resonance control on the mix bus

the mix that sounds fine until you listen for ten minutes

you finish a mix. everything sounds balanced. the vocal sits well, the drums have punch, the bass fills the low end. then you listen for ten minutes and something starts to bother you. a tightness in the upper midrange. a fatigue that builds. you turn the volume down, and the problem gets better, but the mix loses its energy. you have been here.

this is mix bus resonance. it is not one track’s fault. it is the accumulation of spectral energy across every source that shares the 2-5 kHz range.

mix bus resonance control is one of the most misunderstood tools in the mixing and mastering chain. used well, it is invisible: the mix sounds the same, but more comfortable. used poorly, it strips the life out of everything it touches.

how resonances compound

a single vocal with a 3 dB peak at 3 kHz is barely noticeable. add an acoustic guitar with its own 2 dB peak in the same range. layer in cymbals with natural brightness at 3-5 kHz. now the bus has a cumulative energy build-up in the range where your ears are most sensitive.[^1]

this is not constructive interference in the physics sense. the tracks are not phase-coherent at specific frequencies. what happens is spectral energy accumulation: the power spectral densities of uncorrelated sources add together.

the key point for mixing: each source contributes its own energy at 3 kHz. three tracks with mild peaks in the same range do not triple the problem, but they do push the combined energy several dB higher than any individual source. your ears weight the 2-5 kHz range heavily, so even a few dB of accumulation in this range sounds noticeably harsh.

fix the sources first

the most effective mix bus resonance control happens before the bus.

if a vocal has a harsh resonance at 3 kHz, fix it on the vocal channel. if an acoustic guitar has a boxy ring at 2.5 kHz, notch it on the guitar channel. if cymbals are overly bright, treat them on the drum bus or individual channels.

three reasons this is better than bus processing:

1. precision: you can be aggressive on one source without affecting the others. a 5 dB notch on a vocal does not touch the snare
2. specificity: you know what you are fixing. on the bus, a 3 kHz peak could be vocal, guitar, synth, or all three. on a channel, you know exactly what is wrong
3. headroom: fixing problems at the source means less energy arrives at the bus. the bus resonance may disappear entirely without any bus processing

gentle bus processing

after source-level treatment, some harshness may remain. this is the residual build-up that only appears when all tracks play together. this is where bus resonance control earns its place.

the rules are different from channel processing. on a channel, you are treating one source. on the bus, you are treating everything: the punchy kick, the warm bass, the airy vocal high end, the stereo reverb tails. heavy settings damage all of these simultaneously.

settings that work

• depth: 1-2 dB maximum reduction. if you need more than 2 dB on the bus, something is wrong at the source level
• frequency focus: 2-5 kHz. do not let bus processing act on the low end (it will thin the bass) or the extreme highs (it will dull the air)
• sensitivity/threshold: set high enough that the processor only catches the worst peaks. quiet passages should pass through untouched
• speed: moderate to slow attack and release. fast settings on a full mix can cause pumping artifacts

mid/side processing for mastering

in mastering, you receive a finished stereo mix. you cannot fix individual sources anymore. a harsh vocal buried in a full mix is much harder to target than the same vocal soloed on its channel.

mid/side processing helps. by splitting the stereo signal into mid (center) and side (stereo difference) components, you can target center-panned sources independently.

vocals sit almost entirely in the mid channel. so do kick drums, bass, and snare. hard-panned guitars, stereo reverb tails, and wide synths live primarily in the side channel.

this means you can apply resonance suppression to the mid channel’s 2-5 kHz range without affecting the stereo width of guitars or the spaciousness of reverb.

practical mid/side approach

• process the mid channel only (or with significantly more depth than the side)
• focus on 2-5 kHz where vocal harshness concentrates
• use even gentler settings than standard bus processing (0.5-1.5 dB max)
• check mono compatibility after processing: mid-channel changes affect the center image

the level-matching trap

the single biggest mistake in evaluating bus resonance processing is comparing processed vs bypassed without level matching.

