Welcome back. Today we’re doing crash design for drum and bass, from scratch, inside Ableton Live using Session View and resampling. This is an intermediate workflow, and the goal is simple: stop relying on whatever random crash happens to be in the sample pack, and start building crashes that actually fit your drop, your snare, and your vibe.

In DnB, a crash isn’t just a cymbal. It’s a moment of impact. It tells the listener, “new phrase, new energy,” and if you do it right, it glues the whole drop together without washing out your drums. And the secret weapon here is resampling, because once you print a version, you can go way harder with processing and experimentation, without getting lost or overthinking.

Alright, set up first.

Go to Session View. Set your tempo somewhere realistic, like 174 BPM. Now create three MIDI tracks and name them: CRASH_NOISE, CRASH_METAL, and CRASH_HIT. Then create one audio track called RESAMPLE_PRINT.

Open your I/O section. On RESAMPLE_PRINT, set Audio From to Resampling. Set Monitor to Off. That “Monitor Off” is important because it avoids feedback loops and weird doubling. Then arm RESAMPLE_PRINT, but don’t record yet.

Also, tiny workflow tip: treat resampling like version control. We’re going to keep our CRASH_BUILDER group as the clean source, and every time we print, we label it like v01, v02, v03. That way you’re never doing undo archaeology at 2 AM wondering where the clean version went.

Now we build the three layers: air, metal, and impact.

First layer: noise. This is the wash, the breath, the air moving.

On CRASH_NOISE, drop in Operator. In Operator, we only need Oscillator A. Set Osc A to white noise. Now shape the amp envelope. Give it a tiny attack, like 3 to 10 milliseconds, just so it doesn’t click. Decay can be around 1.8 to 3.5 seconds. Sustain all the way down. Release somewhere like 0.8 to 2.5 seconds. We’re aiming for a tail that feels like a real crash in a fast mix, not a ten-second ambient pad.

After Operator, add Auto Filter. Set it to a high-pass, 12 dB slope. Put the frequency around 250 to 500 Hz. We’re clearing out low junk because crashes don’t need to compete with kicks, subs, and the body of the snare. Add a bit of resonance, like 0.6 to 1.2, just to give it a touch of edge. If you want extra bite, add a small envelope amount so the filter opens slightly at the start.

Optional, but really useful for making it feel wide and alive: Chorus-Ensemble. Slow rate, around 0.15 to 0.35 Hz, amount 20 to 40 percent, mix 15 to 30 percent. Keep it subtle; we can always exaggerate later after printing.

Now make a one-bar MIDI clip on this track. Put a note at C3. The note length can be anywhere from a quarter note to a full bar. Don’t stress it; our envelopes are doing most of the work.

Second layer: metallic tone. This is the “cymbal material,” the shimmer that tells your ear “metal.”

On CRASH_METAL, load Collision. Start from any metallic-ish preset, then tweak. Bring up the noise or exciter so it actually hits, not just rings. Pick a material that’s bright to start, because it’s easier to darken later than to invent highs that aren’t there.

Now add Resonators after Collision. Yes, another resonant layer on top. Turn on three to five resonators. Tune them like a bright cluster rather than musical notes. Try something like 2.5k, 3.2k, 4.1k, 5.4k. Think shimmer bands, not “this is a chord.” Set decay around 0.8 to 1.8 seconds. Dry/wet around 30 to 60 percent.

Here’s a realism tip: if this starts sounding too tonal, like it’s “singing a note,” slightly offset the resonator frequencies by a tiny amount, and make the lowest resonator decay a bit shorter than the top ones. Real cymbals don’t ring evenly. That unevenness is what makes them feel natural.

Now add Saturator. Drive 2 to 6 dB, soft clip on. Then EQ Eight. High-pass around 250 to 400 Hz. If it’s harsh, dip a bit around 7 to 10 kHz, like 1 to 3 dB. We’re going for hype, not dentist drill.

Make a MIDI clip similar to the noise layer. Trigger a note, similar length.

Third layer: transient hit. This is the “tick” or “snap” that makes the crash readable on small speakers and in loud drops. A lot of crashes disappear because they’re all tail and no point.

