Alright, let’s build a dub delay return in Ableton Live that hits like classic drum and bass ear candy… without turning your drop into a blurry soup.

This lesson is all about dub delay send automation with restraint. Meaning: we’re not leaving delay on. We’re doing quick, intentional throws that feel like performance gestures. Think of it like flicking an aux send on a DJ mixer for one word, one stab, one fill… then getting out of the way so the drums and bass stay arrogant and clean.

We’re aiming for that call-and-response vibe you hear in jungle, rollers, and half-time: little echoes that answer the phrase, create motion, and pull you into the next section.

First, quick context. This works best in a typical DnB project: 172 to 176 BPM, a drum bus with kick and snare doing the heavy lifting, hats and breaks adding energy, and a bass group that you really don’t want smeared by time-based effects. If you need something to test, grab a vocal chop or a chord stab and place it right at the end of a 4-bar phrase. That’s the perfect target for a throw.

Now step one: create the return track.

Go to Create, insert return track. Rename it DUB, all caps if you want it to feel official. Set the return fader to 0.0 dB. That’s important. We’re going to keep the return consistent, and control intensity from the sends and automation, not from constantly riding the return fader like it’s a volume fix.

Now build the device chain on the return. Stock Ableton devices only.

First device: Echo. This is the heart.
Set it to Sync mode. Start with a time of one quarter note for that classic dub spacing. If you want more jungle urgency later, you can try three-sixteenths or one-eighth dotted, but start with one-fourth so you can clearly hear what your automation is doing.
Feedback: keep it conservative, around 20 to 35 percent. We’re building something you can trust.
Dry/wet: 100 percent, because this is a return.
For character, keep Noise low, and keep Wobble very low, like 0 to 5 percent. Just enough to feel alive, not enough to sound seasick.

Second device: Auto Filter.
Set it to band-pass. This is the secret sauce for “real dub” behavior, because real dub delays are rarely full-spectrum. They’re mid-focused, like a separate little radio in the background.
Aim the frequency somewhere like 800 hertz up to 3.5k as your general zone. Set resonance around 1.2 to 2.0. Do not let it whistle. A tiny bit of drive is fine, maybe 2 to 5 dB, but keep it tasteful.

Third device: Saturator.
This thickens the repeats so they feel intentional instead of thin and plasticky.
Use Analog Clip or Soft Sine. Drive around 2 to 6 dB. Then trim the output so you’re not tricking yourself with loudness. You want richer, not louder.
And yeah, turn on Soft Clip if the repeats sometimes spike.

Fourth device: Compressor, with sidechain.
This is how we keep the delay from stepping on the kick and snare.
Enable sidechain. Choose Audio From your drum bus or kick and snare group.
Ratio around 3 to 1.
Attack: 5 to 15 milliseconds. That lets a little transient through so it still feels punchy.
Release: 80 to 160 milliseconds. You want a groove-y duck, not a long wash.
Then set the threshold so you’re getting about 2 to 6 dB of gain reduction when the drums hit.

Optional fifth device: Reverb, but only as micro-space.
Small to medium size, decay around 0.6 to 1.2 seconds, high cut somewhere like 5 to 8k, and keep the wet low, like 10 to 20 percent. You’re not trying to add a reverb layer; you’re just making the echoes feel placed in a space.

Quick teaching note: this order matters. Echo creates the repeats, filter shapes them, saturation gives attitude, sidechain keeps drums dominant, and reverb is optional glue.

Now step two: safety. Dub delays can get out of hand fast, especially when you start automating sends or feedback.
At the very end of the return chain, add a Limiter. Set the ceiling to minus 0.5 dB. This is your “oh no” button. It doesn’t replace good gain staging, but it saves you from a feedback accident that suddenly nukes your master.

Step three: decide what deserves delay.
In drum and bass, delay is best on vocal chops, stabs and chords, little percussion fills, maybe a snare hit here and there.
What you generally avoid is sending your sub or your full bass bus to delay. That’s how you lose weight and get low-end mud and weird phasey movement.
The rule of restraint: delay throws should feel like events. Punctuation. Not a constant layer.

Now the core technique: send automation throws.

Pick one track to start. Let’s say a vocal chop. Find its send to the DUB return. In many sets that’s Send A, but use whatever return slot your DUB track lives on.

Set the default send level basically off. Minus infinity is fine. Or keep it super low like minus 24 dB if you want a tiny constant room, but for this lesson, I want you to feel how powerful “off by default” is.

Press A to show automation lanes. In arrangement view, choose the vocal track, then select the automation parameter: Send to DUB.

Now let’s draw a few practical shapes you’ll use all the time.

First shape: the single-word throw.
On the last eighth note or last quarter note of a phrase, ramp the send from off up to around minus 12 to minus 6 dB, just for that tiny moment, then drop it straight back down to off.
The goal is that the delay continues after the vocal stops. That’s the whole illusion. You’re not delaying the whole phrase; you’re throwing the last word into space.

And here’s an extra pro detail: don’t always make it a square spike. If it’s a vocal, try a tiny ramp up over the last sixteenth note. It feels more “thrown,” like a hand movement, not a switch.

Second shape: the end-of-4-bars echo tail.
At the end of bar 4, 8, 12, 16… do a quick send spike, maybe up to around minus 9 dB, for an eighth note or quarter note, then right back down.
This creates forward motion without adding more elements. It’s one of the cleanest ways to make your arrangement feel like it’s breathing.

