Vocal Compression Guide: Settings, Techniques & Plugins
Vocals live or die by compression. Too little and they disappear in the mix. Too much and they sound strangled, lifeless, and processed. The difference between amateur and professional vocal production almost always comes down to how compression is applied — and in what order. This guide covers everything: settings, compressor types, genre-specific approaches, plugin recommendations, and the serial compression techniques used by professional mix engineers.
Start vocal compression with a ratio of 3:1–4:1, attack around 10–20ms, release around 80–120ms, and lower the threshold until you see 3–6dB of gain reduction on the loudest phrases. Use makeup gain to restore level. Add a second compressor for density and character. Automate volume separately to handle phrase-to-phrase differences before the compressor sees them.
Why Vocal Compression Is Different
The human voice is the most dynamically unpredictable instrument in any mix. A skilled singer can shift 20–30dB between a whispered verse and a belted chorus within a single take. Even within a single phrase, the level difference between consonants and vowels can be 10–15dB. No fader move or static EQ can manage this — only compression can track and respond fast enough.
But the voice is also the most scrutinised element in any production. Listeners are wired to hear the human voice above all else, and the moment compression becomes audible — pumping, over-squashed consonants, unnatural decay — it breaks the emotional connection with the listener. Vocal compression must be effective and invisible. That is the core challenge.
The professional solution is never a single compressor doing all the work. It is a chain — automation, then two compressors, then potentially limiting — each stage handling a specific aspect of control, so no single stage has to work too hard.
The Vocal Compression Signal Chain
Before touching a compressor, understand the order of the full vocal processing chain. Compression does not happen in isolation.
Each stage has a specific job. Volume automation rides large phrase-to-phrase differences before the compressor sees them — this means the compressor only has to manage the remaining micro-dynamics within each phrase, working far less aggressively. The first compressor does the heavy levelling. The second adds density and colour. Creative EQ and effects come last, after the signal is already controlled.
Core Compression Parameters — Vocals Specifically
Threshold
Set the threshold so you are getting 3–6dB of gain reduction on the loudest phrases of the vocal. Watch the gain reduction meter while the vocal plays. The needle or bar should be moving — but only on peaks, not constantly. If the compressor is in gain reduction on every single syllable at all times, the threshold is too low. Lower it gradually until peak phrases trigger 4–5dB GR and quieter phrases trigger 0–2dB.
Ratio
For sung vocals: 2:1–4:1 for transparent levelling. For pop lead vocals: 3:1–5:1. For rap and hip-hop: 4:1–8:1 — the percussive delivery benefits from a firmer clamp. Never start above 6:1 on a vocal without a specific reason. High ratios kill the natural breath and expression that makes a vocal performance human.
Attack
Vocal attack time is one of the most critical settings and the most commonly wrong. Too fast (under 3ms) and the compressor catches the initial consonant of every syllable — words start sounding muffled and dull, the vocal loses intelligibility and presence. The sweet spot for most sung vocals is 10–25ms. This lets the hard consonant (the "T", "K", "P" sounds that define intelligibility) pass through before compression kicks in. Rap vocals can handle slightly faster attack (5–15ms) because the delivery is more rhythmically driven.
Release
Release time controls how quickly the compressor lets go between compressed moments. Too slow and the compressor stays in gain reduction between syllables, sucking out the space and making the vocal sound dense and airless. Too fast and you get distortion artifacts or rapid pumping on sustained vowels. The starting point for vocals is 80–120ms. For faster-paced rap or spoken word, 50–80ms. For slow ballads where you want the compressor to breathe with the phrasing, 150–200ms. Auto-release modes (available on most modern plugins) track the programme material and are a legitimate shortcut — use them if you are not hearing obvious issues.
Makeup Gain
After compression, the output signal is quieter than the input — the compressor has reduced gain on peaks. Makeup gain brings it back up. The most important rule: level-match before making any compression decisions. A compressed vocal at the same loudness as the dry vocal is genuinely easier to judge. Boosting 3–4dB of makeup gain makes anything sound better, which tricks you into thinking the compression is helping when it may be hurting.
