Parallel saturation in plain English
Put a saturator straight on a track and everything you hear has been through it. Parallel saturation splits the sound in two instead. One path stays clean and the other goes through a saturator driven much harder than you'd ever use on its own, then the two are mixed back together with the saturated copy usually some way under the original.
Saturation does two things at once when you push it. It adds harmonics, new frequencies at whole-number multiples of each note, and it rounds off the peaks, which squashes the front of every drum hit. On a single saturator you can't get much of the first without the second. In parallel you can, because the clean path still has every transient at full height and the copy underneath brings the harmonics.
It's the same routing as parallel compression, and it shares one of its side effects, which the next section works through. If drive, curves and harmonics are new to you, start with what is saturation?; this page assumes you know roughly what a saturator does and goes into what changes when you run one alongside the clean sound.
Parallel saturation in numbers
Drive is gain before the curve, and a saturation curve flattens out near the top. Push a quiet note and a loud one into it and the quiet one comes out much louder than it went in, while the loud one can't get any louder than the top of the curve. So a hard-driven copy has had its dynamic range squeezed, the same way a heavily compressed copy has.
Follow two notes through the lab's soft clipper with 24 dB of drive. A quiet note peaking at −40 dBFS sits low enough on the curve to be raised almost the full 24 dB, and comes out of the copy at −16 dB. A loud hit peaking at −1 dBFS runs into the top of the curve and comes out at 0 dB, just 1 dB higher. In the original the two were 39 dB apart; in the copy they're 16 dB apart.
Blend the copy in at 30% and the loud hit comes out at −0.7 dB, only 0.3 dB above where it started, because at that level the clean sound and the copy are nearly the same size. The quiet note comes out at −25.3 dB, which is 14.7 dB higher than before. The gap between them has shrunk from 39 dB to about 25 dB, and nearly all of that came from turning the quiet note up.
So parallel saturation brings up quiet detail, like ghost notes and the room sound between drum hits, as well as adding harmonics. Turn the drive in the figure down and that lift fades, because the copy stops squeezing; at 0 dB the amber line sits almost on the dashed one. If you want the colour without the lift, use gentler drive on the copy and more of it in the blend.
The harmonics come down with the mix. On the lab's 110 Hz test sine, the valve curve at 24 dB of drive puts its strongest harmonic, the 3rd, at −11 dB below the note. Blend it in at 50% and that bar drops to −18 dB, at 30% to −23 dB and at 10% to −33 dB. Every harmonic falls by about the same amount, so the balance inside the copy stays put and only its level changes.
Heavy drive blended low, or light drive at full mix
If blending only lowers the harmonics, you might wonder why you wouldn't use less drive at 100% mix and get the same thing. The difference is in the shape of what gets added, and in which parts of the sound get it.
Take the lab's soft clipper on the 110 Hz sine. With 2 dB of drive at 100% mix, the 3rd harmonic comes out at −22 dB below the note, the 5th at −42 dB and the 7th at −61 dB. Drive it 24 dB and blend it in at 20%, and the 3rd lands at −21 dB, almost the same, while the 5th is up at −26 dB, the 7th at −29 dB, the 9th at −32 dB and the 11th at −35 dB. Added up, the heavy blend's harmonics come to about 11% of the note's level against 8% for the light drive, so the totals aren't far apart. The heavy copy spreads its share much further up the spectrum, and those upper harmonics are what you hear as edge and bite.
The second difference shows up on real music. With light drive at full mix, only the loudest peaks reach the bend in the curve, so quiet notes come through almost clean and loud notes get the colour. A hard-driven copy saturates everything, quiet notes included, and then the clean path puts the transients back on top. On drums, that means the ghost notes get as much grit as the backbeat while the backbeat keeps its crack.
Light drive at full mix suits sounds you want warmer without changing how they move, like a vocal or a mix bus. Heavy drive in parallel suits sounds you want denser and more aggressive while keeping the attack, which describes most drums, bass and plucked synths.
