Module 4.1 · Foundation Track · Signal-chain order
Compression
After this lesson you'll know what every compressor knob actually does — and why it's the most-misunderstood tool in mixing.
Compression isn't a knob. It's a feeling — the sense that a vocal is right *here*, close to your face, that a snare hits with weight, that the quiet parts of a song still feel present alongside the loud ones. When a record sounds "modern" or "professional," compression is doing 30% of that work, invisibly, on every element.
It's also the tool most beginners get wrong, because the "right amount" of it is almost never audible by itself. You don't hear compression — you hear what it makes possible. A vocal that sits forward without poking out. A bass that locks with the kick instead of floating. A mix that translates from your monitors to a phone speaker without losing its impact.
This module gives you both — the felt understanding (use the visualizer below to see a kick getting flattened in real time, then A/B the bypass to hear the difference), and the technical fluency (every parameter, what it does, when to reach for it).
Going deeper
What compression is actually doing
A compressor is an automatic level controller. When the signal exceeds a threshold (a level you set), the compressor turns it down by a ratio (e.g., 4:1 means for every 4 dB over the threshold, only 1 dB comes out). The speed of the turn-down is the attack; the speed of the recovery is the release. The total amount of turn-down is gain reduction (GR), measured in negative dB.
Mathematically that's all it is. Pedagogically, the felt picture matters more: imagine someone politely riding the fader on a vocal — pulling down the screamed words, pushing up the whispered ones. That's compression, automated. The musical question is *how aggressively* it rides, *how fast* it reacts, and *how soon* it lets go.
The five parameters
| Parameter | What it does |
|---|---|
| Threshold | The level above which compression kicks in. Lower threshold = compressor reacts to more of the signal. Typical: −30 to −12 dBFS for tracking; closer to 0 for mastering. |
| Ratio | How aggressively the compressor turns the signal down. 2:1 is gentle leveling. 4:1 is the workhorse range. 10:1+ is heavy compression. ∞:1 is limiting (nothing past the threshold). |
| Attack | How fast the compressor reacts to a signal exceeding the threshold. Fast attack (under 1 ms) clamps every transient — kills drum punch but controls peaks. Slow attack (20–50 ms) lets the transient through, then catches the body — drums punch harder in compression because the click stays loud while the body gets reduced. |
| Release | How fast the compressor lets go after the signal drops below threshold. Fast (10–100 ms) = audible pumping. Slow (200–1000 ms) = invisible leveling. Auto-release follows the program content and is often the safest choice. |
| Makeup | Adds gain back after compression to match the bypass level. Without makeup, A/B is unfair — compressed always sounds quieter. Always level-match before judging. |
In your DAW
Every major DAW ships with stock compressors that cover most of these topologies. Use the stock plugin first — it does 90% of the job. Reach for paid character emulations only when you specifically want hardware behavior the stock can't replicate.
| DAW | Stock plugin | How to dial it for the chain above |
|---|---|---|
| GarageBand | Compressor | Single plugin, 5 modes. Pick Studio FET for 1176-style snap, Classic Opto for LA-2A leveling smoothness, Studio VCA for SSL-bus glue. |
| Logic Pro | Compressor | Same engine as GarageBand with 7 circuit types. Top-right knob switches between Studio FET, Studio VCA, Vintage FET (1176), Vintage Opto (LA-2A), Vintage VCA, Classic VCA, Platinum Digital. Logic's stock comp is genuinely pro-grade — many engineers mix entire records on it. |
| Ableton Live | Compressor + Glue Compressor | Standard Compressor for general use. Glue Compressor is a built-in SSL G-bus emulation — use it for mix-bus glue. (Most other DAWs require a paid plugin for this; Ableton ships it free.) |
| Pro Tools | Dyn 3 Compressor/Limiter | Transparent VCA-style stock. For character (FET, Opto, vari-mu) you'll want Avid Heat, UAD, or Waves. The free Tokyo Dawn Records TDR Kotelnikov is a solid free supplement. |
| Reaper | ReaComp | Free, transparent VCA. Reaper users typically supplement with TDR Kotelnikov (free) or character plugins. |
| FL Studio | Fruity Compressor / Fruity Limiter | Fruity Limiter has a transparent compression mode plus brick-wall limiting in one plugin. |
Free-first rule: if your DAW's stock compressor has a circuit-type knob (Logic, GarageBand) or a dedicated bus comp (Ableton's Glue), you already have what you need. Spend money on character emulations later when you can hear what's missing — not before.
The three topologies (and why they sound different)
Every compressor — hardware or plugin — uses one of three core circuit designs. The topology determines its character as much as its parameters do, and modern plugins emulate specific hardware. Knowing which is which lets you reach for the right tool:
- FET (UREI 1176) — fast, punchy, bright. Reacts in microseconds. Best on drums, vocals, bass DI — anything that needs attitude. The Waves CLA-76 and UAD 1176 are emulations.
- Optical (Teletronix LA-2A) — slow, smooth, program-dependent. Best on vocals, bass — anything that wants gentle leveling. The Waves CLA-2A and UAD LA-2A emulate it.
- VCA (SSL G bus comp, dbx 160, API 2500) — transparent, controllable. Best on bus compression and the mix bus glue. The Cytomic "The Glue" is the classic SSL emulation.
- Vari-mu / tube (Fairchild 670, Manley Vari-Mu) — slow, harmonic, mastering-grade. Best on the mix bus and mastering. Premium and slow.
The take-away: when you want punch, reach for FET; when you want smoothness, reach for optical; when you want glue, reach for VCA. Most engineers stack them — a fast FET for transients followed by a slow optical for averaging.
The two-stage vocal chain
The classic Chris Lord-Alge approach to a pop/rock vocal is two compressors in series:
- Stage 1 — Leveling: an LA-2A (optical, slow) doing 3–4 dB of GR. Smooths the overall dynamics so the second comp gets a consistent input.
- Stage 2 — Character: an 1176 (FET, fast) doing 2–3 dB of GR. Adds presence and bite without crushing the voice.
Total GR around 5–7 dB. Each comp does its job; neither overworks. This is the move on virtually every modern pop, rock, country, and worship vocal — and you can replicate it in any DAW with stock plugins or free emulations. It's the single most useful chain in mixing.
"When a vocal sits forward and feels effortless, two compressors are doing the work — neither one alone could pull it off." — a paraphrase of every working pop/rock engineer
Common mistakes
- Too much GR. Over 8 dB on a single comp usually sounds bad. Use multiple comps in series with smaller amounts each, or use parallel compression.
- Not level-matching. A compressed signal at a louder output sounds "better" by perception alone. Always match output level to bypass before judging.
- Wrong topology for the source. A slow optical comp can't catch a fast snare transient. A fast FET comp pumps audibly on a sustained pad. Match the comp to the source.
- Fast release on a bus comp. Sounds like obvious pumping. Bus comps want slow attack, slow release, gentle GR (1–3 dB).
- Compressing problems instead of fixing them. If the source has wild dynamics from poor mic technique or performance, compression can't save it. Fix the recording first.
When you reach for parallel compression instead
If you want the density of heavy compression without losing transients, send the source to a parallel bus, crush that bus 10+ dB, and blend it underneath the dry signal. The dry source keeps its transients intact; the parallel return adds body and forward presence. This is the "New York compression" technique, and it's how nearly every modern drum bus and lead vocal gets its bigness without sounding squashed. We'll cover this in detail in Module 11.1 — Mix Bus Glue & Color.