resonance suppression reduces energy in the 2-5 kHz range. this makes the processed signal quieter. quieter signals sound smoother, not because the processing is working, but because of equal-loudness contour effects: the 2-5 kHz sensitivity peak is most pronounced at lower listening levels, so a quieter signal sounds disproportionately less harsh in this range.[^2]

this means you will always perceive the processed version as “smoother” even if the processing is doing nothing useful. your brain is lying to you. the only way to evaluate honestly is to level-match the bypassed signal to the processed signal, within 0.5 LUFS.

when not to process the bus

not every mix needs bus resonance control. if the individual tracks are well-recorded and well-treated, the bus may be clean. processing a clean bus introduces latency, uses CPU, and risks over-processing for zero benefit.

signs you do not need bus processing:

• fresh ears say it is fine: take a break, come back, and listen. if the mix is comfortable at moderate volume for 10+ minutes without fatigue, leave it alone
• the harshness is one source: if soloing tracks reveals one harsh vocal or one ringy guitar, fix that source. the bus is not the problem
• you are compensating for monitoring: harsh monitoring (bright headphones, untreated room) makes everything sound harsh. fix the monitoring, not the mix

signs you do need it:

• fatigue builds over time: the mix sounds fine for two minutes but becomes uncomfortable after ten. this is cumulative resonance in the 2-5 kHz range
• no single source is responsible: every track is individually acceptable, but the sum is harsh. this is the accumulation pattern that bus processing addresses
• the master is too bright but EQ dulls it: a broad EQ cut removes brightness but also removes clarity. a resonance suppressor targets only the peaks, preserving the overall balance

frequently asked questions

frequently asked questions

should you use a resonance suppressor on the mix bus?

only after fixing individual tracks first. mix bus resonance suppression affects every source simultaneously, so heavy settings damage the overall frequency balance. if harshness persists after treating individual vocals, guitars, and cymbals, gentle bus processing (1-2 dB max reduction, focused on 2-5 kHz) can clean up residual build-up.

how do resonances compound across multiple tracks?

when multiple tracks share energy in the same frequency range (e.g., 3 kHz from vocals, guitars, and cymbals), the energy accumulates at the bus. three individually acceptable tracks with mild peaks become a mix with a painful resonance problem. fixing sources first prevents this compounding.

what is mid/side processing for mastering resonance?

mid/side processing separates the stereo signal into center content (mid) and stereo content (side). since vocals sit primarily in the mid channel, you can target vocal harshness in the 2-5 kHz range without affecting hard-panned instruments or stereo reverb tails. this is more surgical than processing the full stereo signal.

why does quieter always sound smoother when A/B testing?

your ears are most sensitive between 2-5 kHz (ISO 226:2023). processing that reduces energy in this range makes the signal quieter, and quieter signals sound less harsh purely due to equal-loudness effects. always level-match to within 0.5 LUFS before comparing processed and bypassed signals.

how much resonance suppression is too much on the mix bus?

if you can hear the processing as an effect (dulling, thinning, or a sense of "something missing"), you have gone too far. mix bus resonance suppression should be invisible: the mix sounds the same but more comfortable. 1-2 dB of targeted reduction in the 2-5 kHz range is typically the maximum before side effects become audible.

references

a note from the developer

art restorers have a rule: remove what does not belong without damaging what does. mix bus resonance control is the same discipline.

the temptation when building SMOOTH was to make the algorithm aggressive: heavy suppression, fast response, dramatic before-and-after. but on the bus, that approach destroys as much as it fixes. the real work was learning restraint. gentle settings, slow smoothing, focused frequency ranges. months of listening before i trusted it on a master. the defaults in SMOOTH are deliberately conservative because of this: it is easier to push harder than to undo damage.

if your approach to bus processing is different from what i described here, or if you have a level-matching workflow that works better, i want to hear it. jonas@kernaudio.io.