On CRASH_HIT, you’ve got two options. If you want pure synthesis, load Drum Synth Clap or Drum Synth Snare. Set it short. Decay around 80 to 200 milliseconds. Tone down a bit so it doesn’t turn into a cheesy clap vibe.

Then shape it. If you’ve got a transient shaper, great. If not, fake it with Saturator and EQ. Add EQ Eight: high-pass 200 to 300 Hz, and a small boost around 2 to 5 kHz to make the attack speak.

Alternative option is a micro-sample: a rimshot, click, tiny stick hit, trimmed to like 10 to 50 milliseconds. The key is: it’s not a snare layer, it’s just the “pointer” that tells your ear where the crash starts.

Make a very short MIDI note, like a sixteenth or even shorter.

Cool. Now we glue.

Select all three MIDI tracks and group them. Name the group CRASH_BUILDER.

On the group insert Glue Compressor. Attack around 3 milliseconds, release on Auto, ratio 2:1. You’re not trying to smash it; you just want 1 to 3 dB of gain reduction so the layers move together.

Then EQ Eight on the group. High-pass around 200 to 350 Hz. If you want a little extra sparkle, maybe a gentle high shelf at 10 to 12 kHz, one or two dB, but only if it actually needs it.

Then Utility. Push width to something like 120 to 160 percent. Don’t go insane. Also, if you’ve got Bass Mono, turn it on and set it around 200 Hz. Even though we’re high-passing, this keeps the stereo image stable.

Now trigger the clips together in Session View. You should hear something that already feels like a cohesive crash, not three separate sounds.

Before we print, let’s do a quick gain staging check. This matters a lot for distortion later. Aim for peaks on the group around minus 10 to minus 6 dBFS. If you print super hot near zero, everything you do afterward turns into brittle fizzy pain. Headroom equals better distortion. Always.

Now the fun part: resampling.

Arm RESAMPLE_PRINT. Create an empty clip slot. Hit Session Record in the top transport. Then launch your crash scene. Record one to two bars so you capture the whole tail. Stop.

Now immediately, right-click the recorded clip and choose Crop Sample. Consolidate if needed. Rename it to something like Crash_Print_01, or Crash_Master_v01. Name it like you’ll reuse it later. Because you will.

This is your commit point. From here on out, we can do aggressive processing without fear.

Now let’s shape the print into something mix-ready for DnB.

First, cleanup and tone.

Put EQ Eight on the printed audio track. High-pass around 250 to 450 Hz depending on how busy your drop is. If your mix is thick and rolling, you’ll probably go higher. Then sweep around 6 to 10 kHz and listen for that painful “shhhk” region. If it’s piercing, cut 1 to 4 dB.

Then add Auto Filter. Try a low-pass 12 dB. Set the cutoff around 6 to 9 kHz as a starting point, and use a little envelope amount so it opens at the beginning. That gives you that “whoosh then open” feel, like the crash blooms rather than just blasting static. Add a touch of drive if it helps.

Now controlled aggression.

If you have Roar, you can use it, but Saturator works perfectly. Drive around 3 to 9 dB, soft clip on. Then Drum Buss: drive 5 to 15 percent, crunch 0 to 10 percent. Usually keep Boom off for crashes unless you specifically want a huge cinematic slam, because Boom can add low energy that just gets in the way.

Then a Limiter at the end. This is just spike control. Adjust gain so it taps 1 to 2 dB on the loudest transient. If you’re doing 6 dB of limiting, you’re probably killing the crash and making it small.

Now space, but DnB-safe space.

Add Hybrid Reverb or Reverb. Decay around 1.2 to 2.8 seconds. Predelay 10 to 25 ms so the transient stays clean. High cut around 7 to 10 kHz so the reverb doesn’t add extra hiss on top of your hats. Dry/wet: keep it modest, like 8 to 18 percent. In dense music, less is more.

After the reverb, add another EQ Eight and high-pass around 500 Hz. That’s a big one. Reverb low end turns your mix into fog.

At this point, you have a solid printed crash. Now we iterate like pros.

Print variations. Fast.

