Pan when you want to put a sound somewhere. It works on any part and costs nothing in mono, which is why almost everything in a mix should be placed with it.
Reach for a Haas delay when a mono part, such as a single guitar or synth line, sounds too narrow and panning it would only move the narrow sound somewhere else. Keep it off the bass and the lead vocal, and listen to it in mono before you commit.
At a glance
| Property | Panning | Haas delay |
|---|---|---|
| What differs between the speakers | Level | Time |
| Usual setting | Anywhere from hard left to hard right | 5 to 30 ms on one side |
| To put a sound fully on one side | About 15 to 20 dB of difference | About 1 ms |
| Width | None; the sound stays a point | Wider, with no echo on most sounds |
| Correlation on a mono part | Stays at +1 | Falls, and can dip below 0 |
| In mono | Level may shift up to 3 dB, tone unchanged | Comb filter: a row of notches |
| On headphones | A level difference between the ears | Can feel pinned to the earlier ear |
| Best for | Placing every part of a mix | Widening a few mono parts above the bass |
How panning places a sound
A pan control splits one signal between two speakers and sets how much goes to each. Turn it left and the left speaker gets more while the right gets less, and both play the sound at exactly the same moment. Your ears hear the louder speaker as the direction, so the sound sits somewhere between them, closer to the loud one.
It takes a surprisingly large difference to move a sound all the way across. Somewhere around 15 to 20 dB between the speakers puts it at one of them, and a few dB only nudges it off the centre. Hard left on the pan control of a mono track means the right speaker gets nothing at all.
How much each side drops when the sound sits in the middle is set by the pan law, which stops a sound getting louder or quieter as it moves. Most DAWs default to −3 dB or close to it. The details, and what each law does in mono, are in panning laws.
How the Haas effect places a sound
The Haas effect, also called the precedence effect, is the way your hearing treats two copies of a sound that arrive close together. Within about 30 ms for speech and most music, and less for sharp clicks and drum hits, it fuses them into one sound and takes its direction from whichever copy came first. Play a mono part from both speakers at the same level, delay one side, and the part moves towards the side that plays first.
Small delays move it the way a pan does. By about 1 ms the sound has gone most of the way to the early speaker, and past that it stops moving and starts to spread. Between roughly 5 and 30 ms you hear one sound that seems wider than the speaker it comes from, which is the range people mean by a Haas delay. The Haas effect lesson covers the history and the echo threshold in full.
How your ears use level and time
Your ears are about 18 cm apart, and they use two clues to tell where a sound is. A sound from the left reaches the left ear first, by up to about 0.7 ms, and it’s also louder there, because your head blocks some of it on the way to the right ear. Timing gives the clearer clue for low frequencies. The shadow of your head works better on high ones, since short wavelengths can’t bend round it as easily; Lord Rayleigh described this split between the two clues in 1907.
Speakers complicate it, because both your ears hear both speakers. A level difference between the speakers turns into a timing difference at your ears, mostly at low frequencies, which is why a pan pot can place a sound convincingly with level alone. Alan Blumlein’s stereo patent of 1931 was built on that idea, and almost every pan pot since works the same way.
A Haas delay starts from a timing difference between the speakers. Up to about 1 ms it works like a pan and moves the sound. Past that the sound can’t go any further than the early speaker, so the extra delay stops changing where it is and makes the two speakers play signals that no longer match, which your ears hear as space around the sound.
Why a delay sounds wider than a pan
A panned mono part is the same waveform in both speakers, just at two levels. At every instant the two speakers push and pull together, so the sound stays a single point wherever you put it. The correlation meter shows this as +1, and the goniometer as one straight line that tips towards the louder side.
A delayed part puts a different moment of the waveform in each speaker. At 10 ms the right speaker is still playing what the left played a hundredth of a second ago, and over a whole guitar chord the two only partly line up. That’s what you hear as width, and the correlation meter drops to show it: on the guitar in the lab it reads about +0.5 at 10 ms, and on the drum loop at 20 ms it gets close to 0. A real double-tracked part gives you the same kind of width, because two separate takes never quite match.