Compressor Types and Their Vocal Characters
The type of compressor — its circuit topology — determines the character it imparts on a vocal as much as the settings do. Understanding the four main types lets you choose the right tool before reaching for the knobs.
VCA Compressors — Fast, Punchy, Precise
VCA (Voltage Controlled Amplifier) compressors are the fastest and most versatile. They respond quickly to transients, have a definite knee, and give you precise control over attack and release. On vocals, VCA compressors add presence and punch — they grab peaks hard and let go cleanly. The SSL G-channel compressor, the 1176, the dbx 160 — these are all VCA designs. In plugin form: the Waves SSL E-Channel, FabFilter Pro-C 2 in VCA mode, Slate Digital VBC FG-Grey.
Use a VCA as your second compressor for adding density and forward presence after an optical first stage. Or use it alone when you want a controlled, modern, high-presence vocal sound.
Optical Compressors — Smooth, Musical, Natural
Optical compressors use a light element and photoresistor to control gain. The result is a program-dependent response — the compressor's behaviour changes based on the characteristics of the incoming signal, not just a fixed ratio. This produces extremely natural-sounding compression that is very difficult to over-do. The Teletronix LA-2A is the definitive optical compressor — it has been on more lead vocal recordings than almost any other piece of hardware in history.
Optical compressors have a slower, more musical release that tracks the natural decay of the voice. They are ideal as the first stage in a serial chain — handling the bulk of levelling without sounding like compression. Plugin equivalents: Waves CLA-2A, UAD LA-2A, Pulsar Mu (for tube optical character).
FET Compressors — Aggressive, Characterful, Fast
FET (Field Effect Transistor) compressors are the category the 1176 defines. They are extremely fast — faster than optical and most VCA designs — and impart a distinctive harmonic character from the FET circuitry. At moderate settings they add punch and presence. Pushed hard, they saturate and add grit. The "all-buttons-in" mode of the 1176 is one of the most celebrated vocal sounds in recording history.
FET compressors work best on vocals that need to cut through dense mixes — heavily produced pop, rock, hip-hop. They add excitement and aggression. They are not the tool for delicate, exposed ballads where transparency is the goal.
Variable-Mu (Tube) Compressors — Warm, Thick, Vintage
Variable-mu compressors like the Fairchild 670 and Manley Variable Mu use tubes to control gain reduction. They are the slowest type — not suitable for catching fast transients — but they add a warmth, depth, and harmonic richness that no other compressor type replicates. On vocals, a tube compressor in the chain (even at low settings) thickens the sound and adds a sense of body that sits beautifully in dense productions.
Plugin equivalents: Pulsar Mu, UAD Fairchild 670, Waves Kramer Master Tape (for saturation-adjacent thickness).
Serial Compression — The Two-Compressor Approach
Using two compressors in series is the professional standard for lead vocals. The principle: share the total compression workload between two stages so neither compressor has to work hard enough to audibly compress. The result is a vocal that sounds controlled without sounding processed.
Stage 1 — Levelling (Optical or Light VCA)
The first compressor handles the large-scale dynamics. Set it to catch the biggest peaks — the chorus that's 10dB louder than the verse, the belted high note that shoots above everything else. Ratio: 2:1–3:1. Attack: 20–40ms. Release: auto or 100–200ms. Target: 4–6dB GR on the loudest moments. This stage should be as transparent as possible — an optical compressor like the LA-2A is ideal here because its program-dependent response sounds natural even at meaningful gain reduction amounts.
Stage 2 — Density and Character (VCA or FET)
The second compressor works on a signal that has already been levelled by Stage 1. Because the big peaks are gone, Stage 2 can be set more aggressively without pumping. Ratio: 3:1–6:1. Attack: 5–15ms. Release: 60–100ms. Target: 2–4dB GR continuously. This stage adds density, presence, and the character of its circuit topology. A 1176 or SSL-type VCA here gives a forward, controlled, professional vocal sound.
