What is a chorus effect?

Chorus in plain English

When two guitarists play the same part, they're never perfectly together. One is a few milliseconds behind the other, and their tuning drifts apart by a few cents as they bend into notes and press harder on the strings. That small, constantly changing mismatch is why two players sound bigger than one player turned up. A chorus effect imitates it with copies of your sound that arrive a little late, where the delay keeps moving so the copies drift slightly in and out of tune with the original.

The movement comes from an LFO, a low frequency oscillator: a wave too slow to hear that turns the delay time up and down for you. Rate sets how fast it moves, and depth sets how far. Mix the moving copies back in with the dry sound and you get the thick, shimmering quality of a chorused clean guitar or synth pad.

Chorus sits in the same family as flanging and phasing, and the chorus vs flanger vs phaser lesson puts the three side by side. This page stays with chorus, starting with why a moving delay changes pitch at all.

How a chorus works

A chorus delays its copies by somewhere around 10 to 30 ms. Below about 10 ms, the frequencies the copy and the original cancel between them are spaced far enough apart to hear as a hollow sweep, which is flanging. Much above 30 to 40 ms your ear starts to hear the copy as a separate event, a double or a short echo. In between, the ear fuses the copy and the original into one sound, which the Haas effect lesson explains, so what you notice is the pitch drift and the width.

Many chorus plug-ins use more than one copy, and each copy is called a voice. The lab uses three, each with its LFO a third of a cycle behind the last, so they're never all moving the same way at once. Most of the time one copy is going sharp while another is going flat, and that spread of slightly different pitches is what reads as several players. Ensemble effects take the same idea further with more voices, or with a slow and a fast LFO moving each delay together.

Why a moving delay changes the pitch

While a delay is getting longer, the copy coming out of it is being stretched, so it plays slightly flat. While the delay is getting shorter, the copy is squeezed and plays slightly sharp. It's the Doppler effect, the same thing that drops the pitch of a siren as it passes you, and it's the reason a chorus sounds detuned when there's no pitch shifter anywhere in it.

How far the pitch moves depends on how fast the delay is changing. If the LFO swings the delay A seconds either side of its centre at f hertz, the fastest change is 2π × f × A seconds of delay per second, and the pitch shifts by that fraction. Take the lab's chorus at 40% depth, which swings each delay 4 ms either side of 20 ms, running at 0.75 Hz: 2π × 0.75 × 0.004 comes to about 0.019, so each copy runs up to 1.9% fast or slow. In musical terms that's about 32 cents, where a cent is a hundredth of a semitone, so roughly a third of a semitone.

Rate and depth multiply, so doubling either one doubles the drift. When a chorus sounds seasick you can halve either knob to halve the wobble, and the choice changes the character: a lower rate keeps the swing and slows the drift down, while a lower depth keeps the speed and narrows the swing.

Anything under about 10 cents is hard to hear as a pitch change at all and comes across as width. Somewhere around 25 to 50 cents is the lush, obvious chorus of a pad. Past 50 cents, a quarter of a tone, most parts start to sound out of tune.

Reading the lab

The lab opens on the chords loop with nothing on it. The cards along the top switch between the dry sound, the chorus, a vibrato and a flanger, and each card keeps its own settings, so you can flip between them without losing anything. On a keyboard, the keys 1 to 4 do the same.

The wide screen compares the sound after the effect with the original, with time running left to right and frequency up the side. A chorus leaves a faint, shifting texture with no clear bands, because its copies move too much to cancel the same frequencies for long. Switch to the flanger and the difference is plain: dark bands that rise and fall together.

Under it, the LFO screen draws the delay of each copy over a couple of cycles around the moment you're hearing, with the bright line for the first copy and the two fainter ones for the others. Next to that is the spectrum at this instant, and the Right now panel reads out the delay range, the pitch drift in cents and the correlation between left and right, all measured from the loop as the lab renders it.

Every option plays at the same loudness as the dry sound, so the chorus can't win by being louder. You can upload your own sound or record eight seconds of one if you'd like to hear chorus on that instead, and nothing leaves your device.

The controls

Chorus plug-ins vary more than compressors do, but nearly every one has a rate and a depth, with the depth sometimes called intensity or amount. Logic's Chorus and Ableton Live's Chorus-Ensemble have both, plus a mix control. FL Studio's Fruity Chorus has a Depth knob and calls its rate LFO frequency, but it has no mix knob, only a Wet Only switch for when it sits on a send.