What panning and the Haas effect do in mono
Mono adds left and right together and plays the sum from one speaker, or from both at once. A single Bluetooth speaker does this, and so do many club systems and some phones, so every stereo trick you use gets heard in mono by someone.
A panned part survives because both copies are in step. Add a loud copy of a waveform to a quiet copy of the same waveform and you get a louder copy of it; the tone can’t change, only the level. With the −3 dB pan law, a centred sound comes out about 3 dB louder in mono than one panned hard to a side, which nudges the balance a little.
A Haas part summed to mono is a sound added to a late copy of itself, which is a comb filter. Any frequency the delay puts exactly half a cycle out of step cancels. The first notch lands at 1 ÷ (2 × delay), and the rest follow every 1 ÷ delay above it. With 10 ms that means notches at 50 Hz, 150 Hz, 250 Hz and on up every 100 Hz, about 200 of them before 20 kHz, and the part goes thin and phasey.
| Setting | Left against right | Correlation | In mono |
|---|---|---|---|
| Pan L 60 | About 10 dB louder on the left | +1.00 | About 1 dB quieter, no notches |
| Delay R 0.5 ms | Same level, right 0.5 ms late | +0.60 | About 1 dB quieter, notches from 1 kHz, every 2 kHz |
| Delay R 10 ms | Same level, right 10 ms late | +0.47 | About 1.3 dB quieter, notches from 50 Hz, every 100 Hz |
The overall level in mono hardly differs between the three, so a level meter won’t warn you about a Haas delay. The damage is in the tone, and short delays are no safer than long ones. Half a millisecond puts the first notch at 1 kHz, in the middle of the range where a guitar or a voice is easiest to hear, and at that spacing each notch is wide enough to change the colour of the whole part.
Panning and delaying a stereo sound
Everything above assumes a mono part, one signal sent to both speakers. A stereo track, like a drum loop or a synth pad printed in stereo, already has different material on each side, and both tools behave differently with it.
On most stereo tracks the pan control works as a balance control. Turned left, it lowers the right channel, and some panners, including the one in this lab, also fold part of the right channel into the left so nothing disappears. Nothing is delayed either way, so a stereo track that was safe in mono stays safe after you pan it. Ableton Live and Logic Pro both have a split or stereo pan mode as well, which sets where each channel lands and narrows a wide sound towards one side more cleanly.
Delaying one side of a stereo recording rarely helps. It’s already wide, so the extra time mostly smears the image and adds a comb filter to whatever the two sides shared, which on a drum loop is usually the kick and snare.
When to use panning and when to use the Haas effect
| Job | Panning | Haas delay |
|---|---|---|
| Place drums, keys and guitars across the mix | Yes | No |
| Kick, bass and lead vocal | Keep them centred | No |
| Widen one mono guitar or synth line | Only moves it | Yes, 10 to 20 ms, then check mono |
| Backing vocals with no second take | Pan pairs if you have them | Works; check it in mono |
| Short percussion like shakers and claps | Yes | Short times only, or it flams |
| A mix that has to work in mono | Safe | Test every one |
Pan almost everything. It places a part without changing it, so you can set up a whole mix with pan pots and know the mono version will sound like the same mix with a slightly different balance.
The Haas delay is for the odd mono part that should feel bigger than one point. If you can record a second take, do that and pan the two takes apart, because two performances give you width without a comb filter. When all you have is one take, a Haas delay is a fair substitute on a part that sits above the bass and isn’t the focus of the song.
How to use a Haas delay and panning together
The two work well side by side: the delay gives the part its width, and level, the thing panning controls, decides where the widened part sits.
- Split the partDuplicate the mono track. Pan the original hard left and the copy hard right, so you can adjust each side on its own fader.
- Delay one sideDelay the copy by 10 to 15 ms with your DAW’s track offset, such as Track Delay in Ableton Live, or with a plain delay plug-in set to one repeat, no feedback and 100% wet.
