I’ve made growl presets for a living since 2022, and my first fifty attempts were embarrassing. They buzzed. They rasped. But they never talked. If you’re stuck at the angry-sawtooth stage, this is the exact serum 2 dubstep growl recipe I use for our packs at Monosounds.
The trick isn’t one magic knob. It’s a chain: a vowel-ish wavetable, the right warp mode, formant movement, distortion into phaser, and a resampling loop that stacks character in layers.
Serum 2 makes all of this easier than Serum 1 ever did. The new spectral oscillator alone gives you growl material that didn’t exist before March 2025. Let’s build one from scratch.
What Makes a Serum 2 Dubstep Growl Talk
A growl reads as talking because of formants — the resonant peaks that make an “oh” different from an “ee”. Your ears are wired to hear them. Move formants rhythmically and almost any tone turns vocal.
So the whole game is this: generate a harmonically rich sound, then push its formants around in time with your track. Every step below serves that one goal.
One ground rule first. Growls live in the mids, roughly 150 Hz to 2 kHz. The sub is a separate, clean layer — I cover that split in my Serum 2 808 and sub bass guide. Never run your sub through the growl’s distortion.
Start With a Vowel-ish Wavetable
Skip the default saw. Scroll the factory tables and sweep the WT Position knob on each one. You’re hunting for anything that sounds like a mouth opening and closing as you sweep. If it already sounds vocal before any processing, you’ve found your starting point.
Better yet, make your own. Record yourself saying “yah-woh-yoi” into your phone, trim the file, and drag the .wav straight onto oscillator A. Serum 2 converts it into a wavetable on the spot. I wrote a full walkthrough in my custom wavetables guide; the short version is that steady-pitch source material converts cleanest.
My starting settings: unison 1 while designing (mono growls hit harder and save CPU), octave down one, level around 75%. Then route an LFO to WT Position before you touch anything else. Static tables never talk.
Warp Modes: Bend and FM Do the Heavy Lifting
The warp section is where a static tone starts to move. Two modes matter most for growls.
Bend squashes the waveform toward one edge. Around 30-50% it adds a nasal snarl to almost any vowel table. Modulate the warp amount with the same LFO you put on WT Position, but at half the depth. Two movements slightly out of sync is what makes the sound chew instead of wobble.
FM is the aggression dial. Set oscillator B to a plain sine one octave below A, switch A’s warp mode to FM from B, and push the amount. At 15-25% you get a throaty rasp. Past 40% it turns metallic and screamy — great for drops, harsh for anything else.
If you’d rather hear this dialed in than read about it, open a few patches from our Grumble Skrillex-inspired preset pack and check the warp amounts. Almost every patch leans on FM warp somewhere.
Here is one of those patches, so you can hear where the warp settings land.
Formant Filter Movement
Now the filter. Serum 2’s filter menu includes formant types that behave like a vocal tract — cutoff becomes a vowel selector instead of a brightness control.
Load one, then modulate cutoff with a slow LFO or an envelope. Keep the range small. The shift from “oh” to “ah” is a tiny knob move; sweeping the full range sounds like a filter demo, not a voice.
Resonance between 20-40% keeps the vowels defined. Push it higher and the patch starts whistling.
Now stack the motion: WT Position moving, warp moving, formant moving — each from a slightly different modulation source. That’s three layers of vowel movement, and a good serum 2 dubstep growl usually talks at this stage, before any distortion touches it.
Distortion Into Phaser: The Classic Chain
Order matters more than exact settings here. Distortion first creates new harmonics; a phaser after it carves moving notches into those harmonics. Flip the order and you just get fizz.
Here’s my default chain in the FX tab:
| Slot | Effect | Settings that matter |
|---|---|---|
| 1 | Distortion | Overdrive mode, drive 40-60% |
| 2 | Phaser | Rate 1/2 sync, feedback 60%, mix 50% |
| 3 | EQ | High-pass at 120 Hz, dip 3 dB near 400 Hz |
| 4 | Compressor | Multiband, tame the 2-5 kHz band |
| 5 | Hyper/Dimension | Mix 15-25% for stereo width |
Serum 2 runs multiple FX buses now, so I route the growl oscillators to one bus and keep anything clean away from it. You can also modulate individual effect parameters, which Serum 1 couldn’t do — an LFO on phaser feedback adds a second layer of talk for free. I break the whole FX section down in my Serum 2 effects guide.
