saturation on the mix bus

the invisible glue

your mix sounds good. every track sits where it should. the levels are right, the EQ is clean, the compression is controlled. but when you listen to a reference track, their mix sounds like one thing and yours sounds like twenty things playing at the same time.

the difference is often mix bus saturation. a carpenter does not pour glue over the table. they apply it at the joints. same principle here. gentle, nearly invisible harmonic colouring across the full mix that makes individual elements feel like they belong together. when every source passes through the same nonlinear process, they share harmonic content. the kick drum, the vocal, the guitars all develop overtones from the same transfer curve. that shared harmonic fingerprint is what engineers call “glue.”

the key word is gentle. if you can hear the saturation when you bypass it, you are probably using too much.

how much is enough

the sweet spot for mix bus saturation is narrow. too little and it does nothing. too much and it damages the mix.

in practice, most experienced engineers work with total harmonic distortion (THD) levels between 0.1% and 0.5% on the mix bus. for context, THD below 0.1% is generally inaudible. THD above 1% is clearly audible as coloring or distortion. the mix bus sweet spot sits right at the threshold of perception: enough to create the shared harmonic content, not enough to draw attention to itself.

in terms of drive settings, this usually means running the mix bus signal into the saturator at or slightly below unity. you are not “pushing” into the nonlinear region the way you would on a vocal or drum bus. the signal should barely enter the curved part of the transfer function.

character on the mix bus

not all saturation types work equally well on the mix bus. the choice of character matters more here than on individual tracks, because the bus carries the entire frequency range and any coloring affects everything.

tape (symmetric, odd harmonics)

tape-style saturation uses a symmetric transfer function, typically a tanh (hyperbolic tangent) curve. symmetric nonlinearities produce primarily odd harmonics: 3rd, 5th, 7th. these add presence and density without shifting the tonal balance. tape is the traditional choice for mix bus saturation because its effect is relatively neutral. it adds cohesion and gentle peak taming without obviously “warming” the mix.

tube (asymmetric, even harmonics)

tube-style saturation uses an asymmetric transfer function that generates even harmonics: 2nd, 4th, 6th. even harmonics are octave-related and sound warm and musical. on individual tracks, tube saturation adds body and richness. on the mix bus, that warmth can become too much. the even harmonics shift the entire tonal balance toward the low-mids, which can make a mix sound thick and congested rather than glued together.

tube saturation on the mix bus works best when used very conservatively, or when the mix is thin and needs warmth as well as cohesion.

transformer (blended)

transformer-style saturation produces a mix of even and odd harmonics with a character distinct from both tape and tube. it tends to add a subtle edge and presence. on the mix bus, it sits between the neutrality of tape and the warmth of tube.

the intermodulation problem

on an individual track, saturation is straightforward. a bass note at 100 Hz goes through the transfer function and generates harmonics at 200, 300, 400 Hz. clean, predictable, musical.

on the mix bus, dozens of frequencies enter the nonlinear function simultaneously. the math changes. when two frequencies pass through a nonlinearity at the same time, they interact and generate intermodulation products: new frequencies at the sum and difference of the originals.

a bass note at 100 Hz and a vocal at 500 Hz produce sum tones (600 Hz) and difference tones (400 Hz). add a guitar at 800 Hz and you get additional products at 900, 700, 300, and 200 Hz. every pair of input frequencies creates new energy at new frequencies. with a full mix containing hundreds of simultaneous frequency components, the intermodulation products accumulate rapidly.

the low end is where intermodulation causes the most audible damage. bass frequencies are high in energy, so their intermodulation products are also high in energy. these products land in the low-mids, creating muddiness and thickness that was not in the original mix. this is why aggressive mix bus saturation makes the low end sound congested even when the bass and kick were clean to begin with.[^1]

the solution is restraint. gentle drive levels keep intermodulation products below the audible threshold. some saturation plugins offer frequency-dependent drive, applying more saturation to the midrange and highs while leaving the bass relatively clean. this is one advantage of spectral-aware saturation: the processor can control exactly how much nonlinearity is applied per frequency region.

frequency-dependent saturation

treating the entire spectrum uniformly on the mix bus is a compromise. the low end has high energy and generates the most intermodulation. the highs have low energy but are sensitive to aliasing artifacts. the midrange is where saturation adds the most musically useful density.

frequency-dependent saturation addresses this by varying the drive level across the spectrum. the concept is simple: apply less saturation to bass frequencies (reducing intermodulation), full saturation to the midrange (where harmonics add density and presence), and controlled saturation to the highs (adding air without aliasing).[^2]

some plugins implement this as a “tilt” or “focus” control. others let you set the drive per frequency band explicitly. the effect is the same: the saturation character adapts to the spectral content rather than treating every frequency the same.

where it goes in the chain

the position of your saturator in the mix bus signal chain changes what it does. there is no universally correct answer, but there are clear trade-offs.

