What is stereo width?

Stereo width in plain English

A stereo track is two channels, one for each speaker. If both channels carry exactly the same signal, you hear one sound placed in the middle, sometimes called the phantom centre because no speaker is actually there. Make the channels different and the sound starts to spread out. A hi-hat louder on the left sits to the left, and a pad with different notes moving in each channel fills the space between the speakers and can seem to reach past them.

Stereo width measures that difference. A narrow sound has channels that are nearly the same, and a wide one has a lot that's unique to each side. Width belongs to the whole stereo picture too: a mix with the drums in the middle and guitars hard left and right is wide, even if every single track in it is mono.

Width costs you something in mono. When a stereo mix plays on one speaker, the two channels are added together, and anything that was different between them partly or fully cancels.

Mid and side: the numbers behind width

Any stereo signal can be split into two other parts. The mid is what the channels share, (L + R) ÷ 2, and the side is how they differ, (L − R) ÷ 2. Add the side to the mid and you get the left channel back; subtract it and you get the right. Nothing is lost by switching between the two descriptions, which is why mid and side are the standard way to think about width.

Seen this way, width is the level of the side compared with the mid. Mono is a sound with no side at all. In most mixes the side sits well below the mid, and the full mix in the lab measures 6.5 dB below. When the side gets as loud as the mid, the left and right channels have nothing in common on average, and that sounds very wide.

A width control in a DAW works on this split. It leaves the mid alone and turns the side up or down. On a percentage scale, 100% leaves the sound as it was, 0% removes the side and 200% doubles it, which is the same as raising it by 6 dB.

The last column is the gap between the stereo version and the mono version. The mono version is the mid on its own, and it doesn't change at all as you turn the knob. What changes is how much extra the stereo version carries on top of it, and that extra is exactly what a mono speaker throws away.

Reading the lab

On this page the lab has Width, Pan, Delay R and the Mono button. The other controls belong to the phase lesson and are greyed out here.

The round screen is a goniometer. It plots each moment of the sound by how far it pushes the left and right speakers, turned so that a sound identical in both channels stacks into a vertical line. Width shows up as spread to either side of that line: a narrow sound is a thin vertical streak, a wide one is a fuzzy ball, and a shape that leans over towards horizontal has more side than mid.

Below it, the correlation meter turns the same picture into a number between −1 and +1, with a strip underneath that shows the reading across the whole loop. The three meters beside it give the left, right and mono sum levels. The Right now panel explains what your settings do to the loop you picked, from numbers measured on that loop as it runs through the same signal path you hear.

Your own mix makes a good test. Load it with + Your sound (it stays in your browser) and the Right now panel tells you its correlation and how much it loses in mono before you touch anything.

How to measure stereo width

The correlation meter is the closest thing to a width meter that most DAWs have. When the two channels are about as loud as each other, its reading follows directly from the mid and side levels: it's (mid power − side power) ÷ (mid power + side power). So +1 means no side at all, 0 means the side is exactly as loud as the mid, and anything below 0 means the side is louder.

The same two levels tell you what mono will cost. The mono version keeps the mid power and loses the side power, so a side 6 dB below the mid costs about 1 dB in mono, and a side as loud as the mid costs 3 dB. Here are the loops in the lab, measured as they were supplied:

A single number hides a lot, though. A mix can measure +0.6 overall while its bass is wide and its vocal is mono, and that's a much bigger problem than the reverse. Watch the strip under the correlation meter as the loop plays, and listen in mono to find out where the side signal lives. The correlation meter lesson covers reading the needle and the goniometer in more detail.

Where does stereo width come from?

Panning mono tracks

Most of the width in a typical mix comes from panning. Each track is mono, and the pan control sends it to the two speakers at different levels. A guitar panned left and a keyboard panned right make a wide mix while each of them stays a single, whole sound. In mono both come back to the middle at a slightly different level, set by your DAW's pan law, and their tone is untouched, which makes panning the safest kind of width.

Double tracking

Record the same part twice and pan the takes hard left and right. The two performances differ by a few milliseconds here and a few cents of pitch there, and those differences keep changing, so the part sounds big without a fixed delay between the sides. In mono the takes add up into a thicker version of the part. It's the standard way to make rhythm guitars and backing vocals wide.

Stereo microphone techniques

A pair of microphones can record width directly. A coincident pair, such as XY, puts two capsules at the same point angled apart, so the channels differ only in level and the recording sums to mono with no comb filtering. A spaced pair puts the mics some distance apart, so sound reaches them at different times; that gives a more open, airy width and some comb filtering in mono. A mid/side pair records the mid and side directly, with a forward-facing mic and a sideways figure-of-eight, so you can set the width after the session by turning the side mic up or down.

Effects

Stereo reverbs, ping-pong delays and stereo chorus all send different signals to each side, which is why they make things wider. Synths do it with unison, stacking detuned copies of each voice and spreading them out across the speakers, as the unison lesson explains.

Stereo width and mono

Plenty of people still hear music in mono. A single Bluetooth speaker adds the two channels together, and so do many club and venue systems and any phone that has only one speaker. Whatever the side signal carried is lost there, so the more of your mix lives in the side, the more of it those listeners miss.

