Reverb & Delay,
Demystified

What reverb and delay actually are

Reverb simulates the reflections of a real or virtual space: thousands of overlapping echoes decaying over hundreds of milliseconds to several seconds. Delay is a discrete repeat of a signal, offset in time: feed the output back into the input and the echo repeats itself, each pass slightly quieter.

Panning places elements left and right. Volume places them forward and back. Reverb and delay place them in a room.

Reverb anatomy: every parameter explained

Pre-delay

The gap in milliseconds between the dry sound and the start of the reverb tail. Without it, reverb smears onto the attack transient immediately. With 20–30 ms, the transient lands clean before the reverb kicks in, so the element stays present even with the reverb level set high.

Early reflections

The first discrete echoes (5–80 ms before the dense tail builds). Tightly spaced reads as a small space; widely spaced reads as a large hall. High early reflection level pushes the source forward; low level sets it back. On most plugins: an “ER” or “Early Mix” slider.

Decay time / RT60

How long the tail takes to drop 60 dB from its peak. The main character control.

Long decays (1.5 s+) belong on pads and atmospheres. Percussion, bass, melodic leads: 0.3–1.2 s. Long reverb on fast elements buries transient detail and kills the groove.

Diffusion

Diffusion controls how quickly reflections in the reverb tail blend into a smooth wash. Low diffusion: individual echoes stay distinct. You can hear them as separate repetitions before they blur. High diffusion: the tail is immediately smooth and dense, no discrete echoes.

Damping

Simulates high-frequency absorption as the tail decays. Low damping = bright and clinical. High damping = warm, dark tail. Most musical material wants medium-high damping.

Modulation

Subtle pitch fluctuation that prevents metallic ringing in the tail. Without it, fixed delay line lengths in the algorithm produce artificial resonances. Too much and you get obvious pitch-wobble on held notes.

Algorithmic vs. convolution reverb

Algorithmic reverb

Generates a synthetic reverberant field from DSP: delay lines, allpass filters, comb filters. No specific physical space captured; everything adjustable in real time. Automate decay, modulate size, perform it live. That flexibility is the whole point.

Convolution reverb

Captures the acoustic behavior of a real space via an impulse response (IR), recorded as a balloon pop or sine sweep in the actual room. Convolving your signal with the IR applies that space's acoustic fingerprint to any audio. Far more realistic than algorithms. Hundreds of IR libraries exist.

For electronic music, algorithmic is the workhorse. Convolution earns its place on acoustic instruments, film scoring, or when you want the specific character of a legendary plate that someone's already IR-captured.

Reverb in a mix: what actually works

Rule 1: reverb goes on a send, not an insert

Insert reverb on every track puts each element in its own private space, and the mix ends up sounding like a stack of isolated recordings rather than a performance. Send/return uses one shared reverb instance, so everything that routes to it shares the same acoustic space.

Send/return setup in Ableton Live:
  1. Create a new audio return track (Cmd+Option+T or right-click → Insert Return Track)
  2. Insert the reverb plugin on the return track as the first effect
  3. Set the reverb's dry/wet to 100% wet — the dry signal stays on the original track
  4. On each track that needs reverb: click "Send" for the return track, set the amount
  5. Adjust the return track's fader to set the overall reverb level in the mix

Send/return setup in Logic Pro:
  1. In the Mixer, click the + icon to add a Bus
  2. Add your reverb plugin to the Bus channel strip
  3. Set the reverb's mix to 100% wet
  4. On each channel you want to reverb, click a Send and route it to the same Bus
  5. Adjust the Send amount per channel and the Bus fader for overall level

Key principle: reverb return fader controls how wet the whole mix sounds.
Individual send amounts control how much each element is processed.

Pre-fader vs. post-fader sends

Post-fader (the default): the reverb send tracks the channel fader. Pull the channel down and less goes to reverb. Pre-fader: the send is independent of the fader, so you can pull the dry signal to near-silence while the reverb tail keeps ringing. That's the “ghost reverb” effect on drum breakdowns. Use it deliberately.

