Multiband compression splits your signal into frequency bands and compresses each band independently. Use it when your dynamic problem is frequency-specific — the low end is too uneven but the top end is fine, for example. Don't use it as a default compressor; a single-band compressor is the right tool for most mixing situations. Multiband earns its place on the master bus for final tonal balance control, on bass instruments with specific frequency problems, and on vocal buses where low-mid buildup needs frequency-targeted control.
Updated May 2026 by the MusicProductionWiki Team
Multiband compression is one of the most misused tools in music production. It's powerful, it's present in almost every DAW and mastering chain, and it's routinely applied where it doesn't belong — turning mixes that just needed a single-band compressor and some EQ into over-processed, frequency-fragmented signals that sound dense without feeling alive.
The concept is seductive: independent control over every frequency range sounds like more control, and more control sounds like a better result. In practice, more control means more ways to make something sound wrong. Multiband compression requires a clear understanding of what the tool actually does, when it solves a problem that a standard compressor can't, and — critically — what it sounds like when it's being applied where it shouldn't be.
This guide covers all of it: how multiband compression works at the technical level, the correct use case for each application context (mixing vs. mastering vs. individual sources), how to set every parameter, the mistakes that are most damaging to a mix, and specific settings for the instruments and contexts where multiband compression genuinely earns its place.
How Multiband Compression Works
A standard (wideband) compressor processes your entire audio signal as one unit. When the signal crosses the threshold, every frequency in the signal gets compressed equally — the low end, the midrange, and the high end all receive the same gain reduction at the same time.
A multiband compressor first splits the signal into separate frequency bands using crossover filters. Typically three to five bands, though some processors offer more. Each band then passes through its own independent compressor with its own threshold, ratio, attack, release, and makeup gain. After processing, the bands are summed back together into a single output signal.
The critical consequence: a loud transient in the low end — a kick drum hit, a bass note peak — only triggers the low band's compressor. It has no effect on the high band's compressor. A sibilant peak in the high end only triggers the high band. This frequency-isolated compression is what makes multiband a specialist tool rather than a universal upgrade over single-band compression.
Multiband compressor signal flow: crossover splits the signal, each band is compressed independently, then bands are summed at the output.
There's an important difference between two types of crossover filters used in multiband compressors: minimum phase (standard) and linear phase. Standard crossovers are more natural-sounding and have faster transient response, but introduce phase shift at crossover points that can affect stereo imaging and transient integrity. Linear phase crossover filters split the signal into bands without introducing phase shift — this preserves the phase relationships of the original signal. Use linear phase mode for mastering and mix bus applications where phase accuracy matters most; use minimum phase for mixing individual sources where latency and transient feel are more important.
When to Use Multiband Compression (And When Not To)
The most important skill with multiband compression is knowing when to reach for it — and when to put it back on the shelf. The majority of mixing situations do not require multiband compression. That statement will feel counterintuitive if you've been told that multiband is a professional upgrade over single-band compression. It isn't. It's a different tool for a different job.
Use multiband compression when the dynamic problem is frequency-specific. That means: one frequency range is too dynamic while others are fine. Classic scenarios include:
- A bass guitar with a boomy, uncontrolled low end but well-behaved mids and highs
- A vocal bus where low-mid resonance (200–400 Hz) is inconsistent but the top end is stable
- A master bus where the bass is causing the limiter to pump while the rest of the mix is controlled
- A drum bus where the low end of the kick is the only frequency range causing problems
- A mix where hi-hat bleed in the upper frequencies is creating uneven compression when a single-band compressor is applied
Do not use multiband compression when:
- The dynamic problem affects all frequencies roughly equally — a single-band compressor is the correct tool
- You want tonal shaping — use EQ instead
- You can't clearly identify which specific frequency band has the problem — if you can't hear it, you can't fix it with multiband without making everything worse
- You're using it because it's there — this is the most common and damaging reason
Understanding how compression works at a fundamental level is a prerequisite. If you haven't internalized what threshold, ratio, attack, and release do on a single-band compressor, you're not ready to set those parameters correctly across four independent bands. The mistakes compound dramatically.
Setting Crossover Frequencies
The crossover frequencies determine where one band ends and the next begins. Setting them correctly is more important than any individual compressor parameter — a poorly placed crossover can mean the compressor is working on the wrong frequency content entirely.