Rate

Rate is the speed of the LFO in hertz, or cycles per second. Most chorus settings sit between about 0.2 and 2 Hz. Slow rates give a gentle drift you feel more than hear; past about 2 Hz the movement turns into a quick wobble, and by 5 or 6 Hz it's close to the vibrato on a held note. Because rate feeds into the pitch drift, speeding a chorus up without lowering the depth makes it more out of tune as well as faster.

The lab rounds the rate to a whole number of cycles per loop, so the screens and the sound stay lined up. On the 8 second loops that means steps of 0.125 Hz, and at 120 BPM a rate of 0.5 Hz is exactly one cycle per bar. Plenty of plug-ins can sync the rate to tempo, which stops a slow chorus drifting against the groove.

Depth

Depth sets how far the LFO moves the delay. In the lab, 100% swings each copy between 10 and 30 ms, and every 10% adds 1 ms of swing either way. Other plug-ins usually show depth as a percentage with no time attached, so treat the lab's figures as a guide to low, medium and high. Pads and held chords can take a lot of it. Leads and vocals need very little, because the drift is heard against a melody and the ear compares every note with the one before.

Mix

Mix balances the dry sound against the copies. At 50% they're at equal level, which gives the strongest beating between them, the slow rise and fall you hear when two notes are almost in tune. Plenty of pedals have no mix knob at all and run at a fixed blend; the Boss CE-2, for one, has only rate and depth. Lower settings of 20 to 30% thicken a part without the effect drawing attention to itself.

At 100% there's no original left. The copies only blend with each other, the pitch drift becomes much easier to hear and most of the thickness disappears.

Width and spread

Stereo choruses run the copies on the left and right with their LFOs out of step. The lab's Spread control at 100% puts the two sides half a cycle apart, so when the left copies are at their longest delay the right ones are at their shortest. Each speaker gets a different version of the drift, and a mono part opens out into a wide stereo one. The correlation figure in the panel tracks it: 1.00 means the two sides are identical, and lower numbers mean they're less alike. On the chords loop, Spread at 100% takes it from 0.97 to about 0.73. The stereo width lesson covers what that does to a mix.

Delay

Some choruses have a delay or pre-delay control that sets the centre of the sweep. The lab fixes it at 20 ms. Shorter centres, around 5 to 10 ms, edge towards flanging and get a more metallic colour, and longer ones, from 25 to 40 ms, sound looser, closer to a second take of the part.

Feedback

A few choruses have feedback, which sends the copies back into the delay. On a chorus it adds a faint metallic ring, so it's usually left at zero or close to it. It does much more to a flanger, and the flanger lesson is the place to hear it.

Voices

The number of copies changes the character as well. One voice with the dry signal gives the sound of a classic analogue pedal, with an obvious, regular drift. Three or more, with their LFOs offset, smooth the drift into a dense shimmer that's harder to pick apart. The lab's chorus uses three, and the count is fixed.

Chorus vs vibrato

Take the original away from a single-voice chorus and what's left is vibrato: one copy whose pitch is moving, with nothing to compare it against. The thickness of a chorus comes from the copy and the original sounding together at slightly different pitches, which makes them beat against each other the way two slightly mistuned guitar strings do.

The lab sets this up so you can hear it. Its vibrato step swings one copy 4 ms either side at 0.75 Hz, the same movement as each copy of the spread chorus before it, so both drift by about 32 cents. The chorus sounds wide and thick, and the vibrato sounds like the same chords played from a slightly warped record. The vibrato lesson picks it up from there.

Chorus, vibrato and flanging are all made from the same part, a short delay with an LFO on it, and what separates them is how long the delay is and whether the dry sound is mixed back in.

Analogue chorus and the BBD sound

The first chorus many guitarists heard was built into an amplifier. Roland's JC-120 Jazz Chorus came out in 1975 with a stereo chorus inside it, and the next year Boss put a version of that circuit into a pedal, the CE-1 Chorus Ensemble. Both made their delay with a bucket-brigade device, or BBD: a chip that passes the signal along a chain of tiny capacitors one step per clock pulse, so the clock speed sets the delay time. Sweep the clock with an LFO and the delay sweeps with it.

BBDs colour the sound. The filters needed around the chip roll off the top end of the copies, and the chip adds a little hiss, so an analogue chorus sounds darker and softer than a clean digital one. Plenty of plug-ins model the chorus in Roland's Juno-60 synth for that colour. The Juno ran its whole output through a stereo BBD chorus switched by two buttons, marked I and II. If your chorus plug-in has a tone or filter control, turning it down gets you closer to the darker analogue copies.

For the sound of a single-voice analogue pedal on a record, listen to the riff on Nirvana's “Come as You Are”, where Kurt Cobain's guitar runs through an Electro-Harmonix Small Clone.