Why This Works
If you set a single compressor to do 10dB of total gain reduction, you hear it — pumping, artifacts, squashed consonants. If you split that same 10dB between two compressors (6dB from Stage 1, 4dB from Stage 2), each compressor is working within its natural range and neither sounds like it is working at all. The combined result is a more controlled, denser, more professional vocal with less obvious compression character.
Genre-Specific Vocal Compression Approaches
| Genre | Ratio | Attack | Release | Target GR | Character |
|---|---|---|---|---|---|
| Pop (sung) | 3:1 – 4:1 | 10–20ms | 80–120ms | 4–6dB | Smooth, controlled, forward presence |
| Hip-Hop / Rap | 4:1 – 8:1 | 5–15ms | 60–100ms | 6–10dB | Punchy, dense, sits on top of the beat |
| R&B / Soul | 2:1 – 4:1 | 15–30ms | 100–150ms | 3–5dB | Warm, dynamic, preserves runs and emotion |
| Rock | 4:1 – 6:1 | 10–20ms | 80–120ms | 4–8dB | Aggressive, cuts through guitars |
| Country | 2:1 – 3:1 | 20–40ms | 100–200ms | 3–5dB | Natural, storytelling delivery preserved |
| Jazz / Folk | 1.5:1 – 2:1 | 30–50ms | 150–300ms | 1–3dB | Transparent — dynamics are the art form |
| EDM / Pop (electronic) | 4:1 – 8:1 | 5–10ms | 50–80ms | 6–10dB | Hyper-compressed, sidechain-ready density |
| Podcast / Spoken Word | 3:1 – 4:1 | 5–10ms | 50–80ms | 4–8dB | Consistent, intelligible at all levels |
Common Vocal Compression Mistakes
Compressing Before Volume Automation
If your verse is 15dB quieter than your chorus and you run both through a compressor, the compressor has to work extremely hard to manage that range — and it will sound like it is working. Always automate phrase-level differences first. Pull the chorus down or ride the verse up before the compressor even hears the signal. Then set your compressor threshold for the remaining variation within phrases.
Slow Attack Killing Consonants — Wait, No. Fast Attack Killing Consonants
Fast attack on a vocal (under 5ms) grabs the initial transient of every consonant — the "T", "S", "P", "K" sounds that define intelligibility. The vocal starts sounding like someone pulled a thin blanket over it. If your compressed vocals sound muffled or dull, try slowing the attack to 15–25ms. You will immediately hear consonants snap back into focus.
Using Ratio to Fix a Threshold Problem
Many beginners raise the ratio when the vocal still sounds uncontrolled — when the real fix is lowering the threshold. A 4:1 ratio with the threshold too high means the compressor barely catches the peaks. The fix is not 8:1 — it is lowering the threshold so the 4:1 compressor catches what it needs to catch. Ratio sets the slope; threshold sets how much of the vocal is above it.
Not Listening in Context
Vocals always sound over-compressed when soloed and under-compressed when buried in a dense mix. Set your vocal compression with the full mix playing, not in solo. The goal is for the vocal to sit and stay in the mix — not to sound perfect in isolation. What sounds like too much compression in solo often sounds like the right amount when the drums and bass are running.
Skipping the Gain Reduction Meter
Always watch the GR meter while compressing. If it is barely moving, the threshold is too high. If it is pinned constantly, the threshold is too low. Target: GR meter active on peaks, returning to zero on quieter phrases. This visual feedback tells you exactly how hard the compressor is working and whether your settings are actually doing what you think they are doing.