A common starting point for a 4-band multiband on a full mix:
| Band | Frequency Range | Primary Content | Typical Dynamic Behavior |
|---|---|---|---|
| Low | 20 Hz – 120 Hz | Sub-bass, kick body, bass fundamentals | Slow-moving, high energy peaks from bass notes and kick hits |
| Low-Mid | 120 Hz – 800 Hz | Bass harmonics, guitar body, vocal chest tones, snare body | Medium speed, often muddy or resonant in problem mixes |
| High-Mid | 800 Hz – 5 kHz | Vocal presence, guitar bite, snare crack, piano attack | Fastest transient content, most sensitive to over-compression |
| High | 5 kHz – 20 kHz | Air, sibilance, cymbals, hi-hat shimmer | Moderate energy, sibilance peaks can be severe |
These are starting points — the optimal crossover points depend entirely on the specific material. The most important rule: set crossovers at natural frequency separations in the content, not at arbitrary round numbers.
To find the right crossover point for a problem band, use a spectrum analyzer while the track plays. Identify where the energy you want to control lives. Set the lower crossover of that band just below that energy cluster and the upper crossover just above it. If you're dealing with a boomy bass around 80 Hz, your low band doesn't need to extend to 200 Hz — tighten it to 20–120 Hz and let the low-mid band handle the 120–500 Hz region where the muddiness lives separately.
The frequency EQ cheat sheet is a useful reference for understanding which instruments live at which frequency ranges before you commit to crossover positions.
Setting Attack and Release Times Per Band
One of the most technically demanding aspects of multiband compression is that attack and release must be set independently for each band — and those settings need to reflect the different transient behavior at different frequencies. High frequencies move fast. Low frequencies move slowly. Setting every band to the same attack and release is one of the most common signs that someone doesn't understand what they're doing with the tool.
Low band (20–120 Hz): Slow attack (30–80 ms) to preserve kick drum and bass transients. The attack needs to be slow enough to let the initial transient punch through before the compressor clamps down. Release should be matched to the tempo — too slow and the compressor doesn't reset between notes, causing sustained gain reduction; too fast and it introduces low-frequency pumping artifacts. Typically 100–300 ms release depending on tempo.
Low-mid band (120–800 Hz): Moderate attack (15–40 ms), release 150–250 ms. This range contains a lot of sustained tonal content — the compressor should react to note-to-note level variation rather than individual transients. If the attack is too fast in this range, you'll kill the body of bass notes and the warmth of guitars.
High-mid band (800 Hz–5 kHz): Moderate-to-fast attack (10–25 ms), release 80–180 ms. This is the most sensitive perceptual range — over-compression here makes everything sound flat and lifeless. Be conservative with gain reduction in this band. The attack should catch peaks without destroying the snap of snare hits and the articulation of vocal consonants.
High band (5 kHz–20 kHz): Fast attack (5–15 ms), short release (40–100 ms). This range contains sibilance and cymbal transients that need faster control. Release should be fast enough to avoid pumping between transients, which in this frequency range is highly audible as a "breathing" artifact in the hi-hat and air frequencies.
The key principle: faster transients in a band need faster attack times, and lower frequencies need slower settings in both directions. The crossover frequencies you've set inform what transient character lives in each band — always set attack and release in the context of what you're actually hearing in that band.
For a deeper understanding of how attack and release affect compression character, the guide on compression on drums covers the transient mechanics in detail.
Gain Reduction Targets and Threshold Setting
How much gain reduction is appropriate on each band depends heavily on the application context — mixing an individual source, mixing a bus, or mastering.
Individual source mixing (bass guitar, vocal, single instrument): 3–6 dB on the problematic band is reasonable when the problem is clear and specific. If you're compressing the low end of a bass guitar to control note-to-note variation in the sub frequencies, 4–5 dB of reduction in the low band is justified. The other bands may need little to no compression if they're already controlled.
Bus mixing (drum bus, vocal bus, instrument bus): 1–4 dB per band. This is a more delicate context because you're processing a blend of multiple sources. Too much reduction in any one band starts to pull the tonal balance in a direction that wasn't in the original arrangement. Set thresholds carefully — you want the compressor engaging only on peaks, not holding down the average level.
Mastering: 1–2 dB per band maximum for gentle glue and tonal balance adjustment. More than that and multiband mastering starts to damage the mix rather than enhance it. If you're seeing 8+ dB of gain reduction on any band in a mastering context, the problem is more fundamental than multiband can fix without causing damage. The mix needs to go back to the mixing engineer, or the approach needs to change entirely.
