Welcome back. This is an advanced sound design lesson in Ableton Live, and we’re going to do something very specific: build authentic 90s sampler tone from scratch using resampling only as the tone generator.

No sampler emulation plugins. No “retro” buttons. The whole vibe comes from a deliberate, repeatable process: control the bandwidth, add conversion-style crunch, print it, then print it again. That’s the key. The sound of jungle and early drum and bass was shaped by limited bandwidth, crunchy conversion, pitch artifacts, noise floors, and the simple fact that audio got recorded and re-recorded over and over.

By the end, you’ll have a resample setup you can reuse in every project, and you’ll know how to apply it to breaks, Reese bass, and atmosphere. Let’s build it.

First, project setup, because if your session is chaotic, your resampling results will be chaotic in a bad way.

Go to Preferences, Audio, and set your sample rate to 44.1 kilohertz. It’s period-correct, and it helps keep your expectations aligned with the kind of bandwidth those records often lived in.

Next: warp discipline. This matters a lot for the “sampler playback” illusion. For drums, you can use Beats warp when you need it for editing, but once you print audio, you’ll often want Warp turned off on the printed clip. For tonal stuff like bass, avoid Complex Pro as your default. If you pitch something, you want it to behave like playback speed, not like modern time-stretch.

Set your tempo somewhere in the drum and bass zone, say 170 to 175 BPM.

Now let’s build the routing. We’re going to make a dedicated resampling track, like a fake sampler input.

Create a new audio track and rename it RESAMPLE BUS. Set Audio From to Resampling. Set Monitor to Off. That “Monitor Off” step is important; it prevents feedback loops and keeps your routing safe. Arm this track when you want to record a print.

Then create another audio track called PRINTED. This is where you’ll drag the recorded clips so you can build a clean little library of generations. Think of PRINTED as your crate: Gen 1, Gen 2, Gen 3, pitched versions, ugly mistakes you’ll use later. Keep it organized.

Now we need the core “90s sampler tone” chain. We’ll do it with stock devices. And here’s a teacher note: you can put this chain on the Master temporarily while printing, which is the strictest way to force commitment. But it’s also easy to accidentally dirty your whole mix. So a safer approach is to route only what you want to crunch into a dedicated group.

Let’s do the safer template method.

Create a group track and name it SAMPLER PRINT GROUP. Any track you want to resample gets routed into this group. Then put the degrade chain on the group itself. That way, your pads, your reference track, your clean sub, whatever else in the session, stays clean.

Here’s the chain order, and I want you to keep it in this order because it’s basically simulating a controlled signal hitting a converter, getting abused, and coming back.

First, EQ Eight. We’re band-limiting before distortion, because that’s how you get bite instead of fizzy pain.

Put a high-pass filter, 24 dB per octave, around 30 to 40 Hz. This is partly about sub control. Deep sub plus bit reduction equals mush and random clipping in the worst place.

Then a low-pass filter, 12 dB per octave, around 12 to 14 kHz. This is the “older top end” move. You can sweep it while listening. The moment the hats stop sounding modern and start sounding like they’re living inside a box, you’re in the zone.

Optional: a small bell dip, maybe minus 2 dB around 300 to 500 Hz if things get boxy after multiple passes. Don’t over-EQ. You’re not mixing yet; you’re shaping an input scene for resampling.

Next device: Redux. This is your bit depth and downsampling crunch.

Start with bit reduction at 12 bits. That’s a sweet spot. If you want nastier jungle, try 10 bits, but don’t start there. For downsample, start around 2.0. You can go up toward 4.0 for more aliasing, but again, the whole point is multiple moderate passes.

Teacher note: if you crush it in one pass, it can sound cool, but it tends to sound like a modern “bitcrusher effect.” The 90s thing is often more like… death by a thousand tiny conversions. Subtle, repeated damage.

Next: Saturator.

Turn Soft Clip on. Drive somewhere between 2 and 6 dB. And trim the output so the level is comparable before and after. Gain matching is everything here, because louder always sounds better, even when it’s worse.

Optional move: turn on Color and center it somewhere around 1 to 3 kHz. That’s where a lot of that crunchy presence lives. Again, don’t go crazy; you can always do a more aggressive pass later.

Optional device: Dynamic Tube. This is for hardware-ish bite, but it can smear transients quickly. Try drive around 10 to 25 percent, bias slightly positive for edge. If your snares suddenly lose snap, back it off or remove it.

Then Utility for gain staging and mono management.

Set your gain so your printed peaks land roughly around minus 10 to minus 6 dBFS. You’re not trying to win loudness; you’re trying to capture character without uncontrolled clipping.

Optional: Bass Mono below about 120 Hz. For drum and bass stability, that can keep your low end tight. Especially useful when you start resampling bass layers.

Limiter is last, only if you need it to catch spikes. Ceiling at minus 0.5. And don’t smash it. If you hear obvious pumping, you’re turning a texture print into a modern loud master. Wrong goal.

Alright. Chain built. Now the real magic: printing in generations.

Here’s the generation workflow. You’re going to resample the same material multiple times, and each time you’ll adjust settings slightly more aggressive, and you’ll commit.