Best Compressor Plugins for Vocals
FabFilter Pro-C 2 — Best All-Rounder
The Pro-C 2 models multiple compressor circuit types (VCA, FET, optical, tube, and more) in one interface, with transparent operation, excellent metering, and a variable knee. The large graphical display shows exactly what compression is doing to your signal. For engineers learning compression, this visual feedback is invaluable. For professionals, the clean, accurate sound makes it the default choice when you want the compression to be invisible. Check price at Plugin Boutique
Waves CLA-2A — Best Optical for Smooth Levelling
Chris Lord-Alge's emulation of the LA-2A is a consistent industry standard on lead vocals. It has only two controls — Peak Reduction (compression amount) and Output (makeup gain) — which forces you to focus on the result rather than the settings. The program-dependent response means it gets harder to over-compress with it, making it a great first compressor in a serial chain. Check price at Plugin Boutique
Universal Audio 1176 — Best FET for Presence and Punch
The UAD 1176 is the reference for punchy, forward vocal compression. The 4:1 setting adds presence and snap. The 8:1 gets aggressive. All-buttons-in produces that distinctive thick, driven character. It is the second compressor in most professional vocal chains — the one that adds character after the optical has done the levelling. Available as a UAD plugin (requires Apollo hardware) or through Universal Audio's LUNA ecosystem. Check price at Plugin Boutique
Waves SSL E-Channel — Best for Modern Pop and Hip-Hop
The SSL E-Channel compressor is a VCA design with a characteristic fast response and modern, punchy sound that sits perfectly in dense productions. The channel strip format (EQ and compressor together) matches how it would be used on an analogue console, and the integrated approach makes it easy to get a cohesive vocal sound quickly. Check price at Plugin Boutique
Klanghelm MJUC — Best Budget Option
The MJUC is a variable-mu tube compressor emulation that punches far above its price point. At low ratios it adds warmth and thickness to vocals without obvious compression character. For producers who cannot afford UAD hardware but want tube warmth in their vocal chain, this is the go-to. Available as a paid plugin with a free lite version (MJUC Jr). Check price at Plugin Boutique
Vocal Compression Cheat Sheet
| Parameter | Starting Point | Too Much Compression | Too Little Compression |
|---|---|---|---|
| Threshold | −18 to −24dBFS | Compressor constantly in GR | GR meter barely moves |
| Ratio | 3:1 – 4:1 | Vocal sounds strangled, no dynamics | Loud notes still jump out |
| Attack | 10–20ms | Consonants sound muffled, dull | Fast peaks still escape |
| Release | 80–120ms | Pumping between words/phrases | Compressor holds too long, airless sound |
| Gain Reduction | 3–6dB peak | Over 8dB from single compressor | Under 1dB — negligible effect |
| Makeup Gain | Match uncompressed level | Boosted too high — artificial loudness | Compressed vocal too quiet in mix |
Exercises: Vocal Compression
🟢 Beginner — One Compressor, One Vocal
Take a lead vocal recording and place one compressor on it — the FabFilter Pro-C 2 or any compressor available in your DAW. Set ratio to 4:1. Play the vocal and slowly lower the threshold until the gain reduction meter shows 4–5dB on the loudest phrase. Set attack to 15ms and release to 100ms. Apply 4dB of makeup gain. Now A/B the compressed versus uncompressed signal at the same loudness (lower the makeup gain to match levels for a fair comparison). Can you hear the difference in consistency? Do the loud phrases sit closer to the quieter ones? Adjust attack until any muddiness on consonants clears. This single-compressor exercise teaches you to hear what compression is and is not doing before adding complexity.
🟡 Intermediate — Serial Compression Chain
Set up two compressors in series on a lead vocal. Compressor 1: optical character (CLA-2A, LA-2A, or any optical emulation). Set Peak Reduction to achieve 4–6dB GR on the chorus. Compressor 2: VCA character (SSL, 1176, Pro-C 2 in VCA mode). Set ratio 4:1, attack 10ms, release 80ms, threshold for 2–3dB GR continuously. Add makeup gain at each stage to compensate for GR. Play the full vocal through the chain and compare: single compressor doing all 8dB of GR versus the two-stage chain sharing the same 8dB. The two-stage version should sound more controlled with fewer obvious compression artifacts. This exercise makes the case for serial compression better than any description.