LFO Shapes for Talking Rhythm
Straight sine LFOs make sirens, not syllables. Draw your own shapes.
My go-to: an LFO with three or four steps at different heights, curved falls between them, synced to 1/2. At 140-150 BPM half-time — where most dubstep lives — 1/2-note movement sits right in the pocket.
Then a second, faster LFO at 1/8 with a small depth on the formant filter. That adds syllables inside each word. Switch it to a triplet grid when you want the classic “yoi yoi” bounce.
Serum 2’s drag-and-drop modulation makes this quick: grab the LFO header, drop it on WT Position, warp, and formant cutoff, then set a different depth for each.
The Resampling Loop
This is the step most tutorials skip, and it’s where the dense, layered pro sound actually comes from. One pass through the synth is never enough. Every serum 2 dubstep growl I ship goes through this loop at least twice.
Render your growl to audio: one sustained note, two to four bars, full FX chain printed. Then bring it back in. Serum 2 gives you two doors: drag the .wav onto an oscillator to slice it into a fresh wavetable, or use Resample to Oscillator and Serum bakes the whole processed sound into a new table without leaving the plugin.
Now your distorted, phased growl is raw material again. Sweeping the new table’s position scrubs through your own processed audio. Add a new warp mode, a new formant filter, new FX. Two or three passes deep, the sound picks up a density you can’t get in a single pass.
One rule: change the modulation rhythm every pass. Use the same LFO shape twice in a row and the layers blur into mush.
The Spectral Oscillator: A Growl Source Serum 1 Never Had
Serum 2’s spectral engine resynthesizes any sample into individual partials you can edit one by one, plus frames you can morph between. For growls, it’s a cheat code.
Feed it a short vocal phrase or one of your resampled growls. Then get destructive: mute clusters of upper partials for a hollow, throaty tone, or shift partials against each other for inharmonic grit. Morphing between frames with an LFO gives you vowel movement without touching a formant filter.
Watch the CPU, though. Spectral costs more than wavetable mode. Keep unison at 3-5 instead of 16, cap your polyphony, and stay at 2x oversampling while designing. Xfer publishes official CPU guidelines on their support page if a patch still chokes your session.
Several bass patches in our free 250+ Serum 2 preset pack use the spectral engine as the mid-growl layer. Open them and check which partials are muted — reverse-engineering finished patches teaches faster than any article.
Serum 2 Growl Questions I Get Asked
Why does my serum 2 dubstep growl sound thin?
Usually it’s a missing sub layer, not a broken growl. Keep the growl high-passed around 120-150 Hz and run a clean sine or 808 underneath. If the mids themselves feel weak, you probably need one more resample pass. Comparing against finished patches helps too — my best Serum 2 bass presets roundup has plenty of reference material.
What LFO rate should I use for dubstep growls?
Start at 1/2 sync for the main movement at 140-150 BPM, 1/8 for inner syllables, and 1/4 triplets for bounce. Anything faster than 1/16 stops sounding like speech and starts sounding like texture.
Do my old Serum 1 growl presets still work?
Yes. Serum 2 loads Serum 1 presets directly, and the upgrade was free for Serum 1 owners. Old growls also make great resampling fodder — render them and rebuild the mids with the spectral engine.
Do I need paid presets to make good growls?
No. Everything in this article works with factory content. Presets are a shortcut for learning, not a requirement — start with our free Serum 2 bass presets, look at how the mod matrix is wired, then build your own from scratch.
Go Make Something Ugly
That’s the whole recipe: vowel table, Bend and FM warp, formant motion, distortion into phaser, a hand-drawn LFO, then resample and run it again. Your first serum 2 dubstep growl from this process will be rough. Your tenth will talk.
If you want working patches to tear apart, grab our free pack with 250+ Serum 2 presets — bass, growls, leads, keys, the lot. Royalty-free, no strings attached. Break them and see how they’re built.
I run Monosounds.studio and make the packs myself. Ten-plus years of producing, 97+ releases, 16,000+ downloads. Every Serum 2 preset we ship passes a CPU check on a mid-range laptop and gets used in a real track before release. If it does not survive a mix, it does not go in the pack.
More about Maxim & the studio →