Let’s do it with a break loop first, because it’s the fastest way to hear the character.

Pick a clean break. Make a tight loop first. Trim start points, remove dead air, get the groove right. If you want, you can do very gentle Drum Buss before printing, like a touch of drive, very low crunch, and usually keep Boom at zero. The goal is just to present a controlled input, not to fully process.

Now route that break track into your SAMPLER PRINT GROUP.

Arm RESAMPLE BUS. Record 4 to 16 bars, whatever you need.

After recording, drag that recorded clip onto your PRINTED track.

Now, crucial step: on that printed clip, right-click and turn Warp off unless you specifically want time-stretch artifacts. Warp off means that when you transpose later, it will behave like sampler playback speed, which is basically the entire jungle pitch aesthetic.

Now you print the printed.

Solo the PRINTED clip, or route it through the same group, and record again onto RESAMPLE BUS. That’s generation two.

And you can repeat for generation three, but here’s the discipline: stop when the transients dull too much. Drum and bass still needs impact. The vibe is grit and glue, not cardboard drums.

Here’s a suggested schedule that works almost every time:

Generation one: 12-bit, downsample 2.0, saturator drive maybe 2 to 3 dB.

Generation two: 11-bit, downsample 3.0, saturator 3 to 5 dB.

Generation three, special FX pass: 10-bit, downsample 4.0 or higher, maybe push the low-pass down further. This is for accents, fills, breakdown moments, or layering quietly underneath.

Now, coach tip: do a calibration pass before you print anything important.

Make a one-bar test clip that includes a clicky rimshot and a sine tone, like an Operator sine at 100 Hz. Run it through Gen 1 and Gen 2. Save those calibration prints in your template. Why? Because it teaches your ears exactly what each pass does to transient edge and low-end stability. It’s like setting exposure before you shoot photos.

And another important coaching idea: try to print with a consistent crest factor, not just peak level.

If one loop hits the chain super spiky and another is already compressed, your generation comparisons will be misleading. A trick is to put a Glue Compressor very lightly before the degrade chain, just 1 to 2 dB of gain reduction, only to normalize transient variance. Then, once you’ve printed, you can remove it. You’re not “compressing for sound,” you’re compressing for consistency of printing.

Now let’s do the classic jungle move: pitching like a sampler.

Take your printed break clip. Make sure Warp is off. Now transpose it.

Pitch down by minus 2 to minus 7 semitones for darker, heavier. Pitch up by plus 2 to plus 5 for rave energy.

When Warp is off, transpose changes pitch and speed together. That’s the vibe. That’s the “this came out of hardware” feeling.

Once you’ve pitched it, resample that pitched version again to lock it in. Now it’s committed audio, not a live playback trick.

Pro move: pitch down for weight, resample, then high-pass slightly to make room for a separate clean hat loop. That’s one of the simplest ways to get huge jungle weight without losing modern clarity.

Now, shaping your printed break after generation two.

Do a quick DnB-focused EQ on the printed audio. High-pass around 40 to 60 Hz to leave space for your sub. If you want more body, a gentle bump around 180 to 220 Hz can work. If it’s ripping your head off, notch 3 to 5 kHz by 2 to 4 dB. That range can get brutal after bit reduction and saturation.

Groove and swing: add it after printing. Use the Groove Pool. MPC-ish grooves, shuffled 16ths, whatever fits. And if you want it really authentic, commit the groove by resampling once more. That “baked swing” feel is very real.

Now let’s talk noise floor, because this is one of the biggest differences between “a bitcrushed loop” and “a believable sampler identity.”

Create a noise track. Use Operator noise, Analog noise, or even recorded room tone. Filter it: low-pass around 8 to 10 kHz, high-pass around 200 Hz. Keep it quiet, like roughly minus 30 to minus 24 dBFS under the drums.

But here’s the key: don’t just overlay noise later. Print it into the audio during one of the generations. Route the noise into the same SAMPLER PRINT GROUP along with your break, and print them together. The noise becomes part of the sample’s glue.

Advanced option: make the noise program-dependent instead of constant. Put a Gate on the noise track with sidechain input from the break. Use gentle settings so the noise blooms with hits. Then print that blend. Now the noise feels attached to the signal like a real system.

Alright, Reese bass. We’re not doing deep synth theory here. Just enough to generate source material, then we’ll let resampling create the tone.

Make a Reese using Operator or Wavetable. Two saws, detuned. Add subtle movement, like a slow LFO to filter cutoff or phase. Record or program long notes, like 2 to 8 bars, and also do some one-shot stabs.

Route it into the SAMPLER PRINT GROUP and print Gen 1, then Gen 2.

Now, post-print shaping on the printed bass track.

Use Auto Filter, low-pass 24 dB. Cutoff somewhere from 200 to 800 Hz depending on whether this is a mid bass layer or more full-range. A bit of filter drive can add weight.

Add gentle saturator if needed, but now you’re seasoning, not destroying.

EQ Eight: carve a little space around 50 to 80 Hz if it’s fighting the kick.

