Multiband vs Wideband Compression Mastering: 3 Key Facts

Use wideband compression for glue. Use multiband when a specific frequency range is misbehaving and EQ alone won't fix it. That's the whole answer. Everything below is how to actually do it.

We've spent a lot of time watching mastering engineers reach for multiband plugins by default, treating them like a Swiss Army knife when a scalpel would do. The result is often a mix that sounds controlled but lifeless, the dynamics squeezed out of frequency ranges that didn't need touching.

This guide covers the real difference, the real settings, which genres actually benefit from multiband, and how to combine both in a serial chain without wrecking your phase coherence.

What's Actually Happening Inside Each Compressor?

A wideband compressor sees one signal. When the input exceeds the threshold, it reduces gain across everything. Kick hits hard at 80Hz? Your 8kHz air comes down with it. That's not always a problem. In fact, on a well-mixed track, it's often satisfying. The whole mix breathes together. That's what mastering engineers mean by "glue."

Multiband splits the signal with crossover filters before compression. Each band has its own threshold, ratio, attack, and release. A 200Hz boom doesn't touch the 4kHz presence. A bright synth stab doesn't trigger low-end gain reduction. The bands process independently, then recombine.

The Phase Problem Nobody Talks About Enough

Here's what most content skips over. Crossover filters introduce phase shift unless you're using linear phase mode. Dynamic (minimum phase) filters are faster and don't pre-ring, but they rotate phase at the crossover points. Linear phase filters preserve phase relationships but add latency and can cause pre-ringing on sharp transients.

In practice: use dynamic (minimum phase) mode for anything with heavy transients, like drums or acoustic instruments. Use linear phase mode for more sustained material, like pads, strings, or when the mix already has crossover filtering baked in from the session. FabFilter Pro-MB gives you both modes. Switch between them and A/B the transient attack. The difference on a snare crack is not subtle.

We ran a dense mix with a heavy kick through Pro-MB in linear phase mode. The sub tightened up nicely. Then we noticed the snare had lost its initial crack, smeared slightly by pre-ringing. Switched to dynamic mode. The snare came back. The sub was slightly less controlled but still acceptable. Phase setting changed the entire character of the transient response.

When Should You Actually Use Multiband?

Most mastering engineers we trust reach for multiband in three situations. Not six. Not ten. Three.

First: a bass note or low-mid frequency is intermittently poking above the rest of the mix. EQ would pull it down permanently. Multiband compression only pulls it down when it exceeds threshold. Smarter.

Second: the mix has high-frequency harshness on certain transients, not consistently, just on hard consonants or bright hits. A high-band compressor with a fast attack and 2:1 ratio catches those moments without dulling the overall air.

Third: genre-specific shaping in EDM, bass music, or anything where the low end needs to behave independently of the top. The pumping low end characteristic of big-room EDM often comes from a low-band compressor set to 4:1 or higher with a release of 80-100ms synced loosely to tempo.

Outside those three situations? Wideband is usually the better call.

The Scenarios Where Wideband Wins

Rock. Acoustic. Jazz. Orchestral. Anything where the mix is meant to feel like one cohesive event. Wideband compression preserves the natural frequency relationships. When the kick hits, the mix responds as a whole, and that response sounds like music.

We mastered a three-piece rock record with wideband only on the bus, a 1.8:1 ratio, 40ms attack, 150ms release, 3dB of gain reduction at peaks. The mix breathed perfectly. Nothing pumped. Nothing lost its edge. We tried multiband on the same track. The bass band tightened up at the cost of the natural relationship between the kick and the guitar low-mids. Technically cleaner. Emotionally duller. We went back to wideband.

What Settings Actually Work?

Standard mastering compression runs 1.5:1 to 3:1. Gain reduction stays at 3-5dB maximum. Go harder than that and you're not mastering anymore, you're producing.

Wideband Settings for Glue

Attack: 30-50ms. You want transients to pass through before gain reduction kicks in. Release: 100-200ms, longer on slower material. Ratio: 1.5:1 to 2:1. Knee: soft. Start with 3dB of gain reduction and back off if it feels tight.

On a pop or R&B master, we'll often set the threshold so the compressor only engages on peaks, sitting mostly at 0-1dB of reduction with occasional 2-3dB catches on the loudest chorus hits. Subtle. Almost invisible. But the track holds together through a playlist better than an uncompressed version.

Multiband Settings for Surgical Fixes

For a typical 3-band mastering setup, try crossovers at 150Hz and 5kHz. That keeps the kick and bass in the low band, the critical midrange in the mid band, and presence plus air in the high band.