Third shape: the fill enhancer.
Take a short drum fill, maybe half a bar with a rimshot or tom or glitch hit. Automate the send so it sits around minus 18 to minus 12 dB only during that fill, then back to off.
This is huge for DnB because it keeps the main groove clean while the fill gets to bloom and feel special.

One important mixing habit here: if a throw feels too loud, lower the peak send amount first, before you touch the return fader. The return fader staying stable means your whole set behaves predictably. If you keep changing the return, every other throw you made becomes a moving target.

Now, optional but powerful: one controlled “dub moment” using feedback automation.

This is not the trick you do every 8 bars forever. This is the tasteful flex you do once to lift a transition.

On the DUB return track, automate Echo’s feedback parameter.
Keep your normal feedback around 20 to 30 percent.
Then, for something like the last bar before a drop, rise to about 45 to 55 percent briefly, and snap back down before the drop hits.
That snap-back is critical. You want the hype, then silence and impact when the drop lands.

And if you do push feedback, tighten your filter. Narrow that band-pass or move it slightly higher. The danger is low-mid buildup. Dub delays love to accumulate energy, and DnB mixes punish you for that.

Next, let’s make the delay sit properly in a rolling mix.

If you need extra control, add EQ Eight before your limiter on the return and high-pass the delay. Start cutting below around 150 to 250 Hz. In heavy rollers, you can go higher than you think, even 200 to 400 Hz, because the echoes are supposed to be smaller than the dry signal on purpose. The dry drums and bass are the main character. The echo is the supporting actor.

Stereo width: keep it subtle. Echo has stereo options; you can widen a touch, or use Utility and set width to maybe 120 to 150 percent. Then do a quick mono check during the throw. If the delay disappears or gets weird in mono, narrow it back down. That’s one of those “sounds amazing in headphones, vanishes in a club” traps.

Now arrangement guidance, because restraint isn’t just sound design, it’s planning.

In an intro, you can do sparse throws on vinyl FX, chords, tiny vocal snippets. In a build, increase throw frequency slightly, and maybe do one feedback lift right at the end. In the drop, keep throws rare. Seriously. Once every 4 or 8 bars is plenty.
A good benchmark: in a full 32-bar roller drop, you might do 4 to 8 throws total. When you hear that number, you should feel the discipline. That’s the whole point.

Here are a few coach rules to make this feel intentional instead of random.

Treat sends like performance gestures, not mix knobs. If you catch yourself “setting” a send level like it’s a static effect, you’re probably overusing it.

Calibrate one reference throw first. Pick one vocal or stab, make one throw you love, and then match everything else to that energy by adjusting send peaks. This keeps your mix consistent and stops you from chasing loudness.

Also, choose one lane of attention per section. For example, in this 8 bars, only vocal throws. Next 8 bars, only stab throws. Two different sources feeding the dub return at once often reads as clutter, not hype. Especially in dense DnB drums.

Watch your danger zones: if your tops are break-heavy and busy, throws can smear transients, so avoid throwing during the busiest hat runs. And if your vocal already has a long tail, throwing it can stack into mush. Sometimes the fix is simple: shorten the clip fade, gate it a bit, then throw the tight version.

Now let’s do the mini practice exercise so you can lock this in quickly.

Build the DUB return chain we just made.
Pick two sources: a vocal chop and a stab, or a vocal and a rimshot fill.
Make a 16-bar loop with a drop at bar 9.

Automation plan:
Put vocal throws at bar 4 beat 4 and bar 8 beat 4, so the last quarter note before the phrase ends. Quick up, quick down.
Put one stab throw at bar 7 beat 3, just an eighth note leading into the drop.

Optional: automate feedback on bar 8 only, up to about 50 percent, then back down to 25 percent right as bar 9 hits.

Then export a quick loop and listen with three questions:
Can you still hear the kick and snare clearly during every throw?
Do the delays feel like punctuation rather than “always on”?
And if you mute the return entirely, do you actually miss the throws? If you don’t miss them, they weren’t doing a job.

Before we wrap, a couple advanced variations you can try later once the basic technique is solid.

You can duplicate the return and make two tempos: one DUB return at one quarter, and another at three-sixteenths or one-eighth dotted. The rule is: only use one at a time. That gives contrast without increasing density.

You can also make one throw bigger without turning it up by automating tone instead of volume. For example, open the band-pass filter slightly for that moment, or add a tiny width bump on the return. Same send peak, but it feels like it expands.

And if you want a “freeze” illusion without riding feedback, you can automate Echo output or even briefly mute the return right after the throw, like a micro-cut. When it’s filtered and saturated, it can sound like a controlled snapshot of the echo.

Recap, clean and simple.

Build a dedicated DUB return: Echo into filter into saturation into sidechain, plus a safety limiter.
Keep the return 100 percent wet, and control intensity with send automation.
Use short throws on phrase endings, fills, and transitions, not constantly.
If you automate feedback, do it briefly, snap it back down, and keep it filtered and ducked.
In heavy DnB, restraint equals impact. Your delay should support the groove, not blur it.

If you want to level up after this, set yourself a throw budget: six throws total across 32 bars, two sources only, and never overlap throws. That one constraint will instantly make your automation feel pro.