🔴 Advanced — Parallel Vocal Compression with Sidechain EQ
Set up a parallel compression send on your lead vocal. Send 100% of the vocal to a separate bus with a compressor set to 8:1, fast attack (5ms), fast release (50ms) — heavy compression. Return this compressed signal at low level and blend it with the dry vocal. The blend adds density and sustain without the heavy compression artefacts of applying that ratio directly. Now add a sidechain high-pass filter inside the compressor at 150–200Hz. This filters the low-frequency content from the sidechain signal (the signal that triggers compression) — so deep vocal chest resonance no longer triggers the compressor. Result: a more consistent vocal response that does not pump when the vocalist hits chest-heavy phrases. A/B with and without the sidechain HPF and note how much more even the compression feels.
Frequently Asked Questions
What are good compression settings for vocals?
A good starting point: ratio 3:1–4:1, attack 10–30ms, release 60–120ms, threshold set to achieve 3–6dB of gain reduction on the loudest phrases. Use makeup gain to restore level after compression.
Should I use one compressor or two on vocals?
Two compressors in series is the professional standard. The first (often optical) handles large peaks with 3–6dB GR. The second adds character and density with 2–4dB GR. This produces a more natural result than one compressor doing all the work.
What attack time should I use for vocal compression?
For most vocals, 10–30ms is the starting point. A slower attack lets consonants through before compression kicks in, preserving intelligibility. Faster attacks (under 5ms) can dull consonants and make vocals sound muffled.
What is the best compressor plugin for vocals?
Top options: FabFilter Pro-C 2 (transparent and versatile), Waves CLA-2A (smooth optical levelling), UAD 1176 (punchy FET character), Waves SSL E-Channel (modern pop and hip-hop), and Klanghelm MJUC (warm tube compression at low cost).
How much gain reduction should I use on vocals?
For most lead vocals, 3–6dB of peak gain reduction. On very dynamic vocalists, up to 8–10dB is acceptable if spread across two compressors sharing the work.
What is the difference between VCA and optical compression on vocals?
VCA compressors are fast and punchy — great for controlled, forward vocals in dense mixes. Optical compressors are slower and smoother — great for natural levelling that does not sound like compression. Many engineers use one of each in series.
Should I compress vocals before or after EQ?
Both orders work and serve different purposes. EQ before compression changes what the compressor responds to — useful for preventing harsh frequencies from triggering over-compression. EQ after compression shapes the final tonal balance. Most engineers use corrective EQ before, creative EQ after.
What compression ratio should I use for rap vocals?
Rap vocals benefit from higher ratios than sung vocals — typically 4:1–8:1. Fast attack (5–15ms) catches peaks. Medium release (80–150ms) tracks the rhythm. Heavy compression on rap vocals is as much an aesthetic choice as a technical one.
How do I compress vocals without killing the dynamics?
Use a lower ratio (2:1–3:1) with a higher threshold so the compressor only catches the loudest peaks. Use a slower attack to preserve natural phrase expression. Then use volume automation to ride quieter sections up manually, rather than relying on the compressor to flatten everything.
What is parallel compression for vocals?
Parallel vocal compression blends a heavily compressed signal with the dry vocal. The compressed signal adds density and sustain. The dry signal preserves the natural dynamic shape and consonant clarity. The result is fuller and more controlled than either signal alone.
Do I need to compress vocals if I use automation?
Automation handles phrase-to-phrase level differences. Compression handles micro-dynamics within phrases — syllable-to-syllable variation. You need both for a polished vocal in a professional mix. Automation first, then compression.
