You've got a vocal that needs presence without harshness, or a drum bus that's punchy in the mids but boomy in the lows. A standard compressor hits everything at once, but multiband compression lets you target specific frequency ranges with different settings. The catch? It's easy to over-process and end up with something that sounds worse than when you started.
Multiband compression works best as a surgical tool for targeted problems, not as a blanket solution for every mix. The key is knowing which frequency splits make sense, how much gain reduction each band actually needs, and when a simpler approach would serve your mix better.
Quick Takeaways
- Start with 3-band splits around 250Hz and 2.5kHz for most sources
- Use different attack/release times for each frequency zone
- Solo each band to hear what you're actually compressing
- Aim for 2-4dB gain reduction max per band for transparent results
- Check your work in bypass - multiband should solve a specific problem, not just sound different
- Consider single-band compression or EQ first for simpler fixes
What Multiband Compression Actually Controls
Multiband compression splits your signal into separate frequency ranges, then applies independent compression to each band before summing them back together. This means you can compress the low mids of a vocal to control boominess while leaving the presence frequencies untouched, or tame harsh drum cymbals without affecting the punch in the lower frequencies.
Each frequency band gets its own threshold, ratio, attack, and release controls. The crossover points determine where one band ends and the next begins. Most multiband compressors use slopes between 12dB and 24dB per octave at these crossover frequencies.
The trade-off comes in complexity and phase relationships. Every time you split a signal into bands, you're introducing phase shifts at the crossover points. Poor crossover design or extreme settings can make your source sound disconnected or artificial, even when the gain reduction meters look reasonable.
Three-Band vs Four-Band: Which Split Works for Your Source
A three-band setup works well for most sources and keeps phase complications manageable. Try crossover points around 250Hz and 2.5kHz. This gives you a low band for controlling boom and muddiness, a mid band covering the fundamental frequencies of most instruments, and a high band for presence and air.
Four-band splits make sense when you need more precise control, especially on full mixes or complex sources like drum buses. A typical four-band setup might use crossovers at 150Hz, 800Hz, and 3kHz. This separates sub-bass, low-mids, upper-mids, and highs into distinct zones.
Here's a practical starting point for different sources:
| Source | Crossover Points | Target Use |
|---|---|---|
| Lead Vocal | 300Hz, 2kHz | Control body without affecting clarity |
| Drum Bus | 150Hz, 800Hz, 3kHz | Balance kick punch, snare snap, cymbal control |
| Bass Guitar | 200Hz, 1kHz | Separate fundamental from harmonic content |
| Full Mix | 120Hz, 500Hz, 2.5kHz | Master bus control for streaming loudness |
| Acoustic Guitar | 250Hz, 2kHz | Manage body resonance and string brightness |
Attack and Release Settings for Each Frequency Zone
Different frequency ranges need different timing settings. Low frequencies carry more energy and change more slowly, so they typically need slower attack times to avoid killing the natural punch. High frequencies move faster and often benefit from quicker response times.
For the low band (under 250Hz), start with attack times between 10-30ms and release times around 100-200ms. This preserves the initial transient while controlling sustained low-end energy. Going too fast in the low band often makes bass and kick drums sound thin or disconnected.
The mid band usually works well with moderate settings: 5-15ms attack and 50-100ms release. This range covers most instrument fundamentals, so you want enough speed to catch level changes without over-pumping.
High frequencies can handle faster compression: 1-5ms attack and 20-50ms release. Cymbals, vocal sibilance, and guitar pick attack live here. Quick response times help control harshness without dulling the source.
How Much Gain Reduction Each Band Actually Needs
The biggest multiband mistake is using too much gain reduction across all bands. If you're seeing 6-8dB of compression on multiple bands, you're probably over-processing. Most transparent multiband work happens with 1-3dB of gain reduction per band, with occasional peaks hitting 4-5dB on problem frequencies.
Watch the gain reduction meters while your track plays. Constant heavy compression across all bands means you should probably use a regular compressor on the whole signal instead. Multiband works best when only one or two bands are working most of the time, with the others catching occasional peaks.