Utility: mono below 120 Hz if your sub content is in this layer. Often, you’ll keep your sub as a separate clean sine, and let the resampled Reese handle the mids.

Now a really powerful advanced technique for Reese: the harmonics reprint layer.

Duplicate your Reese audio. High-pass it aggressively, like 250 to 500 Hz. Push saturation and Redux harder than you’d ever do on the full bass. Resample that. Then low-pass it to tuck it in. Blend it under your cleaner mid layer. This gives you audible grit and attitude without turning your low end into fizz.

Now, let’s push into a few “this is how records were really made” habits.

One is the truncate feel. Old editing workflows often had slightly abrupt tails. Do it intentionally: tight fades, like 1 to 5 milliseconds on drums. Slightly too-short snare tails. Then resample. It instantly reads as chopped-from-hardware instead of modern pristine editing.

Another is micro-room imprint. Put a tiny room reverb before one generation and print it, then remove the reverb afterward. Use something like decay 0.3 to 0.6 seconds, small size, high-pass inside the reverb around 300 Hz, low-pass around 8 to 10 kHz. When you print that, the room becomes part of the sample identity instead of an effect sitting on top.

Now let’s do an advanced variation: clock-jitter illusion using only resampling.

Before printing a generation, add tiny, slow pitch drift. You can automate clip transpose slightly, or put the source in Simpler and modulate transposition by about plus or minus 5 to 15 cents over one or two bars. Then resample. After that, you can remove the modulation. The printed audio keeps that slightly unstable, seasick stability that reads as hardware.

Another advanced idea: different box each pass.

Generation one, emphasize filtering and mild saturation. Generation two, back off the filtering and increase downsample. Generation three, push midrange presence with Saturator Color but reduce bit reduction a little. The point is, it feels like you changed samplers or outputs each time, even though it’s all Ableton.

And here’s a big one for modern DnB: targeted aliasing, not global.

If Redux is wrecking your low end, split the audio into bands. Duplicate the printed clip to two tracks. On the high band track, high-pass around 300 to 600 Hz and go heavier on Redux. On the low band track, low-pass around 300 to 600 Hz and keep Redux minimal. Then resample the sum. Now the grind rides on top while the fundamentals stay stable.

If your breaks lose snap after Gen 2 or Gen 3, use the transient-preserve reprint.

Layer a very quiet clean transient-only track. High-pass it, maybe gate it, so it’s mostly attack. Blend it with your dirty break. Then resample the blend gently. The result feels like one committed sample, but you’ve re-injected attack in a period-credible way.

Now let’s turn this into arrangement power, because generation printing isn’t just sound design. It’s an energy map.

Try a simple 64-bar plan: intro is Gen 1, readable and not too chewed. Pre-drop, bring in a quiet Gen 2 ghost layer for anticipation. Drop A, Gen 2 is main with occasional Gen 3 accents on fills. Breakdown, band-limited Gen 3 plus printed noise bed, so it feels taped. Drop B, back to Gen 2, but every 8 bars introduce a new printed variation: a pitch-print fill, a bandwidth shift, or a new generation blend.

And those pitch-print fills? They scream jungle. Duplicate the break, warp off, pitch it down a few semitones for half a bar, resample that moment, and use it as a transition. You’ll be shocked how “record-like” it feels with almost no effort.

Before we wrap, a note about mistakes.

When a print goes too far and the hats crumble or the kick loses chest, don’t delete it. Label it. Amen_G3_too_sand. Think_G2_pitchminus3_ok. Those “wrong” renders are gold for ghost layers, fills, intros, and breakdown textures.

Now a quick 20 to 30 minute practice routine so this becomes muscle memory.

Pick a break and make a 2-bar loop. Print Gen 1 and Gen 2. Arrange 16 bars: first four bars Gen 1 only. Bars five to eight, Gen 1 plus a low-level Gen 2. Bars nine to twelve, Gen 2 dominant plus a pitched-down fill, warp off, maybe minus 3 semitones. Bars thirteen to sixteen, do a Gen 3 telephone moment: low-pass around 8 kHz, more downsample, then slam back to Gen 2 on bar sixteen.

Add a Reese. Print Gen 1 and Gen 2. Use Gen 2 for the mids, keep sub cleaner.

Then do the final test: turn your degrade chain off. Your printed audio should still be gritty and vibey. If turning the chain off makes everything fall apart, that means you weren’t committing; you were just listening through an effect.

Let’s recap the philosophy.

You built a 90s sampler tone using only stock devices and resampling. The real character comes from moderate degradation over multiple generations: band-limit, bit and downsample, saturate, print, repeat.

Warp off and transpose gives you true pitch-and-speed behavior, which is essential for jungle artifacts.

Noise should be printed into the audio for glue, not sprinkled on top at the end.

And for modern impact with vintage character, think parallel: a clean transient layer for snap, and a dirty resampled layer for body and attitude, then resample the sum so it becomes one committed “sample.”

If you tell me what subgenre you’re aiming for, deep rollers, techstep, ragga jungle, and what break you’re starting from, I can suggest a generation schedule with exact settings and a 32-bar arrangement plan that matches the typical brightness and density of that style.