Multiband compression splits your audio into frequency bands and compresses each one on its own, so you can tame a boomy low end or a harsh top without squashing everything in between. The short version of how to use multiband compression: only reach for it when a normal compressor would over-react to one part of the spectrum, set gentle ratios, and move one band at a time by ear. It is a surgical tool, not a default — most tracks are happier with plain EQ and a single-band compressor. This guide covers when it earns its place, sensible starting settings, and the moves that keep it transparent.
Violet Recording is reader-supported — we may earn a commission from links on this page, at no extra cost to you.
What multiband compression actually does
A standard compressor looks at the whole signal and reacts to whatever is loudest. If a vocal has a booming low-mid resonance, the compressor clamps down every time that boom hits — pulling the entire vocal down with it, including the parts you liked. Multiband compression solves that by dividing the signal into two, three, four, or more bands using crossover points, then giving each band its own threshold, ratio, attack, release, and make-up gain. Now the boom triggers compression only in the low band, and the rest of the vocal sails through untouched.
Think of it as a compressor and a dynamic EQ having a baby. It reacts to level like a compressor, but only within a chosen frequency range like an EQ. That makes it perfect for problems that are both tonal and dynamic — issues that come and go rather than sitting there constantly. If a problem is constant, static EQ is simpler and cleaner. If it only shows up on loud notes or certain words, multiband is the right call.
When to use it — and when to leave it alone
The most common good uses are a low end that pumps or booms only on certain notes, sibilance or harshness that spikes on loud passages, and glue on a full mix or master where the low end needs tighter control than the mids. It also shines on a bass guitar whose low strings jump out, or a drum bus where the kick’s sub swings wildly in level.
Leave it alone when a plain high-pass filter or a single EQ cut would fix the issue. If a track has a constant 250 Hz mud, just cut 250 Hz — don’t spend a band chasing it dynamically. Over-using multiband is how mixes end up lifeless: every band gets clamped, transients vanish, and the whole thing sounds flat and small. Reach for the simplest tool that solves the problem, and keep the multiband for jobs a broadband compressor genuinely can’t handle.
How to use multiband compression step by step
Here is a reliable order of operations. Treat every number below as a starting point you adjust by ear, not a rule.
1. Set your crossover points first. Fewer bands are easier to control, so start with two or three. On a full mix, common crossovers land somewhere around 100–120 Hz (splitting sub/low from the body) and around 2.5–5 kHz (splitting the mids from the highs). Nudge these until each band captures the problem you’re targeting and nothing more.
2. Find the offending band solo. Most plugins let you audition one band at a time. Solo each band and listen for where the boom, honk, or harshness lives. That’s the band you’ll compress; the others often need no processing at all.
3. Start gentle on ratio and threshold. A 2:1 to 3:1 ratio is plenty for most musical multiband work. Lower the threshold until the meter shows just 1–3 dB of gain reduction on the loudest hits. If you’re pulling 6 dB or more out of a band, either the source needs EQ first or you’re asking too much of one band.
4. Set attack and release to the material. On low bands, use slower attack (roughly 30–100 ms) and slower release so you don’t kill the punch of a kick or bass note. On high bands taming sibilance or harshness, faster attack (a few milliseconds) and faster release catch the spikes without dulling everything. Watch that the release lets the band recover between hits rather than pumping.
5. Compensate the gain and A/B honestly. Compression lowers level, so use make-up gain to match the band’s output, then bypass and compare at equal loudness. If the “after” only sounds better because it’s louder, you haven’t improved anything. Trust the version that sounds more controlled at the same volume.
Sensible starting settings for common jobs
For a full mix or master, aim for tiny, mostly-invisible reduction — 1–2 dB per band on peaks, low ratios around 1.5:1 to 2:1, and crossovers that let the low band tighten the bass while the mids and highs stay open. If you are pushing loudness toward a streaming target, remember that most platforms normalise around −14 LUFS integrated as a rough reference, so you don’t need to crush every band to compete. Our free LUFS and loudness tools make it easy to check where you land.
For a vocal with a dynamic low-mid boom, one low band with a crossover under ~300 Hz, a 3:1 ratio, medium attack and release, catching a couple of dB on the boomy words, usually does it. For bass guitar, a low band that steadies the fundamental (roughly 60–120 Hz) keeps the level even without dulling the string attack in the upper band. For de-essing, a high band focused around 5–9 kHz with fast attack works, though a dedicated de-esser is often simpler for that one job.
Where multiband fits in the chain
Order matters. On a channel, corrective EQ usually comes first — remove constant problems statically so the multiband only has to handle what’s genuinely dynamic. Then place the multiband, then any broadband compression or tone shaping after it. On a master, multiband typically sits after any broad glue compression and before the final limiter, doing gentle spectral balancing rather than heavy lifting. If you’re new to signal flow, our guide to EQ and compression fundamentals and the walkthrough on how to mix vocals are worth reading alongside this.
Most DAWs ship with a capable stock multiband compressor, and stock does the job on the large majority of tracks — don’t assume you need to buy anything. That said, when you’re doing detailed mastering work and want precise, low-artifact bands with clean metering and dynamic-EQ flexibility, a dedicated third-party multiband can be a genuine upgrade over basic stock plugins. If you’d rather hand the final polish to an engine that handles multiband balancing for you, an online mastering service like LANDR is a fair option to compare. For more on that decision, see what mastering actually involves, how loud your master should be, and our roundup of online mastering services.
Common mistakes that make it sound worse
The biggest one is using too many bands and too much reduction, which drains the life out of a mix. The second is chasing a constant tonal problem with dynamics when a static EQ cut is faster and cleaner. Third is ignoring make-up gain and fooling yourself with a louder A/B. Fourth is crossovers set in the wrong place, so a band grabs energy you wanted to keep — always solo and listen. Finally, don’t multiband-compress on autopilot on every bus; ask each time whether a plain compressor or EQ would do the same job with fewer side effects. Restraint is the whole skill here.
Frequently asked questions
Is multiband compression the same as a dynamic EQ?
They overlap but aren’t identical. A dynamic EQ applies a moving EQ curve that only engages when a frequency gets loud, and it tends to sound more transparent for surgical moves. Multiband compression splits the signal at hard crossovers and compresses whole bands, which is better for glue and level control across a range. For a single narrow problem, dynamic EQ is often cleaner; for broader dynamic balance, multiband wins.
How many bands should I use?
As few as solve the problem. Two or three bands cover most mixing and mastering needs and are far easier to keep transparent. More bands mean more crossover phase interaction and more chances to over-process. Only add a band when you can point to a specific issue it’s fixing.
Should I use multiband compression on every track?
No. Most individual tracks are better served by EQ and a single-band compressor. Save multiband for sources with a genuinely dynamic frequency problem, for buses, and for gentle spectral control on the master. Used everywhere, it flattens your mix; used sparingly, it’s a precision fix.
Want the wider picture? Start at the mixing and mastering hub and the beginner’s guide to mixing your first song to see where this technique fits in a full workflow.