Variation one: dark crash. Low-pass the print to around 6 to 8 kHz and add distortion. Print it. Crop. Rename it Crash_Dark_v01.

Variation two: reverse crash. Take the clip and reverse it. Add reverb, maybe slightly bigger than normal. Then resample that. Now you have a reverse swell.

And here’s an advanced move: two-stage reverse. Reverse the crash, add a big reverb like 50 to 80 percent wet, resample it, then reverse that resample again. The result is a swell that rises into the hit but still lands clean, because the end becomes tight again after the second reverse.

Variation three: short impact. Shorten the tail with fades, emphasize the transient layer, or even layer a “tick-first” version. A tick-first trick is duplicating the print, high-passing it at 2 to 4 kHz, trimming it to 80 to 200 ms, distorting it, and blending it quietly under the main crash. That makes the crash readable on phone speakers and in chaotic drops.

Variation four: wide shimmer. Add more Chorus-Ensemble, widen a bit with Utility, and print it.

Now, a crucial stereo check. After you widen, temporarily set Utility width to 0 percent. If your crash collapses into thin papery nonsense in mono, that width was coming from phase smear rather than useful stereo energy. Back off the chorus or unison-style effects, and instead try subtle M/S EQ: keep the mid punchy and mono-safe, and let the sides carry the high shimmer.

If you want to go even more advanced, do mid-only aggression and side-only shimmer. Make an Audio Effect Rack with two chains: MID and SIDE. On the MID chain, add Saturator or Drum Buss and keep it controlled. On the SIDE chain, keep it cleaner, add Chorus-Ensemble and maybe a high shelf. That gives you a crash that punches in the center while sparkling on the edges. Very modern, very mix-friendly.

Quick note on warp strategy for printed audio: if you want a natural cymbal tail, turn Warp off and transpose carefully. If you want a designed synthetic texture, try Complex or Complex Pro, mess with formants a bit, and then print it again. That “print and commit” step is what keeps your workflow moving forward instead of endlessly tweaking.

Now let’s talk placement, because the best crash sound in the world still fails if you put it in the wrong spot.

Classic DnB placements:
Use it on bar 1 of the drop. Obvious, but essential.
Then use a variation every 16 bars to mark phrases.
Use a reverse swell right before the drop, but consider leaving the actual downbeat cleaner.
Try call-and-response with fills: put a short crash at the end of a fill, not at the start, so you don’t stack brightness on the busiest moment.

Here’s a mix clarity trick: after the initial impact, automate the crash down by 1 to 2 dB for the first half bar of the drop. You keep the statement, but you don’t smear that first kick-snare cycle.

Also, don’t underestimate the power of subtracting. Mute hats for an eighth note on the downbeat where the crash hits. Suddenly your crash sounds twice as big, with zero extra processing.

Common mistakes to avoid, quickly:
If there’s too much low-mid between 200 and 800 Hz, it’ll sound like cardboard and it’ll mask your snare.
If you over-widen, the stereo image destabilizes right when the drop hits.
If you build up harshness around 7 to 10 kHz after saturation, you must tame it with EQ.
If tails are too long, they pile up and smear the groove.
And don’t print too hot. Keep that headroom. It’s the difference between “expensive crash” and “fizzy spray.”

Alright, mini practice exercise.
Build one crash using the three-layer method.
Resample it as your master.
Then print three variations: bright, dark, and reverse.
Drop them into a simple 32-bar DnB loop. Put a crash on bar 1 and bar 17, and a reverse crash leading into the next phrase.
Then listen at low volume. Low volume is the truth test. Does the crash still read? And is it stepping on your snare crack?

Recap.
You built a crash from noise, metal, and transient impact.
You used Session View resampling as a fast iteration workflow: generate, process, resample, re-process, commit.
You shaped it with EQ, saturation, controlled reverb, and limiting so it works in dense DnB.
And you created a small palette of variations so your arrangement feels intentional, not copy-pasted.

If you tell me what substyle you’re aiming for, like liquid, jump-up, neuro, or jungle, and what region your snare really speaks in, I can suggest a target tail length and a safe frequency window so your crash hits hard without fighting the drums.