How to Compress Vocals

Vocal compressor settings — attack time, release time, ratio, and the right amount of gain reduction — are one of the most confusing topics in mixing. Get them wrong and the vocal feels either lifeless and squashed or wild and unstable. Get them right and the vocal sits forward in the mix without ever sounding over-processed.

This guide breaks down three vocal compression techniques that working engineers use on commercial records: a one-knob Rvox approach, a series compression chain built around an 1176 and a CLA-2A, and a hybrid method that uses Waves Vocal Rider in front of a slower compressor. Each method has a clear use case, exact starting settings, and a defined gain reduction target so you can pick the right one for the vocal in front of you.

Before Any Compressor: Manually Tame the Peaks First

Before reaching for any vocal compressor — Rvox, the 1176, the CLA-2A, or anything else — the most important step is manual peak control. Use clip gain or volume automation to attenuate the loudest syllables so the compressor receives a controlled input signal. Without this step, a single hot transient can pin the compressor into 12 dB of gain reduction while the rest of the phrase barely triggers it.

The goal is simple: feed the compressor a signal that already lives inside its sweet spot. Once the loudest words are tamed by hand, every method below becomes more predictable, more musical, and far less dependent on luck.

Technique 1: Waves Rvox — Single-Knob Vocal Compression

Waves Rvox is one of the most popular vocal compressors among industry mixing engineers because it is ruthlessly simple. There is essentially one slider, and pulling it down increases gain reduction. The entire workflow is: tame peaks manually, drop Rvox on the vocal, and pull the slider until the meter sits between 3 and 7 dB of gain reduction.

That 3 to 7 dB range is the Rvox sweet spot. Below 3 dB, the compressor barely engages and the vocal dynamics stay uneven. Above 7 dB, the vocal starts to lose body and feel squashed. When the input is already controlled, watching the waveform after Rvox should show a smooth, even shape — not a flat line and not a series of spikes.

Use this method when the vocal performance is already fairly consistent and you want a fast, in-your-face result without spending time on multiple compressors.

Technique 2: Series Compression — FET Into Opto for Peaky Vocals

Series compression is the right answer when the vocal has a lot of dynamic range — quiet verses, loud hooks, sharp transient consonants. The idea is to use two compressors back-to-back, each doing a different job. The first is a fast peak compressor, typically a FET-style unit like the Waves CLA-1176 or the Arturia FET. The second is a slow, smooth opto compressor, classically the Waves CLA-2A.

Stage one — the FET compressor — handles transients. Set attack and release to the fastest setting and the ratio to 12:1 or 20:1. At those ratios the unit behaves almost like a limiter, catching the loudest peaks. Aim for a maximum of about 7 dB of gain reduction on the hardest hits.

Stage two — the opto compressor — smooths the result. Drop a CLA-2A (or any opto-style compressor) after the FET stage and aim for 2 to 3 dB of gain reduction. Make sure no plugins sit between the two compressors. The chain has to be uninterrupted to behave like a true series compression.

If you do not own these specific plugins, the same FET-into-opto philosophy translates to stock plugins in Logic Pro, Ableton Live, and Studio One — pick a fast-attack stock compressor at a high ratio for stage one, then any slow opto-style stock compressor at a 2 to 3 dB target for stage two.

Pro Tip: Use the Arturia FET to Cap Gain Reduction

Even with manual peak control upstream, some vocal performances still throw the FET compressor into more gain reduction than you want. The Arturia FET compressor solves this with a built-in gain reduction limit — a feature that lets you cap the maximum amount of compression the unit can apply, regardless of input.

Setting that cap to your target — for example 7 dB — keeps the compressor in its musical range no matter how aggressive the singer gets on a particular take. It is one of the most underrated features in any modern 1176-style emulation.

Pro Tip: The Waves L1 Limiter Trick From Ilya Salmanzadeh

Ilya Salmanzadeh, a Grammy-winning music producer from Max Martin’s team, revealed on Instagram that he uses the Waves L1 Limiter on Ariana Grande’s vocals to catch peaks. The L1 has an instant attack and a brick-wall ceiling, so it shaves the loudest moments without colouring the body of the vocal.

