Every producer I know has a folder of vocal chops. And most of them still bounce each chop out of a sampler, one by one, before it ever touches a synth. That workflow made sense five years ago. It doesn’t anymore.
Serum 2 puts a Sample engine, a Multisample engine, granular, and spectral resynthesis inside every oscillator. So serum 2 vocal chops can live inside the synth now: pitched, filtered, modulated, and resampled without leaving the plugin once.
We at Monosounds have been building vocal patches in Serum 2 since it shipped in March 2025. Here are the five methods I actually use, with the settings I start from every time.
Serum 2 Vocal Chops Start in the Sample Oscillator
This is the fastest route. Take a short, dry vocal phrase – one or two bars, no reverb baked in. Switch OSC A’s engine from Wavetable to Sample, then drag your .wav onto the oscillator. Order matters here: if you drop the file while the oscillator is still in Wavetable mode, Serum converts it into a wavetable instead. Sometimes you want that. Not yet.
Two settings matter right away. First, keytracking. Turn it off and the vocal plays at its original pitch on every key, which is what you want for one-finger chop stabs. Leave it on and the sample repitches chromatically, the classic playable-chop sound. Second, the sample start point. Slide it forward to the beginning of a vowel and you have a brand new chop without cutting anything.
Now make it move. Put a fast-attack envelope on the amp with a decay around 200-400 ms and every key press becomes a tight stab. Then map an LFO or a macro to the sample start position, and each note grabs a different slice of the phrase. That one modulation route replaces about half of what I used to do in a sampler.
Map Chops Across Keys with Multisample
One sample repitched across the whole keyboard falls apart fast. Past four or five semitones the vowels smear, the consonants stretch, and you’re deep in chipmunk territory. Most serum 2 vocal chops that sound cheap sound cheap for exactly this reason.
The Multisample engine fixes it. It loads SFZ files, which means you can map a different chop to each key zone. My workflow: cut 8-12 chops from a single vocal take, keep them in one folder, and map each file to its own key. Now C3 is an “oh”, D3 is a “hey”, E3 is a reversed breath. You play the chops like drums, and every zone stays in its natural pitch range.
SFZ sounds intimidating but it’s a plain text file – a minimal mapping is a few lines per sample, and there are free tools that generate one from a folder. It’s the same logic behind the playable patches in our GO2 beatmaking pack: a mapped patch survives every new project, while a bounced chop folder gets lost in three months.
Formant Shifting Without the Chipmunk Effect
Repitching a vocal moves its formants along with the pitch. Up sounds tiny, down sounds like a monster. Sometimes that’s the whole point – half of future bass runs on chops pitched up three or four semitones, and if that’s your lane, my future bass in Serum 2 tutorial covers the chord side of that sound.
But when you want pitch movement without the cartoon voice, Serum 2 gives you two good options. Option one: convert the vocal to a wavetable after all. Drop the .wav onto a Wavetable-mode oscillator, then use the Formant warp mode. Warp shifts the vowel character while the played note controls pitch independently. It sounds synthetic rather than natural, but that robotic vowel tone is a genre staple for a reason.
Option two is the Granular engine, which cuts the link between pitch and time completely. That one deserves its own section.
Granular Vocal Textures
Set the oscillator engine to Granular and load the same vocal. Instead of playing the file straight through, Serum 2 fires off tiny overlapping grains of it – up to 256 at once. Pitch and playback position become independent, so you can hold a chop at any pitch, for as long as you want, without stretch artifacts.
My starting point for a vocal texture: grain size around 80-150 ms, enough density that grains overlap, and a slow LFO drifting the grain position through the vowel section of the phrase. Add a small amount of random spread to the position and a static one-shot turns into a breathing pad. Pitch a second granular oscillator up an octave at low level and it reads as ghostly harmony.
Stacked vocal layers like this are the kind of source the granular engine handles well.
Watch the CPU though. Dense grains plus big unison gets heavy fast. I keep unison at 1-3 on granular oscillators, drop oversampling to 2x while sketching, and freeze the track once the part is locked. If you want the engine explained knob by knob, I wrote a full Serum 2 granular tutorial that goes much deeper than this section can.