If you find yourself needing heavy compression on the low band, check if your room is making you over-compensate for bass buildup. If the high band is constantly working hard, you might have a harsh recording that needs EQ more than compression.
When Single-Band Compression Works Better
Before reaching for multiband compression, try a regular compressor with a slow attack time. This often handles peak control and glue without the phase complexity of frequency splitting. Many sources that seem to need multiband processing actually respond better to simple compression plus targeted EQ.
A vocal that's boomy and bright might benefit more from a standard compressor set with a 10ms attack and 100ms release, followed by a gentle EQ cut around 400Hz and a small boost around 2-3kHz. This approach keeps the natural phase relationships intact while addressing the same frequency balance issues.
Single-band compression also works better when you need the natural interaction between frequency ranges. A drum bus often sounds more cohesive with one compressor that treats the entire kit as a unit, rather than multiband processing that separates kick, snare, and cymbal compression.
The Solo Band Check That Prevents Over-Processing
Most multiband compressors let you solo individual bands to hear what each frequency range sounds like after processing. Use this feature to catch problems before they affect your final mix.
Solo the low band and listen for thinness or unnatural pumping. The low frequencies should still feel solid and connected to the source. If the low band sounds weak or disconnected, try a slower attack time or less compression.
Solo the high band to check for dulling or artificial smoothness. Cymbals and vocal air should retain their natural character. If everything sounds too controlled or flat, back off the ratio or raise the threshold.
When you solo the mid band, listen for the core character of your source. This band usually contains the fundamental frequencies that define the instrument. Over-compression here makes everything sound lifeless and small.
Common Multiband Mistakes
Avoid these settings that usually backfire:
- Using the same attack/release times across all bands
- Setting crossover points too close together (less than an octave apart)
- Applying heavy compression to all bands simultaneously
- Using multiband to fix problems that EQ would handle better
- Setting thresholds so low that compression never stops
DAW Stock Multiband Tools: Logic, Ableton, and Pro Tools
Logic Pro's Multipressor offers four bands with visual feedback and good crossover controls. Start with the "Vocal" preset and adjust the crossover points to 300Hz and 2kHz. Set gentle ratios (2:1 to 3:1) and watch that you're getting only 2-3dB of gain reduction per band.
Ableton's Multiband Dynamics works well for both compression and expansion. Try the "Master Gentle" preset as a starting point, then adjust the frequency splits and ratios for your source. The below/above threshold controls let you apply different processing above and below your set level.
Pro Tools includes the System 5 MC and Avid Channel EQ with dynamics. The MC offers transparent processing with smooth crossovers. Use moderate ratios and slower attack times than you might expect - Pro Tools multiband tools tend to sound more natural with conservative settings.
For any DAW stock tool, bypass the processing frequently to make sure you're improving the source rather than just making it sound different. The goal is solving a specific problem, not adding generic multiband character.
Crossover Frequency Planning: Where to Split and Why
Choose crossover points based on the natural breaks in your source material, not arbitrary frequency numbers. For vocals, the transition between chest voice and head voice often happens around 250-400Hz, making this a good low crossover point. The presence range starts around 2-3kHz, creating a natural high crossover.
Drum sources benefit from crossovers that separate kick fundamentals from snare fundamentals. Try 150-200Hz to isolate the kick, then 800Hz-1kHz to separate snare body from snare snap and cymbal content.
Avoid placing crossover points right on top of important harmonic content. If your bass guitar's fundamental note is 80Hz, don't put a crossover at 80Hz. Move it to 120Hz or 60Hz to keep the fundamental energy in one band.
Guitar sources often work well with crossovers around the harmonic break points. Try 250Hz to separate body resonance from midrange clarity, and 2kHz to separate warmth from presence and air.
Reference Comparison: Multiband vs Simple Processing
Pick a reference track in your genre and listen specifically to how the elements sit together frequency-wise. Does the vocal sound like it was multiband compressed, or does it have the natural cohesion of simpler processing?