Reading the lab
This is the saturation lab from the saturation page, cut down to the four curves that suit a parallel copy: valve, tape, soft clip and hard clip. The bitcrusher and downsampler controls are locked, and a Low cut knob beside Tone filters the copy only. The lesson starts with heavy valve drive on the drums at 100% mix, then brings the mix down so you can hear the clean hits come back.
On the big screen, the pale shape is the sound after the drive, before the curve, and the coloured shape is what comes out, which below 100% mix is the blend. The transfer curve draws the saturated copy as a solid line, and once the mix is under 100% it adds a dashed line for the blend, so you can see how much straighter the blend is than the copy alone. The harmonic bars measure the blend on a pure 110 Hz sine.
The figure to watch in the panel on the right is the crest factor, the gap between the peak level and the average level. Saturation shrinks it by rounding the peaks. When the blend is working, the crest factor sits between the copy's figure and the original's, and the closer it gets to the copy's, the more the hits are being rounded off. Auto gain is on, so the blend always plays at the loudness of the clean loop, and A/B flips between the two.
Mix knob or a send
You can set this up with the saturator's own mix knob or with a send to a separate bus. Both give the same sound once the levels are right.
With a mix knob
Many saturators have a control that blends the processed and clean signals inside the plug-in. Ableton Live's Saturator calls it Dry/Wet, and FabFilter Saturn 2 and Soundtoys Decapitator call it Mix. Turn the drive up past where it sounds good, then pull the mix down until the original comes back on top. Because both paths live inside one plug-in, they always stay in time with each other.
Mix knobs don't all behave alike, though. The lab's is a plain crossfade that blends the copy as it comes off the curve, louder for the drive. Some plug-ins turn the copy down to make up for the drive before it reaches the mix, and then you'll need more of it for the same effect. That's why a mix percentage from one plug-in is only a rough starting point on another.
With a send to a separate bus
The other way is a send to a return or aux channel with the saturator on it at 100% wet. The track plays as normal, and the bus fader decides how much of the copy you add. It's the only option with a saturator that has no mix knob, and it's the better one when you want to process the copy further.
A bus lets you put things after the saturator that only the copy goes through: a high-pass filter, an EQ cut on a harsh frequency, even a compressor. Several tracks can also share one bus, which is how a whole drum kit gets the same saturated copy. A linear 30% mix is the same blend as a bus fader about 7.4 dB under the clean track, and the parallel compression lesson has a table of the other settings.
How to set up parallel saturation
This is the bus method on a drum group. The same steps work for bass, synths and vocals, with the settings from the table further down.
- Make a return or aux busIn Ableton it's a return track, in Logic an aux, and in FL Studio a mixer insert. Call it something like Drums Sat.
- Send the drums to itAdd a post-fader send from the drum group at 0 dB, so the copy follows your drum levels.
- Pull the bus fader all the way downStarting from silence lets you hear the copy arrive, which is easier to judge than a loud blend from the first second.
- Put a saturator on the bus and drive it hardSolo the bus and push the drive until the drums sound thick, flattened and a bit too much. Pick the curve by ear: tape for density, valve for a fuller low mid, a clipper for bite.
- Filter the copyStill soloed, add a high-pass filter after the saturator if the low end gets woolly, and a low-pass around 6 to 10 kHz if the top fizzes.
- Unsolo and raise the bus faderBring it up slowly under the clean drums. Stop when the kit sounds thicker and the hits still have their crack, then back it off a decibel or two.
- Check it at matched loudnessMute and unmute the bus while turning the drum group down to make up for the level the copy added. If the drums only sound better when they're louder, pull the bus back.
Filtering the saturated copy
Filtering is where parallel saturation does things a single saturator can't. Whatever you do to the copy after the curve changes only the part of the sound you added, so the clean track keeps its full range.