Practical Exercises
Set Your First Vocal Compressor
Open your DAW and load a vocal track with dynamic range (loud phrases and quiet phrases). Insert a compressor plugin. Set the ratio to 3:1, attack to 15ms, and release to 100ms. Play the vocal and slowly turn down the threshold until you see 4–5dB of gain reduction on the loudest words. Now turn on makeup gain to bring the overall level back up so the compressed vocal matches the original in loudness. Play it back. You should hear the vocal sitting more evenly — the loud parts are tamed, quiet parts are lifted. This is basic levelling compression. Save this setting as a preset.
Build a Two-Compressor Vocal Chain
Insert two compressor instances on your vocal track in series. On the first compressor (levelling stage), use a ratio of 4:1, attack 10ms, release 80ms, and set threshold to get 3–4dB of reduction. On the second compressor (character stage), try a slower attack (30–40ms) and a higher ratio (6:1–8:1), aiming for 2–3dB of reduction. Toggle between listening to just the first compressor, just the second, and both together. Decide: Does the first compressor handle too much work? Is the second adding glue or obvious pumping? Adjust ratios and thresholds so each compressor does one job cleanly, then export a 30-second clip comparing no compression, single compressor, and the full chain side-by-side.
Automate, Compress, and Match Professional Dynamics
Record or load a vocal with phrase-to-phrase level differences (quiet verse, loud chorus, whispered ad-lib). First, draw volume automation rides on a duplicate vocal track to tame the biggest peaks before any compressor touches it — aim for a more consistent starting level. Route this automated vocal to a new track with your two-compressor chain (levelling + character). Set aggressive compression: first compressor at 4:1 with 5dB reduction, second at 7:1 with 3dB reduction. Now compare the final result to a professional reference vocal in the same genre. Is your vocal still dynamic and emotional or over-processed? Adjust automation depth and compressor settings to match the reference's tonal character and dynamics. The goal: compression that controls without strangling. Document your final threshold, ratio, and attack/release values for future sessions.
Frequently Asked Questions
The first compressor (optical/VCA) handles heavy levelling to control dynamic range across phrases, while the second compressor (VCA/tube) adds density and character to the vocal tone without aggressive gain reduction. Using two compressors prevents any single stage from working too hard and becoming audible, which is critical for maintaining vocal clarity and emotion.
Volume automation rides handle large phrase-to-phrase dynamic differences before the signal reaches the compressor, meaning the compressor only needs to manage micro-dynamics within each phrase. This approach allows the compressor to work much less aggressively while still achieving professional control, keeping the vocal sound natural and transparent.
Begin with a 3:1–4:1 ratio, attack of 10–20ms, release of 80–120ms, and adjust the threshold until you see 3–6dB of gain reduction on the loudest phrases. Use makeup gain to restore the overall level after compression, then evaluate how transparent the effect sounds before adjusting further.
The human voice is the most scrutinized element in any mix, and listeners are naturally wired to detect it above all other elements. If compression becomes audible through pumping, over-squashed consonants, or unnatural decay, it breaks the emotional connection with the listener, so vocal compression must control dynamics without drawing attention to itself.
A skilled singer can shift 20–30dB between a whispered verse and a belted chorus within a single take, and within a single phrase, consonants and vowels can differ by 10–15dB. This level of unpredictability is why no fader move or static EQ alone can manage vocals—only compression can respond fast enough.
Corrective EQ like high-pass filtering should be applied after gain staging and volume automation but before the first compressor. This ensures the compressor is responding to the corrected signal and not over-reacting to frequencies that will be removed anyway, improving overall control and efficiency.
Creative EQ, saturation, and other effects should be applied after compression is complete. This order ensures the signal is already dynamically controlled before creative processing, preventing effects from being inconsistently applied across loud and soft vocal sections.
A single compressor attempting to handle both levelling and character work becomes audible and aggressive, often resulting in pumping or unnatural sound. The professional approach distributes the workload across two specialized compressors plus automation, so each stage handles a specific aspect of control with transparency.