A/B your processed track against the reference, focusing on whether your multiband work helps the source sit in the mix better. If your processing makes individual elements sound more controlled but the overall mix feels less cohesive, try backing off the multiband settings or switching to single-band compression.
Many professional mixes use multiband compression sparingly - often just on the mix bus for final loudness control, or on specific problem sources like overly dynamic vocals or drums recorded in boomy rooms. The reference might sound better because it uses less processing, not more.
To level-match your comparison, use a plugin like Moozix's AI mixing tools to ensure you're comparing loudness-matched versions. Louder often sounds "better" initially, which can make you think heavy multiband processing is working when you're just hearing the volume increase.
Stem Preparation: Multiband Settings for Automix Tools
When preparing stems for online mixing services or AI tools, multiband compression can help create more consistent levels across frequency ranges. But avoid heavy processing that removes the natural dynamics completely.
For vocal stems heading to automix processing, gentle multiband compression with 1-2dB gain reduction can help control resonant frequencies without over-processing. Set conservative crossovers around 300Hz and 2.5kHz with slow attack times to preserve natural vocal dynamics.
Drum stems benefit from light multiband work to balance the relationship between kick, snare, and cymbals before automix processing. Use minimal gain reduction (1-3dB max) and preserve the natural punch with slower attack times on the low band.
Export your stems with multiband processing printed in, but also keep unprocessed versions. This gives the mixing engineer or AI tool options to work with. Include notes about what multiband settings you used and why - this helps the next stage of processing make better decisions.
When to Skip Multiband and Use EQ Instead
Static frequency problems often respond better to EQ than multiband compression. If your vocal is consistently too boomy around 400Hz, a gentle EQ cut fixes the problem without the phase complications of multiband processing.
Multiband compression works best for dynamic frequency problems - issues that come and go depending on the musical content. A singer who gets boomy only on certain notes, or drums that get harsh only when played loudly, are good multiband candidates.
If you find yourself setting very slow attack and release times to avoid pumping artifacts, consider whether a static EQ cut would solve the same problem more simply. Multiband compression that barely moves isn't adding much over careful EQ work.
For mix bus processing, many engineers prefer EQ plus single-band compression over multiband compression. The mix bus often needs gentle glue and level control more than frequency-specific dynamics processing. Save multiband for sources where you really need independent control over different frequency ranges.
Common Questions About Multiband Compression Settings
How many bands should I use for vocal processing?
Three bands work well for most vocals. Use crossovers around 300Hz and 2kHz to separate body, clarity, and presence. More bands add complexity without much benefit unless you're fixing specific resonant frequencies that need isolated control.
Why does my multiband compression make everything sound disconnected?
You're likely using too much gain reduction across multiple bands or setting crossover points too close together. Try gentler settings (1-3dB max per band) and space crossovers at least an octave apart to preserve natural frequency relationships.
Should I use different attack times for each frequency band?
Yes, low frequencies need slower attacks (10-30ms) to preserve punch, while high frequencies can handle faster response (1-5ms) for controlling harshness. Mid frequencies typically work well with moderate settings around 5-15ms attack time.
When should I use single-band compression instead of multiband?
Use single-band when you need natural frequency interaction or when the problem is consistent rather than frequency-specific. Many sources sound more cohesive with simple compression plus EQ than with multiband processing.
How do I know if my multiband crossover points are right?
Solo each band and listen for natural-sounding frequency content. Avoid placing crossovers directly on fundamental frequencies of your source. The low band should sound solid, the mid band should contain the core character, and the high band should have clear detail without harshness.
Can I use multiband compression on my mix bus?
Yes, but use very gentle settings - typically 1-2dB gain reduction maximum per band. Mix bus multiband works best for final loudness control rather than corrective processing. Many mixes benefit more from EQ plus single-band compression on the bus.
Hear what these choices do to your own song.
Upload stems or a finished track, choose a reference direction, and compare a private Moozix mix before you export anything.