Taking the fizz off the top
Heavy drive adds harmonics a long way up, and the highest ones can sound fizzy or harsh, especially from a hard clipper. A low-pass filter on the copy, somewhere between 3 and 10 kHz, takes them out while the clean path keeps the cymbals and the air. The lab's Tone control does this: it's a low-pass after the curve, and it filters only the saturated copy, so below 100% mix the clean sound passes it by.
Keeping the sub clean
On bass, the low end of the copy is the problem. Heavy drive makes it dense and loud, and stacked on the clean sub it turns muddy. A high-pass filter on the copy, after the saturator, fixes that. Set it between 100 and 250 Hz and the copy adds only harmonics, while the clean track carries the sub.
The high-pass has to come after the saturator, because the harmonics are made from the low notes, and a high-pass before it removes the notes they'd have come from. A 55 Hz bass note gives the saturator harmonics at 110, 165, 220 Hz and up. A 150 Hz high-pass after it turns down the 55 Hz note and most of the 110 Hz harmonic, and lets the 165 Hz one and everything above through. The lab's Low cut knob is this high-pass, and it only touches the copy.
Your ear works out the pitch of a note from its harmonics even when the fundamental itself is quiet or missing. That's why this helps bass on phone speakers and small laptop speakers, which play little or none of the sub: the notes stay easy to follow there, and on big speakers the sub is still the clean one you started with.
Starting settings for the copy
These assume a saturator with a linear mix knob like the lab's, and the drive figures are in the lab's decibels. On a plug-in with a 0 to 10 drive knob, solo the copy and turn it up until it sounds like too much. On a bus, a 20% mix works out at about 12 dB under the clean track, 30% at about 7 dB under and 40% at about 3.5 dB under.
| Source | Curve | Drive on the copy | Filter on the copy | Mix |
|---|---|---|---|---|
| Drum group | Tape or soft clip | 18 – 30 dB | Low-pass 6 – 10 kHz | 20 – 40% |
| Snare or room mic | Hard clip or valve | 15 – 25 dB | Low-pass 8 kHz | 15 – 30% |
| Bass | Tape or valve | 18 – 30 dB | High-pass 100 – 250 Hz | 25 – 50% |
| Lead vocal | Valve or tape | 12 – 20 dB | High-pass 200 Hz, low-pass 6 – 8 kHz | 10 – 25% |
| Synth or keys | Valve or soft clip | 12 – 24 dB | Low-pass 8 kHz | 20 – 40% |
| Mix bus | Tape or valve | 6 – 12 dB | None | 10 – 20% |
Where to use parallel saturation
Drums
Drums get the most from it, because their hits are short and tall and a saturator in series rounds them off first. In parallel, the backbeat keeps its crack and the copy fills in the body, the room and the ghost notes. A drum group sent to one saturated bus is the usual setup, and the drum bus lesson covers what else tends to go there. If the kick is pushing the whole copy into the curve on every beat, put a high-pass at 80 to 120 Hz on the bus before the saturator. This is the one case where the filter goes first, because the aim is to keep the kick out of the copy altogether and leave its low end to the clean drums.
Bass
On bass, the high-pass after the saturator does most of the work. Saturate a copy hard with tape or valve, high-pass it at 150 Hz or so, and blend it in until the notes come through on small speakers. Steps 7 and 9 in the lab are this bass setup without and with the high-pass; press A/B in either to hear how much the blend adds to the upper part of each note.
Vocals
A saturated copy blended low brings a vocal forward without making it sound distorted, because the copy's grit sits under the clean voice. Keep it lower than on drums, at 10 to 25%. The S and T sounds are the part to watch, since they turn harsh fast when they're saturated, and a low-pass on the copy at 6 to 8 kHz or a de-esser in front of the saturator keeps them clean.
Synths and keys
On a mellow synth pad or an electric piano, a hard-driven valve copy at 20 to 40% thickens the sound and helps it hold its place in a busy mix. Chords need more care. Saturating several notes at once also makes sum and difference tones between them, called intermodulation, which aren't in tune with the chord and turn muddy quickly. Blending the copy in low keeps them down along with everything else the copy adds. Pick Chords in the lab and move the mix between 20% and 100% to hear how fast they build up.
