character's voice might get progressively more delayed as they walk away from camera, simulating acoustic distance.

The Mixing Strategy: When to Use Each

• Reverb for general spatial atmosphere — making everything feel like it's in the same space
• Delay for movement and rhythm — adding interest and dimension with repeating echoes
• Parallel reverb/delay to add effects without losing clarity — blend the effect return with dry signal rather than printing the effect destructively onto the track itself

The best mixes use reverb and delay subtly. Beginning mixers tend to over-reverb everything, turning the mix into an echoey, distant, muddy mess. Professional mixes usually use just 10–20% wet reverb on individual tracks, balanced carefully. The space should support the mix, not dominate it — unless that's specifically what you're going for artistically.