Different reverbs for different elements

Delay anatomy: every parameter

Delay time

The interval between the original signal and its first repeat, expressed in ms or a tempo-synced note division. The formulas:

Quarter note:     ms = 60000 / BPM
8th note:         ms = 30000 / BPM
16th note:        ms = 15000 / BPM
Dotted quarter:   ms = 90000 / BPM
Dotted 8th:       ms = 45000 / BPM
Quarter triplet:  ms = 40000 / BPM
8th triplet:      ms = 20000 / BPM

Example at 128 BPM:
  Quarter note:   468.75 ms
  Dotted 8th:     351.56 ms   ← the most musical delay time
  8th note:       234.38 ms
  16th note:      117.19 ms

Use the calculator below to get every subdivision at any BPM. Click a row to copy the ms value straight to clipboard.

Feedback

The percentage of delay output fed back into the input. 0% = single echo. 50% = 6–8 fading repeats. 99% = near-infinite sustain. 100%+ = runaway buildup: the dub effect.

Wet/dry mix (on a delay insert)

Same principle as reverb: use sends where possible, 100% wet on the return. For slapback or doubling, a low insert mix (15–30% wet) is fine.

Filter on the feedback loop

A LPF and HPF applied to the feedback path (not the output). Each repeat passes through the filter before feeding back, so the tonal character changes with each repetition. Set the LPF to 4–8 kHz and repeats get progressively darker and warmer. That's the tape echo character: the Roland Space Echo RE-201's HF rolloff on every pass through the heads.

Delay types and uses in electronic music

Slapback delay

Single short repeat (50–150 ms, 0% feedback). Adds a room-sized quality to dry vocals without the wash of reverb. Standard in hip-hop and trap.

Slapback settings:
  Delay time:   60–120 ms (unsynced, or roughly 1/8 note at slow tempos)
  Feedback:     0% (single repeat only — no trail)
  Filter:       Slight LPF at 8–10 kHz on the wet signal (removes harshness)
  Wet level:    15–25% wet on an insert, or moderate send on a return
  Stereo:       Keep mono, or use L/R slightly offset (50 ms left, 65 ms right) for subtle width

Dotted 8th delay — the melodic delay

Repeats land between the downbeats, giving melodic lines a syncopated quality that reinforces the groove. Works on synth arpeggios, plucked bass, and leads.

Ping-pong delay

Alternates repeats between L and R. Short time (8th/16th note) for tight movement, quarter note for a wider sweep. In a dense arrangement it adds stereo energy fast. Keep the source centered so the delay has room to move.

The Haas effect — delay as stereo width

Below ~35 ms the ear fuses the delay with the source and reads it as directional width, not a distinct echo. Copy a mono signal to two channels, offset one by 15–30 ms, and it reads as stereo width without actual stereo separation.

Dub delay

High feedback, heavily filtered feedback path, and live automation of the send and feedback levels. Cascading echo that dissolves into a warm wash.

Dub delay settings:
  Delay time:     Quarter note or dotted 8th, synced to BPM
  Feedback:       65–85% (long trail without complete runaway)
  Feedback LPF:   2.5–5 kHz — heavy rolloff makes each repeat significantly darker
  Feedback HPF:   80–150 Hz — prevents low-end buildup in the delay trail
  Wet level:      25–50% on a return — the trail should be prominent, not buried
  Modulation:     Light time modulation (rate 0.3–0.8 Hz) for analog wobble character

Dub technique — live automation:
  1. Start with feedback at 60%, delay mostly dry
  2. Push the send level up to 80–100% on a single hit or phrase
  3. Bring feedback to 85–90% momentarily — allows the trail to bloom
  4. Pull the feedback back to 50% to tame the trail
  5. Cut the send back to normal level
  → Creates the characteristic "bloom and dissolve" of classic dub production

Reverb and delay together: how they interact

The chain order matters: delay before reverb

Delay → reverb: discrete echoes feed into the reverb, each dissolving into its own reverb cloud. Cohesive and three-dimensional.

Reverb → delay:the reverb tail gets delayed too, so washy blobs stack on each other. Usually sounds like a mistake unless that's specifically what you're after.