before compression

saturation before the bus compressor rounds off transient peaks, giving the compressor a smoother signal. the compressor works less hard, produces fewer pumping artifacts, and the combined effect is more transparent. this is the most common position for mix bus saturation.

after compression

the compressed signal has a narrower dynamic range. more of it sits consistently in the saturator’s nonlinear region, producing more even harmonic content. the saturation effect is more consistent but also more obvious. this works when you want the saturation to be a deliberate tonal choice rather than invisible glue.

before EQ

the harmonics generated by saturation get shaped by your bus EQ. you can cut the low-mid intermodulation products or boost the upper harmonic density after the fact. this gives you post-hoc control over the tonal effect of the saturation.

before limiting

this is the default position in many mastering chains. the limiter catches any peaks the saturation creates, and the saturation fills out the dynamic range that the limiter removes. the combination can increase perceived loudness without the aggressive artifacts of pushing the limiter harder.

common mistakes

the loudness trap

this is the most common mistake. saturated audio sounds louder. louder sounds better. engineers crank the drive, compare without level-matching, and conclude that more saturation equals a better mix. it does not. it equals a louder mix that is also muddier, less dynamic, and more fatiguing.

always level-match your A/B comparison. match LUFS, not peak level. if the difference disappears when levels are matched, the saturation was doing nothing useful.

over-processing

gentle saturation adds cohesion. aggressive saturation destroys detail. transient clarity, stereo separation, and dynamic range all degrade with excessive drive. the attack of a snare drum gets rounded off. the separation between instruments narrows as intermodulation fills the gaps. the mix starts to feel flat and lifeless, which is the opposite of glue.

ignoring the context

mix bus saturation is the last link in the chain. if you have already saturated individual tracks, the mix bus sees a signal that is already harmonically rich. adding more saturation on top compounds the nonlinearity. the intermodulation products from individual track saturation become the input to the bus saturation, generating second-order intermodulation products. listen to whether the mix actually needs bus saturation. sometimes it does not.

the A/B protocol

before committing to any mix bus saturation setting, follow this protocol:

1. set the drive to the level you think sounds good
2. level-match the bypassed signal (match LUFS, not peak)
3. bypass the saturator and listen for 30 seconds
4. engage the saturator and listen for 30 seconds
5. ask yourself: “does the mix sound more cohesive, or just different?”

if the answer is “just different,” reduce the drive or remove it entirely. mix bus saturation that only changes the sound without improving it is adding complexity for no benefit.[^3]

frequently asked questions

frequently asked questions

how much saturation should you use on the mix bus?

less than you think. on the mix bus, saturation should be felt, not heard. most experienced engineers use drive levels that produce 0.1-0.5% total harmonic distortion. if you can hear the saturation as an obvious effect when you bypass it, you are using too much. the goal is subtle density and cohesion, not audible coloring.

should you put saturation before or after compression on the mix bus?

both positions work, but they do different things. saturation before compression rounds off transient peaks, giving your compressor a smoother signal to work with. saturation after compression acts on the already-compressed signal, which means more of the signal sits in the nonlinear region and the saturation effect is more consistent. try both and listen. many engineers use saturation before compression for a more transparent result.

why does tape saturation work best on the mix bus?

tape-style saturation uses a symmetric transfer function (tanh curve) that produces primarily odd harmonics. these odd harmonics add presence and density without shifting the tonal balance toward warmth the way even harmonics do. on the mix bus, where you want cohesion without coloring, the neutral character of odd-harmonic saturation tends to be more transparent than tube-style even-harmonic saturation.

what is intermodulation distortion and why does it matter on the mix bus?

when multiple frequencies pass through a nonlinear process simultaneously, they interact and generate new frequencies at the sum and difference of the originals. a bass note at 100 Hz and a guitar at 500 Hz create new energy at 600 Hz (sum) and 400 Hz (difference). on individual tracks this is manageable, but on the mix bus with dozens of simultaneous frequencies, intermodulation products accumulate and can make the low end muddy or the midrange congested.

can you use saturation instead of compression on the mix bus?

saturation provides gentle peak reduction because the transfer function naturally compresses peaks that enter the nonlinear region. but it does not replace a compressor. you cannot control the attack, release, ratio, or threshold of saturation the way you can with a compressor. use saturation for tonal cohesion and gentle peak taming. use compression for deliberate dynamic control. they complement each other.

references

a note from the developer

mix bus saturation was the use case that shaped WARM’s entire architecture. the question i kept hitting: how do you add harmonic content to a full mix without the intermodulation artifacts ruining the low end?

the answer turned out to be frequency-dependent drive. the spectral analysis pipeline controls how much saturation each frequency region receives, so the midrange gets the warmth while the bass stays clean. that single decision, made early, shaped everything else about the plugin. without it, bus saturation on bass-heavy material just turns to mud. (i know because the first ten prototypes did exactly that.)

if your bus saturation workflow is different from what i described, jonas@kernaudio.io. i am always looking for approaches i have not tried.