What is reverb pre-delay?

Pre-delay in plain English

A reverb takes in your sound and sends out a wash of reflections. Pre-delay holds that wash back for a moment before it starts. The dry sound plays straight away and the reverb follows a few milliseconds later, so for that short time you hear the sound on its own.

The gap is usually somewhere between 10 and 80 ms, well under the length of a single syllable, so you rarely hear it as silence. What you notice is that the front of each word or drum hit comes through sharper, and the sound seems nearer to you even though the reverb is just as long and just as loud as before.

Almost every reverb has the control, labelled Pre-delay or Predelay and set in milliseconds; Ableton Live's Reverb and Valhalla VintageVerb both have one on the front panel. The other reverb controls are covered in what is reverb?, and this page stays with this one.

Pre-delay in a real room

In a real space, sound reaches you by two kinds of path. The direct sound travels in a straight line from the source to your ears. Every reflection takes a longer route, off the walls or the ceiling, so it always turns up later. The gap between the direct sound and the first reflection is the room's own pre-delay. Concert hall acousticians call it the initial time delay gap, a term Leo Beranek introduced in his 1962 book Music, Acoustics and Architecture.

Sound travels about 34 cm in a millisecond, so you can work the gap out. Say you're 3 metres from a singer and the first reflection to reach you, off the nearest side wall, covers 10 metres on its way. That's 7 metres further than the direct sound, and 7 ÷ 0.343 gives a gap of about 20 ms. Now move to the back of the hall, 20 metres from the stage. The first reflection, off a wall close to your seat, might travel 23 metres, only 3 metres more than the direct path, and the gap shrinks to about 9 ms.

So the gap tells your ears where you are. A long one means you're close to the sound and the walls are far off, and a short one means you're as far from the source as you are from the walls. That's why adding pre-delay brings a sound forward without making the space any smaller, and taking it away pushes the sound back into the room.

Reading the lab

The lab is set up for this one control. Pre-delay, decay, mix and the type of space are free to move. The other knobs, which set the tone and size of the space, are dimmed so the gap stays the only thing changing as you go through the steps, though each step still sets them for you.

On the impulse response screen, the pre-delay is the shaded strip on the left, ending in a dashed line with its length written in milliseconds. Everything the reverb does starts to the right of that line: the amber early reflections and the coloured tail. Turn the pre-delay knob and the whole shape moves later while the reverb itself stays exactly the same, which is the clearest picture of what the control does. Even at zero there's a short space before the first amber line, about 15 ms in this hall, because the walls of any space are some distance away; pre-delay adds its gap on top of that. The time axis is stretched at the start, so 20 ms takes up more of the width than you might expect, and the shape looks a little squeezed as it moves right.

On the big screen, watch the start of each phrase or hit. With no pre-delay the coloured reverb rises together with the pale dry shape, and with 50 ms or more it starts visibly later. The Right now panel describes the setting you last touched for the sound you've picked, and A/B switches the reverb off at matched loudness. You can also upload your own sound or record eight seconds with your microphone; it stays on your device.

Keeping the front of each sound clear

The front of a sound carries a lot of what makes it recognisable. In a vocal that's the consonants, such as the t of “take” or the k of “cold”. They're short and much quieter than the vowels around them, which makes them the first thing a reverb covers up. A drum's transient is loud, but it's over in a few milliseconds, and a reverb that starts on top of it blurs its edge.

With no pre-delay, the reverb is already building while the consonant or the transient is still sounding, so the two overlap and the reverb masks part of it. Give the reverb a 20 to 50 ms head start and the front of the sound has finished before the wash arrives.

Your hearing helps here. When a copy of a sound arrives within roughly 30 ms of the original, your brain folds the two together and hears a single sound, coming from wherever the first one came from. This is the precedence effect, and the Haas effect lesson covers it. It's why 20 ms of pre-delay sounds like a cleaner start to the same sound and never like a gap. A reverb is quieter and more spread out than a clean copy, so the folding holds for longer than that; somewhere around 80 to 100 ms it gives way and the reverb turns up as an event of its own.

How much pre-delay should you use?

Use the shortest setting that gets the front of the sound clear of the reverb. For a lead vocal in a hall or plate that's often 30 to 60 ms; for drums it's usually less. This table gives you a sense of what each range sounds like.

A long, dense reverb usually wants more pre-delay than a short one, because there's more wash sitting on the front of every sound. Tempo pulls the other way. In a fast, wordy part, a pre-delay about as long as one syllable means the reverb from each word starts right on the front of the next, and the blur has only moved along by one word. A syllable in normal speech lasts around a fifth of a second, so on rap and busy pop verses, staying under about 60 ms is safer.

Timing pre-delay to the tempo

Short pre-delays are too brief to land on or off the beat in any way you'd hear, so set anything under about 30 ms by ear. Once the gap gets longer it becomes part of the rhythm, and it sounds tidier on a note value of the song.