So the parts that carry the song belong in the mid, where mono can't touch them: the kick, the snare, the bass and the lead vocal. Pads and reverb can live out wide, along with doubled guitars, because a listener in mono still gets the song when those go a little quieter.

Width goes wrong in mono in two ways, and they sound different. Width from a mid/side control only costs level, so the wide parts get quieter and their tone stays as it was. Width made from delays or phase shifts between the channels also changes the tone, because the sum turns into a comb filter with notches at regular frequencies. That's the hollow, phasey sound of a widened part folding down, and the comb filtering lesson explains where the notches land.

Can you widen a mono sound?

A width control can't do it. A mono recording has identical channels, so its side signal is silence, and turning silence up by any amount leaves silence. That's why Width at 200% does nothing to the nylon guitar in the lab.

To widen a mono sound you have to give each side something different. The quickest way is a delay of about 5 to 30 ms on one side, which is the Haas effect; a stereo chorus or a stereo reverb also works. Each of these creates a side signal that wasn't in the recording, and each changes what you hear in mono. A second take beats all of them if you can record one.

How do stereo wideners work?

A stereo widener plug-in uses one method or a mix of several, and the method decides what happens in mono. Find out which one yours uses before you trust it on anything important.

In Ableton Live, the Width control on the Utility device is a mid/side width control. Logic's Direction Mixer has one called Spread, and Logic's Stereo Spread plug-in uses the band-splitting method, sending alternate frequency bands to the left and right. iZotope's free Ozone Imager adds a Stereoize mode meant for mono sources. Whichever you use, press Mono with it on and again with it bypassed. The stereo widener vs mid/side comparison puts them side by side.

Keeping the low end in mono

Low frequencies are the worst place for width. They carry most of a mix's energy, so cancelling them in mono takes a lot of level with it, and your ears are poor at telling where very low sounds come from, so spreading them buys you little in stereo. The usual practice is to keep everything below about 100 to 150 Hz in the centre.

Vinyl adds another reason. A record groove moves side to side for the mid and up and down for the side, and loud out-of-phase bass makes the vertical movement big enough to cause trouble in cutting and playback. Mastering engineers cutting vinyl often sum the lowest frequencies to mono for that reason.

Many widener plug-ins have a crossover so they only widen above a frequency you set. You can do the same with a mid/side EQ by cutting the side below 120 Hz or so, which the mid/side EQ lesson shows. Pick Bass in the lab and press Mono to hear why it's worth doing: that loop measures +0.47, a lot of side for a bass line.

How to make a mix wider, step by step

Width comes most reliably from the arrangement and the panning. Wideners come last, on the parts that can afford them.

1. Put a mono button on the masterIn Ableton Live, add Utility as the last device on the master and set Width to 0% when you want to check. In Logic, the Gain plug-in has a Mono button. Leave it off while you work, and press it often.
2. Keep the centre solidPan the kick, snare, bass and lead vocal to the middle. These are what a mono listener needs to hear.
3. Pan everything else with a reasonSpread the hats, percussion, guitars and keys across the field so they stop crowding the centre parts. Hard left and hard right are fine for doubled parts.
4. Record or programme doublesTwo takes panned apart give width that survives mono, and on a rhythm guitar or a backing vocal they beat any plug-in.
5. Widen the parts that can take itPads, reverb returns and effects can go to 120 or 150% on a mid/side width control. Keep the low end out with the plug-in's crossover or a side EQ.
6. Compare at matched levelMore side means more level in stereo, so turn the widened version down until it's as loud as the bypassed one before you judge it.
7. Check it in monoPress the mono button through the loudest section. The centre parts should keep their level and tone; if one goes thin or hollow, find the widener or delay responsible.

How wide should each part be?

Common stereo width mistakes

Widening the bass

A widener or chorus across a bass sounds enormous on headphones, and then the low end goes thin or vanishes on a mono system. Widen a higher layer of the sound and keep the sub in the centre.

Judging width by level

In the lab, the full mix at 200% width is about 1.9 dB louder in stereo than at 100%, and louder usually sounds better at first. Turn the wide version down to match before you decide it helped.

Making everything wide

Width is a contrast between what's in the middle and what's at the edges. Put a widener on every track and nothing stands out to the sides any more; the centre loses its weight and the whole mix sounds vague.

Trusting headphones alone

Headphones give each ear one channel with nothing leaking across, which exaggerates width. On speakers the channels blend in the room, so a mix that feels too narrow on headphones is often fine on monitors.

Never pressing Mono

A widener can sound great in stereo and wreck a part in mono, and the only way to find out is to press Mono each time you add one.

Train your ears to hear width

The quiz under the lab plays a loop with hidden settings and the screens covered, then asks whether it will survive mono. Some rounds only change the width, which never hurts the mono version; others add a delay or a flipped channel, which does. Every round is level matched, so you're judging the stereo picture alone. Listen for a centre that feels solid against one that seems to come from everywhere at once.

Questions people ask

Keep going