Saturation and harmonic distortion: what it adds that an EQ cannot
Why an EQ only moves what already exists while a saturator manufactures new frequencies: even and odd harmonics, the missing fundamental that saves bass on a phone, and when oversampling actually matters.
What an EQ cannot do
An EQ is linear. That means something very specific: it can only change the level of frequencies already present in the signal. If a recording has nothing at 4 kHz, you can boost 4 kHz by twenty decibels and nothing will appear, because there was nothing to amplify.
A saturator is non-linear, and that's the whole difference: it manufactures frequencies that didn't exist. That's where its real usefulness lies, and its danger too, because you don't choose what it manufactures directly — the shape of its curve decides.
The shape of the curve decides everything
A saturator is completely described by one graph: what it outputs for every value that goes in. The diagonal line is doing nothing — output identical to input — and everything that departs from that diagonal is distortion.
- — Soft saturation (tube, tape)
- — Hard clipping (clipper, diode)
There's a practical consequence people miss: the drive knob isn't an amount-of-effect control. Almost always it's the level at which you hit a fixed curve. Turning drive up means hitting the graph further out, where it bends harder. That's why saturation is the process least tolerant of poor gain staging: the same track two decibels hotter enters at a different point on the curve and sounds different.
Even and odd harmonics
The harmonics that appear are exact multiples of the original frequency, and each multiple lands on a specific musical interval. That's what sits behind the adjectives everyone uses.
| Harmonic | Interval above the note | How it reads |
|---|---|---|
| 2nd (even) | Octave | Body, reinforcement. Doesn't read as added |
| 3rd (odd) | Octave + fifth | Density, grip. Starts to be noticeable |
| 4th (even) | Two octaves | Consonant brightness |
| 5th (odd) | Two octaves + major third | Edge, presence |
| 7th onward | Outside the scale | Harshness, a broken quality |
Which family a device generates depends on whether its curve is symmetrical. A curve that treats the positive and negative half-cycles alike produces only odd harmonics; an asymmetrical one adds the even ones. This isn't marketing: it's the technical reason a class-A tube stage — asymmetrical by construction — and a push-pull stage sound different.
| Source | Predominant | Typical use |
|---|---|---|
| Tube (class A) | Even | Vocals, bass, anything asking for body |
| Tape | Even, plus peak compression | Drums, mix bus |
| Transformer | Even in the low end | Low end, bus glue |
| Transistor | Odd | Guitars, synths, aggression |
| Digital clipper | Odd, many and high | Peak control ahead of the limiter |
The fundamental that isn't there
This is the application that solves the most problems and is explained the worst. A phone or laptop speaker doesn't reproduce 50 Hz: there isn't enough cone. If your song's bass lives there, on those devices it doesn't exist, and no EQ fixes it because the speaker can't deliver what it can't deliver.
Saturating the bass produces harmonics at 100, 150 and 200 Hz. Those do come out of the small speaker. And here something happens that looks like magic and isn't: on receiving a regularly spaced harmonic series, the ear reconstructs the fundamental note even though it isn't present. It's called the missing fundamental, and it's why a saturated bass reads as low on a phone while a clean one vanishes.
Aliasing: the part nobody mentions
Distortion generates harmonics upward without end. A digital system can only represent frequencies below half its sample rate: at 44.1 kHz the ceiling is 22 kHz. Harmonics above that ceiling don't disappear — they fold back down.
The problem is where they land. Saturate a 5 kHz tone and its harmonics go to 10, 15, 20, 25 and 30 kHz. The 25 kHz one folds back and appears at 19.1 kHz; the 30 kHz one at 14.1 kHz. Neither is a multiple of 5 kHz, so they're harmonics of nothing: frequencies with no musical relationship to the signal. That's what you hear as metallic and out of tune.
- The brighter the source, the worse: a hi-hat or cymbal has harmonics clearing the ceiling immediately. Bass barely aliases at all.
- The harder the clip, the worse: a clipper generates far more high harmonics than a gentle curve.
- Oversampling is the fix: the plugin works internally at four or eight times the rate, so the ceiling rises and the harmonics fit before being filtered on the way back.
- It costs CPU, so enable it where it matters — bright material, high drive, the master — and not on twenty tracks out of habit.
Where to put it
- In parallel, especially on low end and vocals. You blend the saturated version under the clean one and set the amount with a fader, without committing the original.
- On a group bus, not on every channel. Saturating twelve drum tracks separately stacks the same kind of dirt twelve times; saturating the bus binds them.
- Ahead of the limiter on the master, with a very gentle clip. Taking two decibels of peak off with the curve before the limiter sees it saves the limiter work, and that's audible.
- Never as a last resort to fix an arrangement. If two elements are fighting, saturation makes them fight harder.
On the third point, the LUFS and loudness guide explains why those two decibels change the outcome; on the fourth, the EQ guide has the tool that actually resolves a frequency clash.
Training your ear for this
Saturation is hard to judge because your ear adapts to it in under a minute and stops counting it as distortion. The exercise that works is the opposite of what everyone does: instead of turning it up until you hear it, set it where you like it, keep working, and bypass it twenty minutes later. Your reaction to its absence is the reliable data point.
In Fix The Mix, saturation is one of the six worlds in mission mode, where the exercise is identifying how much has been applied to real material; and the mix diagnosis includes cases where the problem is excess distortion rather than missing EQ, which is exactly the confusion this guide exists to undo.
Frequently asked questions
- What's the difference between saturation, distortion and overdrive?
- None in mechanism: all three bend the waveform and generate harmonics that weren't there. The difference is degree and habit of language. Saturation usually means bending gentle enough not to register as an effect, overdrive the middle ground, and distortion the kind you hear as distortion. There's no measurable boundary between them.
- Why are even harmonics said to sound more musical?
- Because they land on notes consonant with the original. The second harmonic is an octave up, the fourth is two octaves: the same note. Odd harmonics land elsewhere — the third on a fifth, the fifth on a major third — and from the seventh upward they stop matching anything in the scale, which is when distortion starts to sound harsh.
- How much saturation is too much?
- The reliable test is removing it, not adding it. Set it where you like, carry on mixing for twenty minutes, then bypass it suddenly. If it sounds thin without it, it was well judged. If it sounds relieved and cleaner without it, you overdid it. Your ear adapts to distortion very fast and stops counting it as distortion.
- Does it help the bass be heard on a phone?
- Yes, and it's its most useful application. A phone speaker doesn't reproduce 50 Hz, so the bass fundamental simply isn't there. Saturate it and harmonics appear at 100, 150 and 200 Hz, which that speaker does reproduce, and the ear reconstructs the low note from them. It isn't a brightness trick: it's a psychoacoustic effect called the missing fundamental.
- What is aliasing and when should I turn on oversampling?
- Distortion generates harmonics upward without limit. Those above half the sample rate don't fit and fold back down, landing on frequencies that aren't multiples of the original: they sound metallic and out of tune. Turn oversampling on when you saturate bright material — cymbals, hats, airy vocals, the full mix — and at high drive. On a gently saturated bass it barely matters and just burns CPU.
- Why does saturation make something louder without raising the peak?
- Because it rounds the peaks and lifts everything else, so the signal carries more average energy for the same peak: the crest factor drops. That's more perceived loudness without touching the limiter. It's why gentle clipping before the limiter gets you to the same LUFS with less gain reduction.