- Correct the leanThe part will sit towards the side that plays first. Pull that side down a few dB at a time until it sits where you want it; getting it near the centre can take as much as 10 dB.
- Keep the low end outPut a high-pass filter on the late side, somewhere around 150 to 300 Hz, so the lowest notches have nothing to cancel.
- Listen in monoSum the whole mix to mono and listen to the part in context, with nothing soloed. If it goes hollow, try a slightly different delay or turn the late side down.
Any level difference between the two sides also makes the mono notches shallower. With one side 6 dB below the other they’re about 10 dB deep instead of complete, which is often enough to hide them. Bringing the two pan controls in from the edges also changes the lean, but each speaker then plays some of both copies, which puts the comb filter into each speaker in stereo, so keep them hard left and right unless the mix needs otherwise.
Common mistakes
Using a delay where a pan would do
If the goal is to put a part on the left, pan it there. A Haas delay used only to move something costs you a comb filter in mono for a result the pan pot gives for free.
Folding a Haas pair with one pan control
Put a Haas delay inside a stereo track and then turn that track’s balance or stereo pan control, and the late side gets folded into the early one. The two copies now meet inside one speaker, so the comb filter is audible in stereo too.
Judging it on headphones only
Headphones send each side to one ear with nothing leaking across, so a Haas delay can sound quite different from how it does on speakers, often pulled harder towards the early side. Check the part on speakers, or at least in mono, before you decide how much delay it needs.
Mixing from off-centre
Both tools assume you sit halfway between the speakers. Sit about 34 cm to the left of centre and the left speaker’s sound reaches you roughly 1 ms before the right one’s, which is enough by itself to pull the whole mix that way, so any panning or Haas setting you judge from there will be off.
Train your ears
The quiz under the lab plays a loop with hidden settings and asks whether it will survive being played in mono. In the rounds that fail, the right side is a millisecond or so late or flipped upside down; in the others only the width changes, which mono ignores. The screens are covered, so you judge by the picture the sound makes between your speakers or in your headphones. A short delay tends to come across as a lean with a slightly phasey edge, and a safe round sounds solid in the middle.
Questions people ask
Is the Haas effect better than panning?
They do different jobs. Panning places a sound and leaves it alone in mono, so it suits almost everything in a mix. A Haas delay places a mono part and makes it sound wider, which panning can’t do, but it turns into a comb filter in mono, so it suits a few higher parts that you have checked.
Does panning cause phase problems?
Not on its own. A pan control changes the level in each speaker and delays nothing, so the two speakers stay in step and nothing cancels when they’re added in mono. The level can change, though: compared with how they balanced in stereo, a centred sound comes out up to 3 dB louder in mono than one panned hard to a side. That shifts the balance of the mix and leaves the tone alone.
How many milliseconds equal a hard pan?
Roughly 1 ms on a pair of speakers. Around that point a delay has moved a centred sound most of the way to the speaker that plays first, and adding more delay widens the sound instead of moving it further. It takes something like 15 to 20 dB of level difference to move it the same distance with a pan control.
Can I pan a Haas-delayed part?
Yes, and it’s a good way to correct the lean. Turn the early side down a few dB, or move the two sides with the pan controls on their own tracks. Avoid a stereo pan or balance control on a single track that carries both sides, because folding the delayed side into the other channel puts the comb filter inside one speaker.
Why does a Haas delay sound wider than panning?
Because the two speakers play different moments of the waveform, and your ears hear that difference as space around the sound. A panned part plays the same moment from both speakers, just louder in one, so it sounds like a single point placed somewhere between them.
Should I use the Haas effect on headphones?
You can, but check it on speakers too. On headphones each ear hears only one side, so a delay of more than about 0.7 ms is longer than any gap a real sound could make between your ears, and the part can feel pinned to the earlier ear. Panning on headphones gives a level difference between the ears, which sounds closer to the speaker version.