Introduction to Multiband Compression

Multiband compression is one of the most versatile tools available to mixing and mastering engineers. Unlike a standard broadband compressor that applies the same gain reduction across the entire frequency spectrum, a multiband compressor splits the audio signal into two or more discrete frequency bands, each with its own independent compression controls. This allows you to apply dynamic processing precisely where it is needed, without affecting the rest of the mix. For example, you can tame a harsh vocal sibilance around 6–8 kHz without dulling the low-mid body of the voice, or control a boomy snare resonance at 200 Hz without squashing the snare's attack.

Multiband compression has been a staple in commercial hardware units like the dbx 1066 and the TC Electronic Finalizer, and is now standard in almost every major DAW plugin, from FabFilter Pro‑MB to iZotope Ozone Dynamics. Understanding how to use it effectively can dramatically improve the clarity, punch, and balance of your mixes, especially when dealing with instruments or mixes that have stubborn problem frequencies.

What is Multiband Compression?

At its core, multiband compression works by splitting the incoming audio through a bank of crossover filters. Each crossover creates a separate signal path for a defined frequency range—typical splits are low (e.g., 20–200 Hz), low‑mid (200 Hz–2 kHz), high‑mid (2 kHz–8 kHz), and high (8 kHz–20 kHz). Some plugins allow you to set custom crossover points and even use linear-phase or minimum-phase filter slopes.

Once the audio is divided, each band passes through its own dedicated compressor section. You can independently set the threshold, ratio, attack, release, and sometimes even the knee and sidechain filter for each band. After compression, the bands are summed back together to produce a single output. The result is a dynamic processor that can behave completely differently for low frequencies than for high frequencies—something a single wideband compressor cannot achieve.

One critical distinction is the filter type. Linear-phase crossovers preserve transient timing but introduce pre‑ring, while minimum-phase crossovers have no pre‑ring but cause subtle phase shifts at the crossover frequencies. For most mixing tasks, minimum-phase is preferred because it sounds more natural; linear-phase is often reserved for mastering where transient alignment is less of a concern.

Why Use Multiband Compression?

  • Control problem frequencies without collateral damage: A wideband compressor cannot differentiate between a muddy 150 Hz and a clean 1 kHz. Multiband compression lets you specifically reduce the dynamic content of a problematic frequency range—like a masking low‑mid resonance in a vocal track—while leaving the rest of the performance untouched.
  • Preserve natural dynamics: Because you’re only compressing the band(s) that need it, the overall dynamic range of the source remains intact. This is especially valuable for acoustic instruments and vocals where over‑compression destroys nuance.
  • Improve clarity and intelligibility: By taming resonant spikes, you can reduce listener fatigue and bring buried elements forward. For example, applying gentle multiband compression to the 2–4 kHz region of a bus can help dialogue cut through a dense mix.
  • Creative sound shaping: Use aggressive band compression to create pumping effects in specific ranges (e.g., a sidechained low band on a bass to make room for a kick), or to emulate the effect of vintage units that naturally vary compression by frequency.
  • Mastering glue and coherence: In the mastering stage, multiband compression can tame excessive low‑end energy, smooth out harsh high frequencies, and unify the overall spectrum without making the master sound over‑processed.

How to Set Up Multiband Compression Effectively

Step 1: Identify the Problem Frequencies

Before reaching for a multiband compressor, use your ears and a spectrum analyzer to pinpoint exactly where the issue lies. Solo the track and listen for honkiness (around 500 Hz–1 kHz), boxiness (200–400 Hz), sibilance (5–8 kHz), or rumble (20–80 Hz). Make a note of the frequencies that sound harsh, boomy, or resonant.

Step 2: Set the Crossover Points

Most multiband compressors allow you to drag crossover markers on a frequency display. Start with no more than three or four bands. A common setup is: low band (20–200 Hz), mid band (200 Hz–2 kHz), high band (2 kHz–20 kHz). If your problem is strictly in the high mids, you can create a dedicated band just for that range. Keep crossover points away from strong fundamental frequencies (e.g., avoid 100 Hz for a kick) to prevent phase cancellation at the seams.

Step 3: Adjust Threshold and Ratio

Begin by activating the band that contains the problem frequency. Apply moderate compression with a ratio between 2:1 and 4:1. Lower the threshold until you see 2–4 dB of gain reduction on that band. Listen to how it affects the overall track—if the sound becomes too flat or unnatural, reduce the ratio or raise the threshold. For very narrow problems, use higher ratios (up to 8:1) but with very low thresholds to avoid pumping.

Step 4: Fine‑Tune Attack and Release

Attack and release times should be set to match the material. For transient‑heavy sounds (drums, percussion), use a fast attack (1–10 ms) and a medium release (40–100 ms). For sustained sounds (pads, bass), slower attacks (10–30 ms) allow the initial punch through, and longer releases (100–300 ms) avoid distortion. In the high band, a faster release can help reduce sibilance while preserving air. Always listen critically—if you hear “breathing” or unnatural volume changes, your attack/release may be too fast or too slow.