That makes it a valid drop-in replacement for the FET stage in a series chain. Instead of an 1176-style compressor at fast attack and high ratio, place an L1 with a moderate threshold so it only intervenes on the loudest words, then follow it with the same opto-style compressor as before. The result is a slightly cleaner, less coloured peak control before the opto smooths the rest.

Technique 3: Vocal Rider Plus Compressor — The Hybrid Method

The third method is a hybrid that sits between manual automation and pure compression. Open an instance of Waves Vocal Rider on the vocal channel. Set the range sliders to the maximum and minimum positions with the gain in the middle, set the attack to fast, increase the vocal amount, and adjust the target slider so it tracks closely with the input signal — not far to the left, not far to the right.

Vocal Rider automates the channel volume rather than compressing the signal. It increases the lower parts of the vocal and decreases the louder parts, but it does so by riding the fader, not by squeezing the waveform. The vocal becomes more even without ever feeling crushed, and every detail of the performance stays alive.

You can also have Vocal Rider write the rides as automation lanes you can edit by hand. Enable write automation both inside the plugin and on the audio channel, and the rides become editable automation curves you can shape per phrase.

After Vocal Rider, place any compressor with a medium-fast attack and a comparatively slow release time, around 170 milliseconds, aiming for 2 or 3 dB of gain reduction. Because the input signal is already level-stabilised, the compressor never has to work hard. It contributes glue and tone rather than dynamic control, and the result is a vocal that is forward, smooth, and natural at the same time.

When Each Technique Is Wrong

Picking the wrong method for the source is more common than mis-setting the controls. Each of these techniques fails in a specific situation, and recognising those situations is what separates a working engineer from a tutorial follower.

• Rvox is wrong for highly dynamic vocals. If the performance has 12+ dB of dynamic range between the quietest and loudest words, Rvox will either over-compress the loud parts to bring up the quiet ones, or stay transparent on the loud parts and leave the quiet ones unintelligible. Use a series chain or Vocal Rider instead.
• Series compression is wrong for already-tame vocals. If the vocal is consistent and close-mic’d at a controlled level, two compressors in series will flatten the natural breath articulation and remove micro-dynamics that make a vocal feel alive. A single Rvox is enough.
• Vocal Rider is wrong for fast rap delivery. Vocal Rider responds at phrase level, not syllable level. On dense rap performances with rapid level changes inside a single bar, the rider cannot react fast enough and either leaves transients exposed or chases them in a way that sounds pumpy. Use a fast FET compressor in series instead.
• Any of these methods is wrong without manual peak control upstream. Skip the clip-gain step and every method becomes unpredictable, because the compressor’s behaviour depends entirely on a single uncontrolled syllable.

3 Vocal Compression Mistakes to Avoid

These three mistakes are the most common reasons a compressed vocal sounds amateur even when the plugin choice is correct:

1. Setting the threshold by ear without watching the gain reduction meter. Without the meter, you cannot tell if the compressor is pulling 2 dB on most words and 12 dB on one shouted word. Always confirm with the gain reduction meter — and aim for 3–7 dB consistently, not occasionally.
2. Using a fast attack on a singer who relies on consonant impact. A 1176 set to fastest attack will eat the leading edge of every consonant, which thins articulation and pushes the singer back in the mix. If the vocalist depends on attack — rappers, soul singers, anyone with a percussive style — slow the attack down or move the FET stage later in the chain.
3. Stacking too much makeup gain to compensate for over-compression. If you have to push +10 dB of makeup just to get the vocal level back, you are over-compressing. Pull the threshold up until you can match level with under +6 dB of makeup. This single fix solves more "my vocal sounds squashed" complaints than any plugin choice.

Comparing the Three Vocal Compression Techniques

All three methods produce smooth, controlled vocals — but they suit different scenarios:

• Rvox (Technique 1) — fastest workflow. Best for vocals that are already fairly consistent and need an in-your-face character with minimal effort.
• Series compression with the 1176 and CLA-2A (Technique 2) — best for peaky, dynamic vocals where you need real transient control followed by tonal smoothing.
• Vocal Rider plus a slow compressor (Technique 3) — best when you want maximum detail preservation. The rider does the loud-versus-quiet work, and the compressor only has to add glue.