Where to use chorus

Guitar

Clean electric guitar is where chorus gets used most. Chords and arpeggios at 0.5 to 1 Hz, with a moderate depth and an even mix, get a wide shimmer that fills the space between the speakers. On a distorted guitar the usual order is distortion first and chorus after it, so the drive doesn't flatten the movement. Pick Guitar or Nylon guitar in the lab to try it on a real part.

Synths and keys

Pads and polysynth chords take chorus better than almost anything. It does a similar job to detuning several oscillators against each other, which the unison and detune lesson covers, and it's why so many synths have a chorus built in. A supersaw that's already heavily detuned rarely needs more, since chorus on top tends to blur it. Electric pianos suit a lighter, slightly faster setting.

Bass

On bass, keep the chorus to the upper part of the sound, the growl and the pick attack. Below about 150 Hz the drift and the stereo spread work against you: the low end loses focus, and in mono some of it can cancel. Put the chorus on a send with a high-pass filter at around 150 to 200 Hz on the return, so the low end stays dry and centred. The insert vs send lesson shows how to set one up.

Vocals

A lead vocal with obvious chorus sounds out of tune, because the copies drift against the singer. For a little width on a lead, use a light chorus on a send with a slow rate and a low depth, and keep the return well under the dry vocal. Backing vocals and doubles can take a lot more, and a wide chorus on a stack of harmonies pushes them out to the sides, away from the lead.

How to set up a chorus, step by step

1. Decide between an insert and a sendOn a single instrument an insert is fine, with the mix at 30 to 50%. On a send, set the chorus fully wet and use the return fader as your mix, which also lets you filter the return.
2. Start slow and shallowA rate around 0.5 Hz with the depth near the bottom of its range gives you a starting point where the effect is barely there.
3. Bring the depth up until you hear itThen back it off a little. On held notes, listen for the moment the pitch starts to wobble and stay just under it.
4. Set the rate for the feel you wantSlower for a drift, faster for a shimmer. Lower the depth as you speed up, so the pitch drift stays where it was.
5. Widen itTurn up the width or spread and listen in the full mix, where a part that sounds huge on its own often needs less.
6. Check it in monoSwitch your master to mono. If the part drops in level or goes hollow, bring the width or the depth down until it holds up.

Starting settings for common sounds

These are places to begin, chosen so the chorus is clearly doing something. The depth figures are for the lab, so on other plug-ins read them as low, medium or high.

Chorus in mono

A stereo chorus gets its width by making the left and right different, and whatever differs between the two sides partly cancels when they're added together in mono. Suppose that at one instant the left copy of a note is 16 ms late and the right copy is 24 ms late. Summed to mono those two copies are 8 ms apart, which is a comb filter with notches starting at 62.5 Hz and repeating every 125 Hz. The delays keep moving, so the notches move too, and a pad that sounded wide on headphones can go hollow and slightly phasey on a phone speaker.

The fix is usually small: lower the width or the depth until the mono version holds up. A correlation meter helps while you do it, since a reading that stays above zero is a good sign. Step 6 in the lab goes the other way: with full spread and no dry sound, the reading drops to about −0.13, which means the two sides are partly working against each other. The mono compatibility lesson shows what else to listen for.

Common chorus mistakes

Too much depth on a melody

Drift that sounds lush on a pad sounds out of tune on a lead. At 1 Hz with 5 ms of swing, a copy drifts by about 54 cents, more than a quarter of a tone. On leads and vocals, keep the depth low enough that you notice the width before the wobble.

Chorus on everything

When every part is chorused, they all spread into one wide wash and nothing is left in the middle. Choose the one or two parts that need width and leave the rest dry, so the vocal and the bass have a centre to sit in.

Chorusing the sub

A chorus across the whole of a bass smears the lowest notes and can make them drop out in mono. Filter the low end out of the effect, as in the bass section above.

A wet-only chorus on an insert

If a chorus on an insert sounds thin and wobbly where you expected it to sound thick, check the mix. At 100% wet there's no original for the copies to beat against, which is the vibrato you heard in the lab.

Training your ears to hear chorus

The quiz under the lab plays the current sound through a chorus, a vibrato or a flanger with random settings and the screens covered. Its vibrato is fast, 4 to 6 Hz against the 0.75 Hz of step 7, so a quick wobble with nothing getting thicker gives it away. The flanger always has some feedback in the quiz and a metallic edge to its sweep, which leaves the chorus as the one where the sound thickens under a slow drift. Eight rounds take a few minutes, and your best score is saved in your browser.

Questions people ask

Keep going