Pre-delay on reverb as natural delay

Pre-delay set to a note-synced value with a strong early reflections mix gets you the “delay into reverb” character inside a single unit. Pre-delay acts as the discrete echo; the tail is the space it dissolves into.

Using both for layered depth

For leads and vocals: dotted 8th delay at moderate feedback (send at −10 to −15 dB) builds the rhythmic structure, feeding into a medium room reverb (0.8–1.2 s) for acoustic space. The delay keeps it rhythmically alive; the reverb puts it in a room.

Practical recipes

Starting points to tune by ear. All assume send/return routing, reverb at 100% wet unless noted.

Pad reverb — wide, lush, enveloping

Reverb type:   Large hall or algorithmic (Valhalla Room, Fabfilter Pro-R)
Pre-delay:     0–8 ms (pads sustain — pre-delay not critical)
Decay (RT60):  2.5–4.0 s
Early refl:    Low-medium mix (0.3 on a 0–1 scale)
Diffusion:     85–100% (totally smooth tail)
Damping:       50–65% (warm, not overly dark)
Modulation:    Rate 0.8 Hz, depth 12%
Size:          Large (80–100%)
Send level:    −6 to −10 dB relative to dry pad level
Wet on return: 100%

Notes: pads can absorb more reverb than almost any other element because their
slow attack and long sustain hide the onset of the reverb tail. Use a high-pass
on the reverb return at 80–100 Hz to prevent low-end buildup in the return channel.

Snare room — tight, punchy, present

Reverb type:   Plate (Valhalla Vintage Verb plate mode, UAD EMT 140, Fabfilter Pro-R plate)
Pre-delay:     5–15 ms (preserves the snap of the transient)
Decay (RT60):  0.8–1.4 s
Early refl:    High mix (0.6–0.8)
Diffusion:     60–75%
Damping:       55–70% (reasonably warm)
Size:          Medium (40–60%)
Send level:    −8 to −12 dB relative to dry snare

Notes: plate reverb has a particular density and brightness in the 1–4 kHz range
that makes snares sound larger without becoming distant. The decay must be short
enough that the reverb tail has substantially decayed before the next snare hit —
at 130 BPM (quarter note = 461 ms), a 1.4 s decay works but is right at the edge.

Vocal delay — rhythmic, present, musical

Delay type:    Tempo-synced, dotted 8th note
BPM:           Match track BPM (use calculator above)
Feedback:      30–45% (3–5 audible repeats)
Feedback LPF:  6–8 kHz (slightly warm repeats without muddying)
Feedback HPF:  150–200 Hz (prevents bass buildup in the delay trail)
Wet (send):    −14 to −18 dB relative to dry vocal
Stereo:        Mono delay, or very subtle L/R offset (within ±5 ms) for width

Notes: vocal delay should not fight for center-stage with the dry vocal. The rule
is that the repeat should be noticeable when you listen for it, but should not
compete with the intelligibility of the next lyric. Start with the send level lower
than feels right and bring it up. You will almost always find the correct level is
lower than your first instinct.

Dub delay — feedback-heavy, filtered, performative

Delay type:    Tape-style or analog delay emulation (Return/RC-20/Soundtoys EchoBoy)
Delay time:    Quarter note (tempo-synced)
Feedback:      70–80%
Feedback LPF:  3.0–4.5 kHz (each repeat darkens significantly)
Feedback HPF:  120–180 Hz (control low-end buildup)
Modulation:    Rate 0.4–0.7 Hz, depth 15–25% (gentle wow/flutter)
Saturation:    Light — a small amount of tape-style saturation on the feedback path
Send level:    Variable — automate or perform during the track

Notes: dub delay is most effective as a performative effect rather than a static
setting. Automate the send level or feedback in real time — push the send up
briefly on a fill or transition phrase, let the trail bloom and decay, then pull
back. This technique is more effective at creating energy and movement than any
static setting can achieve.

Cheat sheet

Reverb parameter quick reference

Delay parameter quick reference

BPM to ms formula

The three rules