Practical Workflow Summary

1. Tame the loudest vocal peaks manually with clip gain or volume automation before any compressor.
2. For the simplest result, drop Rvox and pull the slider to 3 to 7 dB of gain reduction.
3. For peaky vocals, use a fast FET compressor (1176-style) at fast attack, fast release, ratio 12:1 or 20:1, max 7 dB of gain reduction.
4. Follow the FET stage with a slow opto compressor (CLA-2A-style) targeting 2 to 3 dB of gain reduction, with no plugins between them.
5. If peaks still misbehave, use the Arturia FET compressor’s built-in gain reduction limit, or substitute the Waves L1 Limiter for the FET stage.
6. For maximum detail, place Vocal Rider first with wide range, fast attack, and the target tracking the input.
7. After Vocal Rider, use any compressor at medium-fast attack, ~170 ms release, aiming for 2 to 3 dB of gain reduction.
8. Match the technique to the source: Rvox for consistent vocals, series for peaky vocals, Rider plus compressor for maximum detail.

Frequently Asked Questions

What is the sweet spot for gain reduction on a vocal with Waves Rvox?

Aim for 3 to 7 dB of gain reduction on Rvox. Below 3 dB the compressor barely engages and the vocal stays inconsistent; above 7 dB the dynamics start to flatten and the vocal can sound squashed. Before turning the slider down, manually attenuate the loudest peaks with clip gain so Rvox sees a controlled input and stays in that range.

Why use series compression on a vocal instead of one compressor?

Series compression splits the workload between a fast peak compressor and a slow level compressor. A FET-style unit like the Waves CLA-1176 or Arturia FET catches transients quickly, then an opto-style unit like the Waves CLA-2A smooths the resulting signal. One compressor trying to do both jobs has to choose between a fast attack that thins the body or a slow attack that misses peaks. Two compressors in series let each one do what it is good at.

What attack, release, and ratio should I use on a 1176-style FET compressor for vocals?

Set attack and release to the fastest setting and the ratio to 12:1 or 20:1, which makes the compressor behave more like a limiter on transient peaks. Aim for a maximum of about 7 dB of gain reduction on the loudest words. If peaks push past that range, the Arturia FET compressor has a built-in gain reduction limit that prevents the unit from over-grabbing.

Why use Waves Vocal Rider before a compressor?

Vocal Rider automates the channel volume rather than compressing the signal, so it stabilises phrase-level loudness without crushing transients. When it sits before the compressor, the compressor receives a vocal that is already even, which means it can focus on tone and density rather than fighting wide level swings. The result is a vocal that feels louder and more present without sounding squashed.

Can I replace the FET compressor with the Waves L1 Limiter on vocals?

Yes. Ilya Salmanzadeh, a Grammy-winning producer from Max Martin’s team, has said on Instagram that he uses the Waves L1 Limiter to catch peaks on Ariana Grande’s vocals. The L1 acts as an instant-attack peak catcher, so swapping it in for a 1176-style FET compressor is a valid first stage in a series chain — it shaves the loudest moments before an opto compressor like the CLA-2A smooths the rest.

When should I not use the Waves Vocal Rider on vocals?

Avoid Vocal Rider on dense rap performances with rapid level changes inside a single bar. Vocal Rider responds at phrase level, not syllable level, so it cannot react fast enough on fast rap delivery and tends to either leave transients exposed or chase them in a way that sounds pumpy. Use a fast FET compressor in series for that source instead.

How much makeup gain is too much on a vocal compressor?

If you need more than about +6 dB of makeup gain to match the bypassed level, the compressor is doing too much work. Pull the threshold up until you can match level with under +6 dB of makeup. Stacking +10 dB or more of makeup gain is the single most common reason engineers describe their vocals as squashed even when the plugin choice is correct.