Spectral Ghost Vocals
The Spectral engine resynthesizes your vocal into partials: individual sine components you can edit one by one, mute, or morph between frames. This is where the ghost-vocal sound lives. Resynthesize a sung phrase, thin out the partials, slow the frame movement way down, and you get something that is clearly a voice and clearly not human at the same time.
My usual move: strip the lowest partials, keep the region between roughly 300 Hz and 3 kHz where the vowel information sits, and morph between two frames over four bars. It works absurdly well under sparse drums. I break the whole engine down in the Serum 2 spectral oscillator guide if you want the full picture.
Which Engine for Which Job
Here’s how I decide, patch by patch:
| Engine | Best for | Pitch behavior | CPU cost |
|---|---|---|---|
| Sample | Clean one-shot chops, stabs | Repitches whole sample | Low |
| Multisample (SFZ) | Many chops mapped to keys | One chop per key zone | Low-medium |
| Wavetable + Formant warp | Robotic vowel leads | Pitch and formant separate | Low |
| Granular | Pads, stretched textures | Pitch and time decoupled | Medium-high |
| Spectral | Ghost vocals, morphing vowels | Partial-level control | Medium-high |
No Vocal Sample? Fake One with Formant Filters
You can get a convincing vocal-ish synth with no recording at all. Vowels are just resonant peaks called formants, and if you recreate the peaks, the ear hears a voice.
The recipe: one saw oscillator with light unison – unison 3, detune around 0.08-0.12 – into two band-pass filters running in parallel. Serum 2’s dual filter routing does this natively, no external EQ needed. Tune the first filter somewhere between 300 and 700 Hz and the second between 900 and 2300 Hz, with enough resonance that each one almost whistles. Then map a single macro to move both cutoffs between vowel positions, so turning the knob sweeps from “oo” to “ah” to “ee”.
Two finishing touches sell it. A vibrato LFO on pitch at 5-6 Hz with tiny depth and a delayed onset, because real singers don’t start vibrato instantly. And a Bode frequency shifter on its own FX bus, shifted just a few Hz, for that slightly detuned unnatural-voice quality. If you’d rather reverse-engineer than build, our 250+ free Serum 2 presets include vocal-style keys, atmospheres, and chord patches you can open up and study.
Serum 2 Vocal Chop Questions
How do I install vocal chop presets in Serum 2?
If the pack comes as a .SerumPack file, open Serum 2’s Menu tab and use Import Preset Pack. Otherwise copy the presets into your Serum 2 presets folder – subfolders show up in the browser, so keep packs organized. Serum 1 presets load fine too. Our Serum 2 preset guide covers which packs are worth grabbing first.
Can Serum 2 time-stretch a vocal to my project BPM?
Sort of. The Sample engine links speed and pitch, like an old hardware sampler. The Granular engine decouples them, which handles chops under a bar or two beautifully. For stretching a full vocal to a new tempo, I still do it in the DAW first and load the result – it’s cleaner and cheaper on CPU.
Why do my serum 2 vocal chops sound thin?
Usually it’s the source, not the synth. A chop recorded with room reverb or heavy processing fights every engine. Start dry, high-pass around 100-150 Hz, and put Serum 2’s reverb on its own FX bus so the tail doesn’t retrigger with every note. Layering a quiet saw an octave below the chop also fills things out fast.
Is Serum 2 a paid upgrade for these vocal features?
No. Serum 2 is a free upgrade for Serum 1 owners with lifetime free updates, currently on version 2.1.x. All five oscillator engines are included, nothing is paywalled. Details are on Xfer’s product page.
Start with a Phrase You Already Have
Grab any vocal sitting in your project folder, even a rough take. Sample engine first, granular when it needs movement, spectral when it needs to get strange. Those three cover 90% of serum 2 vocal chops in real tracks; the rest is taste.
And if you want a shortcut, grab our free 250+ Serum 2 presets – royalty-free, no strings attached. Pull the vocal-ish patches apart and steal how they’re wired.
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 →