A useful discipline: set your threshold by watching the gain reduction meter while the loudest, most dynamic section of the track plays. The compressor should engage during peaks but sit idle or near-idle during quieter passages. If the compressor is engaged 100% of the time with constant heavy reduction, you're using it as a tonal shaper (a job for EQ) rather than a dynamic controller.
Understanding compression ratio mechanics helps set appropriate ratios per band. Low ratios (1.5:1 to 3:1) are appropriate for transparent dynamic control in a mastering context; higher ratios (4:1 to 8:1) are acceptable on individual sources with severe frequency-specific dynamic problems.
Multiband Compression vs. Dynamic EQ: Which to Choose
Multiband compression and dynamic EQ are frequently confused with each other because they both apply frequency-specific dynamic processing. They work differently and have different sonic signatures — choosing between them is a meaningful decision, not an arbitrary preference.
Multiband compression applies compression to a frequency band. The full compressor control set applies — threshold, ratio, attack, release, knee, makeup gain — and the band is compressed relative to a fixed threshold. The behavior is compressor-like: the band's level is limited relative to a ceiling, and the dynamic range of that band is narrowed.
Dynamic EQ applies EQ gain changes that are triggered by the signal level. The EQ band boosts or cuts in response to the signal, typically using a sidechain detector. The behavior is EQ-like with dynamic response: the frequency content is attenuated when it exceeds a threshold, but the underlying tonal relationship of the signal is preserved more naturally.
Dynamic EQ is generally more surgical and less colored. Multiband compression is more aggressive and can shape character as well as dynamics. The distinction matters most at the extremes:
- For controlling sibilance on a vocal, a dynamic EQ is typically the cleaner, more transparent choice — it removes exactly the frequency content that's too loud without affecting the compressor's character on surrounding material.
- For controlling a bass guitar's low end that's wildly inconsistent in level, a multiband compressor with a slow-attack low band gives you full compressor control — threshold, ratio, attack, release — which is the right tool for a level-control problem.
- For mastering, dynamic EQ tends to be more appropriate for subtle frequency balance adjustments that respond to the signal; multiband compression is more appropriate when a specific frequency range's dynamic range genuinely needs narrowing.
The full technical comparison is covered in the dynamic EQ vs. multiband compression guide, which goes deeper into the signal path differences and gives side-by-side examples.
The Best Multiband Compressor Plugins
The plugin you choose affects both workflow and sound quality. The differences between top-tier multiband compressor plugins are real and worth understanding before committing to a tool.
FabFilter Pro-MB is widely considered the most transparent and flexible multiband compressor available. Its linear phase mode and dynamic mode switching (which allows upward compression and expansion in addition to downward compression) make it suitable for both mastering and mixing. The interface is exceptionally clear — you can see exactly which frequency ranges are being processed and by how much. This is the standard recommendation for mastering applications. You can read the detailed comparison of FabFilter's EQ line in the FabFilter Pro-Q 3 review to understand their approach to transparency.
Waves C6 is an industry-standard mixing multiband with a straightforward interface. It's been on professional mixing consoles and in professional DAW sessions for over a decade. Less flexible than Pro-MB but extremely well-understood by engineers who've worked with it for years. Good for bus processing where a more character-forward sound is acceptable.
iZotope Ozone Dynamics module (part of Ozone 11 and Ozone 12) is excellent for mastering contexts. It integrates with the rest of Ozone's processing chain and includes intelligent AI-assisted threshold suggestions that serve as useful starting points. The iZotope Ozone 12 review covers the full feature set in detail.
iZotope Neutron also includes a multiband compressor that integrates with track assistant and masking detection — useful for mixing individual tracks where frequency masking between instruments is part of the problem being addressed. See the iZotope Neutron guide for mixing-specific workflows.
DAW-native options: Logic Pro's Multipressor and Ableton Live's Multiband Dynamics are both competent and free to use within their respective environments. They lack the fine-grained visual feedback of FabFilter Pro-MB but are entirely functional for learning and for situations where a third-party plugin isn't available. For Hip-Hop and electronic production workflows in Ableton, the native Multiband Dynamics often appears in drum bus chains — covered in the broader best mixing plugins guide.