Low band: 1.6:1 ratio, 10ms attack, 100ms release. You want to catch bass peaks without killing the kick's initial crack. Mid band: 1.2:1 ratio, 30ms attack, 150ms release. Light touch here. The midrange is where the most audible artifacts appear if you over-compress. High band: 1.1:1 to 1.3:1, 5ms attack, 80ms release. Barely touching it.

For a 4-band setup, crossovers at 80Hz, 250Hz, and 5kHz. The extra split lets you treat the sub-bass (below 80Hz) separately from the upper bass (80-250Hz). Useful on material with inconsistent sub content, particularly hip-hop or bass-heavy electronic music where the sub lives in a narrow range.

Common crossover mistakes: setting the low/mid split too high (above 300Hz puts guitar fundamentals in the low band, which fights your intent), and setting the mid/high split too low (below 4kHz compresses too much of the presence range).

Genre-Specific Guidance: What We Actually Use

EDM and Bass Music

Multiband shines here. The characteristic pumping low end in big-room house comes directly from aggressive low-band compression. Set the low band to 4:1, threshold low enough to catch most of the kick, release at 80-100ms. The kick releases the compressor in rhythm. The bass pumps with the track. That's intentional. It's not a bug.

For neuro or tech production with detailed mid-frequency content, a tight mid-band compressor at 1.8:1 with a fast 8ms attack can control the aggressive mid transients without killing the textural detail that makes that music work.

Hip-Hop and Trap

The 808 is the challenge. An 808 with a lot of sub content will trigger wideband compression hard and pull everything else down with it. Multiband with a low-band split at 100-120Hz, ratio of 2:1, attack of 15ms and release of 200ms controls the 808 without touching the hi-hats or vocal air. We love this setup for trap masters. It tightens the low end without making the mix feel claustrophobic.

Rock and Metal

Mostly wideband. Metal specifically benefits from the cohesive response of a single compressor catching the whole mix. The guitars, kick, and bass all move together. Multiband on metal tends to separate the frequency ranges in a way that sounds clinical. The aggression gets processed out of it. Frustrating, because the instinct is to "fix" the low end, but the mud in a metal master is often intentional density.

Classical and Acoustic

Skip multiband entirely. The natural frequency relationships between instruments are the whole point. Any independent band compression risks destroying that balance. Wideband at the lightest possible setting, 1.5:1 maximum, or no compression at all beyond peak limiting.

Can You Use Both? The Serial Chain Approach

Yes. And on complex material, we often do.

The setup: wideband compressor first in the chain, doing the heavy lifting for glue and overall dynamic control. 3-5dB of gain reduction. Then multiband after it, set much lighter, catching specific frequency problems the wideband introduced or left unresolved.

We used this on a pop-electronic hybrid record. The wideband (a Neve 33609 emulation, about 2.5:1 with 4dB reduction) glued the mix beautifully but left the low-mids slightly bloated. Added iZotope Ozone's multiband after it. Low-mid band set to 1.3:1, barely compressing, just catching the 250-400Hz buildup when the chorus hit. Result: the glue held, the bloom resolved, and we didn't have to EQ out frequencies that were only intermittently problematic.

The order matters. Multiband first sounds different. It treats frequency problems before the wideband sees the signal, which changes the wideband's response. Both orders work. They just work differently. Try wideband-first as your default.

Tools Worth Having in Your Signal Chain

For multiband mastering, FabFilter Pro-MB is the standard. The dynamic/linear phase toggle, the interface clarity, and the M/S processing mode make it the most flexible option in most mastering scenarios. For integrated suites, iZotope Ozone 11 includes both a wideband compressor module and a multiband compressor, so you can build the serial chain inside one plugin and bypass each stage independently for A/B comparisons.

For wideband work, the UAD SSL G-Master Buss Compressor and the Vertigo VSC-2 emulation are both excellent for glue. The Waves SSL G-Master is a more budget-friendly option if you catch it on sale.

For learning the concepts interactively, the Mastering the Mix blog has some of the clearest practical documentation on crossover points and phase coherence. The iZotope Learning Hub is worth bookmarking for free video walkthroughs of their Ozone modules. FabFilter's Pro-MB manual is unusually readable for a plugin manual and covers phase mode tradeoffs clearly.

Also useful: and

Summary

Wideband compression is your default mastering tool. It glues the mix, preserves natural frequency relationships, and sounds musical. Multiband is a problem-solving tool for specific frequency-domain issues that EQ can't fix dynamically. Use a 3-band setup with crossovers around 150Hz and 5kHz as a starting point. Keep ratios at 1.6:1 or below in most bands. Watch your phase mode: dynamic for transient-heavy material, linear for sustained content. And on complex mixes, serial compression (wideband then multiband) gives you the best of both without committing to either extreme.