Step 5: Use Makeup Gain Sparingly

After compression, each band may have a makeup gain control. Instead of automatically boosting, bypass the compressor on each band to compare the level—often you don’t need makeup gain at all. If the compressed band sounds quieter, raise its output gain until the perceived volume matches the bypassed signal. Over‑boosting can reintroduce the problem frequencies you just tamed.

Practical Applications in Mixing and Mastering

Vocals – Taming Sibilance and Resonances

Vocals are prime candidates for multiband compression. A common technique is to isolate the 5–8 kHz band to control sibilance. Set a moderate ratio (3:1) with a fast attack (1–3 ms) and release (50–80 ms) so that the compressor acts only on the harsh “s” sounds, leaving the rest of the vocal open. Similarly, if the vocal has a nasal quality around 1 kHz, create a narrow band there and apply gentle compression (2:1) to smooth it out. This is often more transparent than using a de‑esser alone, because it also reduces dynamic variation in that frequency range.

Bass – Controlling Boom and Clarity

Bass instruments often accumulate low‑mid mud. Set a band from 40–150 Hz and compress with a ratio of 4:1, slow attack (20–30 ms), and fast release (50–80 ms). This tightens the low end without losing the initial finger noise or pick attack. For a more consistent bass part, also set a high band (2 kHz–5 kHz) with light compression to even out the growl frequencies.

Drums – Punch and Resonance Control

On a drum bus, multiband compression can tighten the kick and snare while preserving cymbal sparkle. A low band at 60–120 Hz with high ratio (6:1) and fast attack (5 ms) can keep the kick consistent. A mid band (200–500 Hz) with moderate compression (3:1) can tame boxy snare drums. Be careful with attack times on the high band—too fast can dull the cymbal crashes; use a slow attack (20–30 ms) to let the transient through.

Mastering – Final Polish

In mastering, multiband compression is used with extreme restraint. The goal is not to solve mix problems but to add subtle cohesion. Use 2–3 bands with ratios rarely exceeding 1.5:1 and very low thresholds (1–2 dB of gain reduction). A common approach is to compress the low band slightly (1.5:1, slow attack) to even out bass levels, and a similar gentle touch on the high band (2 kHz+) to smooth out harshness. Always compare your master against a reference to ensure you’re not altering the tonal balance.

Common Pitfalls and Best Practices

  • Over‑compressing: Multiband compression is powerful, but too many bands or too much gain reduction can make your mix sound flat and lifeless. Aim for no more than 3–4 dB of reduction per band, and never compress every band equally—only the ones that need it.
  • Ignoring phase issues at crossover points: Minimum-phase filters can cause phase cancellations around crossover frequencies, especially if you move the compression heavily. To check, solo the summed output with and without the bypass button, or use a linear‑phase mode if available.
  • Compressing frequencies that don’t need it: A common mistake is to use multiband compression on every track. Reserve it for problem sources—if a track sounds great with a single compressor, don’t add a multiband. Unnecessary processing can introduce artifacts.
  • Not using sidechain filters: Many multiband compressors allow you to sidechain specific bands. For example, you can sidechain the low band of a bass synth to the kick drum, creating a clean low‑end separation without affecting the bass’s midrange.
  • Trusting your ears over meters: Meters show gain reduction, but only your ears can tell you if it sounds natural. Always A/B between the processed and unprocessed signal, and listen on multiple systems (headphones, monitors, phone speakers).

Advanced Techniques

Parallel Multiband Compression

Instead of applying multiband compression directly to a track, try routing the signal to a bus with a multiband compressor set aggressively (e.g., 4–6 dB of reduction) and blend it back with the dry track. This adds density and control without sucking the life out of the original. It works exceptionally well on drum buses and vocal tracks.

Sidechaining Specific Bands for Ducking

Use a sidechain trigger (like a kick) to activate compression on only a specific band of another instrument. For example, sidechain the low band of a pad to a kick: every time the kick hits, the low frequencies of the pad duck, creating a pumping low end without affecting the pad’s body or air. This is a common trick in electronic music.

Dynamic EQ vs. Multiband Compression

Both tools can address problem frequencies, but they work differently. A dynamic EQ applies gain reduction only when a specific frequency exceeds a threshold, while a multiband compressor compresses the entire band once the threshold is crossed. For narrow, resonant peaks (e.g., a ringing snare at 400 Hz), a dynamic EQ is often more transparent. For broader frequency imbalances (e.g., a bass that varies wildly in low‑mid energy), multiband compression is better. In some cases, using them together—dynamic EQ for static resonance and multiband compression for dynamic imbalance—gives the best results.

Final Thoughts

Multiband compression is not a cure‑all, but when used judiciously it becomes an indispensable part of any engineer’s toolkit. The key is to treat it as a precision instrument: identify exactly what frequencies are causing trouble, set your bands carefully, and apply only as much compression as needed to solve the problem without destroying the music. Practice on a variety of sources, from vocals to full mixes, and always reference your work on multiple playback systems. With time, you’ll develop the intuition to know when a single‑band compressor will do the job and when only a multiband tool can deliver the control you need.

For further reading, check out Sound On Sound’s guide to multiband compression and the iZotope article on understanding multiband compression. To hear practical examples, watch this video by Produce Like A Pro and experiment with Waves’ free plugin demos.