The Most Common Multiband Compression Mistakes
Knowing what goes wrong with multiband compression is as important as knowing how to use it correctly. These are the mistakes that appear most often in intermediate-level mixes and masters, and each one has a specific cause and fix.
Mistake 1: Using multiband compression as a default compressor. Multiband compression is not a premium version of single-band compression. Applying it to every track because it offers "more control" is the fastest way to make a mix sound over-processed and fragmented. Every track does not need frequency-specific dynamic control. Most tracks need a well-set single-band compressor — sometimes not even that.
Mistake 2: Setting all bands to the same attack and release. This is a direct indicator that the producer doesn't understand the tool. Each band contains different frequency content with different transient characteristics. Identical attack and release settings across all bands ignore this completely and usually result in unnatural-sounding compression in at least two of the bands.
Mistake 3: Using too much gain reduction, particularly on the low band. Heavy low-band compression (8+ dB) on a mix or master removes the natural dynamic relationship between the bass and the rest of the mix. The low end starts to feel separate from the track — it's there, but it doesn't breathe with the music. 1–3 dB is the ceiling for transparent low-band compression on a mix bus.
Mistake 4: Placing crossovers at arbitrary round numbers. Setting crossovers at 100 Hz, 1 kHz, and 10 kHz because they look clean on a display is not analysis — it's guessing. The correct crossover positions depend on where the frequency content you're trying to control actually lives in the signal. Use a spectrum analyzer. Set crossovers based on the material, not aesthetics.
Mistake 5: Ignoring phase artifacts at crossover points. Standard (minimum phase) crossovers introduce phase shift at crossover frequencies. This can cause comb-filtering-like effects at the crossover point when bands are summed, particularly audible on transient-heavy content. In mastering contexts, always audition the linear phase option to compare. On individual sources in a mix, minimum phase is often fine, but be aware of what's happening at the crossover.
Mistake 6: Over-multiband-compressing and calling it "mastering." A mix that's been over-multiband-compressed often sounds "all frequency, no air" — dense and controlled but lacking life and movement. The correct amount of multiband processing in a mastering chain is typically much less than beginners expect. Many excellent masters use no multiband at all; where it's used, the total gain reduction across all bands is rarely more than 2–3 dB at any moment.
Mistake 7: Not A/B testing against bypass. Multiband compression that isn't improving the mix should not be in the chain. Always A/B the processed and unprocessed signal at matched levels. Gain-matched A/B testing reveals quickly whether what you've applied is genuinely solving a problem or just adding complexity. If you can't clearly hear an improvement in bypass, the plugin doesn't belong in the chain.
Mistake 8: Using multiband to fix a mixing problem at the mastering stage. If the low end is pumping because the kick and bass weren't properly balanced during mixing, multiband mastering compression is a workaround, not a fix. It can partially compensate, but it will always compromise something in the process. The correct approach is to return the mix to the mixing stage and address the problem at its source. Mastering multiband is for refinement, not rescue.
Practical Exercises
Single-Band First, Multiband Second
Take any bus in your current project — drums, vocals, or bass — and apply only a single-band compressor. Spend 15 minutes setting threshold, ratio, attack, and release until you're satisfied with the result. Then, and only then, add a multiband compressor and try to solve one specific frequency-range problem you can still hear. If you can't identify a frequency-specific problem remaining after the single-band compressor, don't add the multiband at all.
Crossover Placement by Ear
Load a multiband compressor on a bass guitar track. Before enabling any compression, use a spectrum analyzer to identify where the bass energy peaks — find the fundamental frequency range and the harmonic range above it. Set your crossover between the low band and low-mid band directly at the frequency where the character of the bass changes from fundamental to harmonic content. Now enable the low band's compressor only and dial in 3–4 dB of gain reduction, comparing A/B against bypass until the low end is controlled without losing the note definition above the crossover.
Mastering Multiband at 1 dB Per Band
Take a finished mix and apply FabFilter Pro-MB (or Logic's Multipressor) in linear phase mode on the master bus. Set a strict rule: no band may exceed 1 dB of gain reduction at any point. Work within that constraint to address one specific tonal-dynamic problem in the mix — for example, controlling a low end that pushes the limiter harder than the rest of the mix. Document exactly what problem you identified, how you positioned the crossovers, and what the gain reduction meters showed during the loudest section. Compare the final result against the unprocessed mix at gain-matched levels and assess whether the 1 dB constraint forced better decision-making than you would have applied without it.