One beat lasts 60,000 ÷ BPM milliseconds. Divide that by 4 for a 16th note, by 8 for a 32nd and by 16 for a 64th. At 120 BPM a beat is 500 ms, so a 16th is 125 ms, a 32nd is 62.5 ms and a 64th is 31.25 ms. The 32nd and 64th are the useful ones for pre-delay.

Find the clearest setting by ear first, then move it to the nearest value in the table. If your reverb has a sync option you can pick the note value directly. The same arithmetic sets delay times, and the tempo-synced delay lesson goes through dotted and triplet values as well.

Pre-delay on drums

On drums, pre-delay decides whether the reverb sits behind each hit or on top of it. A plate is the clearest case, because it has no early reflections and starts at full density almost at once. With zero pre-delay that density lands on the crack of the snare and rounds it off. Ten to 30 ms lets the crack through first, and the snare sounds harder while the plate still fills the space between hits.

Go much longer on a busy drum part and each hit starts to sound doubled, a bit like a flam, because the reverb now arrives as a second, softer attack. On drums that's rarely what you want unless the gap is on the grid. Kicks usually get little or no reverb in modern mixes, and a short room with a heavy low cut is the usual choice when they do.

The last drum button puts the gap on a 32nd note of the loop's tempo, which is the grid from the table above.

Pre-delay, early reflections and room size

These three get confused because they all move things around in time. Pre-delay shifts the whole reverb later and leaves its shape alone. Size, on most algorithmic reverbs including this one, spreads the early reflections further apart and slows the build-up of the tail, so it changes the shape of the space itself. The early reflections are the first separate echoes inside that shape, and the early reflections lesson looks at them on their own.

In a real room the gap and the reflections come as a package, because both depend on where the walls are. A reverb lets you pull them apart. A small, tight room with 60 ms of pre-delay is a combination no real room gives you, and it can suit a vocal you want close and clear with only a hint of space. Decay is separate again: it sets how long the tail lasts once it has started, which the decay time lesson covers.

How to set pre-delay, step by step

1. Set up the reverb with pre-delay at zeroChoose the space and the decay, and set the level with the whole mix playing.
2. Listen to the front of the soundFocus on the first consonant of each line, or the crack of each snare hit, and how much the reverb blurs it.
3. Raise the pre-delay slowlyGo up in steps of 5 to 10 ms until the front comes clear of the reverb.
4. Back off if the reverb comes awayIf you start to hear the reverb as its own echo, take 10 to 20 ms off.
5. Snap long settings to the tempoPast about 50 ms, move it to the nearest 32nd or 64th note from the table above.
6. Check the reverb level againWith the words clear of the reverb, it can often come up a decibel or two.

Using more reverb without losing the words

Pre-delay is what lets a vocal carry a lot of reverb and still sound up front. Without it, every extra decibel buries the consonants a little further. With 60 to 80 ms the words arrive first and the hall comes in behind them, so the reverb can go much louder before the vocal starts to sound distant. The last step of the lab sets this up; drag the pre-delay back to zero at the same mix and the words sink.

Ducking works well alongside it. Compress the reverb return with its sidechain fed from the dry vocal, and the reverb dips under the singing then swells back in the gaps. Pre-delay clears the front of each word; ducking clears the rest of the phrase. The sidechain compression lesson shows how to set it up, and the ducked delay lesson does the same thing for echoes.

If a reverb has no pre-delay control, you can add one. Put the reverb on a return track, as described in insert vs send, and put a delay plug-in in front of it set to 100% wet, zero feedback and the time you want. That's how studios did it before digital reverbs had the control built in: the signal went to a tape machine first, and the playback, slightly late, went on to the plate or the echo chamber.

Common pre-delay mistakes

So much that it turns into an echo

Past about 100 ms the reverb stops sounding like part of the space and starts sounding like a copy of the sound. That can be a good effect, but if you didn't mean it, bring the pre-delay down until the reverb attaches to the voice again.

Judging it with the track soloed

On its own, a 60 ms gap can sound like an obvious slap. With the rest of the track playing, other parts fill it and the same setting is barely noticeable. Set the length with the mix running, the same as the reverb level.

Using pre-delay to fix a muddy reverb

Pre-delay clears the front of each sound, but a reverb that clouds the whole mix is usually full of low end. A low cut on the reverb return, somewhere between 200 and 500 Hz, does far more for that than any amount of pre-delay.

Giving every reverb the same gap

A short drum room and a long vocal hall want different settings. If several parts share one reverb and only some of them need a longer gap, set up a second return with its own pre-delay.

Training your ears to hear pre-delay

The quiz under the lab plays a hall with either no pre-delay or 60 ms and asks which one it is. The rounds are set to the same loudness, so all you have to go on is the front of each sound. With no gap the reverb starts on top of it and the consonant or the crack of the hit comes through softer; with 60 ms it arrives clean and the hall follows behind. Once you can hear that soft front, you'll notice it on records too. Your best score out of eight is saved in your browser.

Questions people ask

Keep going