Parallel saturation or parallel compression?
Both lift the quiet parts of a sound under a clean original, so on drums they can sound close. A compressor turns the level up and down over time and, at normal attack and release settings, leaves the shape of each cycle alone. A saturator changes that shape, so its copy carries harmonics. Use parallel compression when you want more body and sustain with no extra colour, and parallel saturation when the part also needs edge or needs to be heard on small speakers. Nothing stops you running both on one drum bus, and saturation vs compression compares the two side by side.
Common parallel saturation mistakes
Judging the blend while it's louder
Adding a saturated copy makes the track louder, and a louder version nearly always sounds better for a few seconds. Turn the group down when you bring the bus in, or use the lab's Auto gain and A/B, so you're comparing tone at the same level.
The two paths falling out of time
Saturators that oversample, which many do to cut down aliasing, often add some latency. If the copy arrives even a millisecond late, it combines with the clean sound into a comb filter, a hollow, phasey tone with a row of notches through the spectrum. Your DAW's delay compensation lines the paths up, so check it's switched on if the blend suddenly sounds thin.
A filter shifting the copy against the clean sound
An ordinary EQ filter delays the frequencies around its cutoff slightly, and in a blend that changes how the copy and the clean sound add up there. It's usually small. If a bass loses weight when you bring in a high-passed copy, move the cutoff lower or pick a gentler slope, and compare with the copy muted.
Too much of the copy
Once the copy is loud enough to set the shape of the peaks, you're back to ordinary saturation with extra routing. If the drums start to sound flat again, pull the bus down until the hits come back.
Saturating the dry path as well
A touch of saturation on the clean track is fine. Heavy saturation there takes away the transients and the clean low end the blend depends on, so the copy has nothing to sit under.
Train your ears
The quiz under the lab plays the current sound through valve, tape or hard clip at 100% mix with the screens covered, and asks which one it is. Knowing the curves by ear makes a parallel bus faster to set up, because picking a curve is how you choose the copy's character. The levels are matched in every round, and your best score is saved in your browser.
Questions people ask
Is parallel saturation the same as parallel distortion?
Yes. It's the same routing with a heavier curve. People tend to say saturation when the copy is driven into tape or valve curves and distortion when it's a clipper, a fuzz or an amp simulator, but either way a processed copy is being mixed under the clean sound.
What mix percentage should I use?
Start at 20 to 40% with the drive set much harder than you'd use on its own. Mix knobs on different plug-ins crossfade in different ways, and some turn the copy down to make up for the drive while others don't, so 30% on one saturator won't sound exactly like 30% on another. Set the final blend by ear at matched loudness.
Does parallel saturation cause phase problems?
It can. If the saturator adds latency, usually because of oversampling, the copy arrives late and the blend turns hollow unless your DAW's delay compensation is on. A filter on the copy also shifts the timing of the frequencies around its cutoff, which changes the tone of the blend there. A plug-in's own mix knob normally keeps both paths in time inside itself.
Should I saturate the dry track as well?
A little is fine, but the dry path is what keeps the transients and the low end clean. If it's saturated heavily too, the blend loses the reason for doing it in parallel, and you may as well use one saturator at a lower mix.
Can I use parallel saturation on the mix bus?
Yes, with a light hand. A tape or valve copy blended 10 to 20% under the whole mix adds density without rounding off the drums. If you can clearly hear it on an A/B at matched loudness, it's too much for a mix bus, because whatever it adds lands on every part at once.
Why does my bass get muddy when I saturate it in parallel?
The copy has its own low end, and heavy drive makes that low end dense and loud, so it piles onto the clean sub. High-pass the copy after the saturator, somewhere between 100 and 250 Hz, so the copy only adds harmonics and the clean track carries the sub on its own.