Everyone downloads wavetables. Almost nobody builds their own. That’s a shame, because Serum 2 custom wavetables are the cheapest way to get a sound nobody else has, and the process takes minutes, not hours.
I’ve been making Serum preset packs at Monosounds since 2022, and here’s an honest number: roughly 4 out of 10 patches in our commercial packs run on tables we made ourselves. Stock tables are fine. But when a bass cuts through a mix in a way you can’t quite place, a custom table is usually the reason.
This guide covers the four ways I actually build tables in Serum 2: dragging a .wav onto an oscillator, drawing in the wavetable editor, the formula generator, and the Resample to Oscillator trick. Plus where the files live on disk, and how to keep a personal wavetable bank from turning into a junk drawer.
Why Bother With Serum 2 Custom Wavetables
Three reasons. First, identity. Every producer with 30 dollars has the same stock tables you do. A table made from your own voice or your own guitar exists nowhere else.
Second, understanding. Once you’ve sliced a vocal into frames and watched what pitch detection does to it, presets stop being magic. You start reading patches instead of just playing them.
Third, the Wavetable engine is only one of five oscillator engines in Serum 2, but it’s still the one you’ll use most. I broke down when to reach for each engine in my guide to Serum 2 oscillator types. Custom tables make that main engine yours.
The Fastest Way: Drag a .wav Onto an Oscillator
The fastest route to Serum 2 custom wavetables is plain drag and drop, and it’s exactly as simple as it sounds. Take any .wav file and drop it straight onto the oscillator’s waveform display. Serum 2 analyzes the audio and slices it into wavetable frames on the spot.
Source material matters more than settings here. My rules after four years of doing this:
- Use a sustained, monophonic sound. One vocal note, one guitar string, one synth stab. 1 to 3 seconds is plenty.
- No reverb, no delay on the source. Tails smear the frames.
- Louder is better. Quiet sources produce weak, noisy tables. Normalize first.
When you drop the file, Serum asks how to slice it. For anything with a clear pitch, pick the pitch-based analysis so the frame length follows the actual cycle of the sound. You get a smooth, in-tune table. Fixed-size slicing on the same vocal gives you a detuned, glitchy result, which is sometimes exactly what a techno lead needs. Try both. It costs you ten seconds.
Then sweep the WT Pos knob slowly. That knob is now a tour through your source sound. Drop an LFO on it, set the rate to 1/2 at 124 BPM, and you have movement that no stock patch has.
Wavetable Editor Basics
The editor is where Serum 2 custom wavetables get cleaned up after import, and where drawn-from-scratch tables get born. Open it from the oscillator section and you’ll see your table as a stack of frames you can step through.
The tools I actually use, in order of frequency:
- Delete frames. Imported audio almost always has junk at the start (the attack click) and the tail (silence). Cut them. A 40-frame table that’s all sweet spot beats a 200-frame table that’s 80% filler.
- Draw with grid snap on. Freehand drawing sounds like noise. Snapped lines and segments give you harmonically sensible shapes.
- Edit harmonics, not the waveform. The spectral view lets you pull individual harmonics up or down. Kill every even harmonic on a saw frame and you’re most of the way to a square. Boost harmonic 7 alone and you get a hollow, bell-ish color you can’t draw by hand.
- Morph between keyframes. Make frame 1 a sine and frame 64 your harmonically mangled saw, let Serum interpolate the frames between, and the WT Pos knob becomes a timbre fader.
If you find yourself wanting to edit partials across a whole piece of audio rather than frame by frame, that’s a different tool for a different job: the Spectral engine in Serum 2 does resynthesis with partial-level editing and is often the better pick for pads from long samples.
The Formula Generator in One Minute
The formula tab lets you type math and get a waveform. A plain sine is one short expression, and each extra term stacks another harmonic on top. I use it for one thing: perfectly clean digital shapes that drawing can’t match. Odd-harmonic stacks for chiptune leads, mostly. If you’re not comfortable with sin() and x, skip it entirely. You will lose nothing. The drag-and-drop and resample workflows cover 95% of real use.
Resample to Oscillator: Bake FX Into a Table
This is the deepest rabbit hole in Serum 2, and my favorite. The idea: build a patch, run it through effects, then use Resample to Oscillator to render the processed sound back into the oscillator as a brand new wavetable. The FX are now baked into the table itself.
Why that’s a big deal: Serum 2 has multiple FX buses, so you can distort, frequency-shift with the Bode shifter, and reverb a sound, bake all of it into a table, and then run that table through a fresh, empty FX chain. Repeat two or three generations and you get tables with grit and character that no amount of knob-turning produces in one pass.
My standard bass recipe: start with a basic saw, Overdrive distortion mode at around 40% drive, a little Bode shifting for inharmonic edge, resample. Then filter the new table, resample again. Second-generation tables like this power a lot of the low end in our Colors preset pack, and it’s the same trick behind most of the patches in my Serum 2 bass presets roundup.
Bonus: baked FX cost zero CPU at play time. A resampled table with distortion built in plays as cheap as any stock table. On a busy project that matters.
Where Serum 2 Wavetables Live on Disk
Serum 2 custom wavetables are plain .wav files with Serum metadata inside. That means you can copy them, back them up, and share them like any other file. The default locations:
| OS | Default Tables folder |
|---|---|
| Windows | DocumentsXferSerum 2 PresetsTables |
| macOS | ~/Library/Audio/Presets/Xfer Records/Serum 2 Presets/Tables |
Don’t trust defaults blindly, though. Installers and older setups move things around. The reliable way: open Serum 2, go to the Menu tab, and use Show Serum Presets Folder. Whatever opens is your real location. The Tables folder sits inside it. If anything looks off, Xfer’s official Serum 2 page links to their support docs on folder setup.
Two habits that save headaches. After copying files in manually, rescan folders from the menu so Serum sees them. And back the Tables folder up somewhere outside your machine, because your DAW project references tables but does not embed them. Lose the folder, and old projects open with silent oscillators.
Organizing a Personal Wavetable Bank
Serum 2 supports subfolders inside Tables, and this is where most people go wrong: they hoard. Five hundred unnamed tables called “resample_final_2.wav” is not a bank, it’s a landfill. I put table hoarding near the top of my list of Serum 2 beginner mistakes for a reason.
Here’s the structure I’ve used for every Monosounds pack since 2022:
Tables/MS_Bass,Tables/MS_Vox,Tables/MS_Keys— one folder per source type.- Names follow prefix, source, character:
MS_Vox_Ah_Dark.wav. Six months later, that name still tells me everything. - A
_Labfolder for experiments. Tables only graduate out of it after they’ve earned a spot in three real patches. Most never do, and that’s fine.
If you want reference material for how finished tables get used, open a few patches from our free 250+ Serum 2 preset pack and check which table each oscillator loads and how WT Pos is modulated. Reverse-engineering working patches taught me more than any tutorial did. There’s a longer walkthrough of that habit in my ultimate Serum 2 presets guide.
FAQ: Serum 2 Custom Wavetables
Do Serum 1 wavetables work in Serum 2?
Yes. Tables are .wav files and the format carried over. Copy your old Serum 1 tables into the Serum 2 Tables folder, rescan, done. Serum 1 presets load in Serum 2 too, so nothing from your old library is wasted.
What’s the best source length for an imported wavetable?
One to three seconds of a sustained, single-pitch sound. Serum’s frames are 2048 samples each and a table holds up to 256 frames, so long files just get truncated or thinned out. Shorter and cleaner beats longer and busier every time.
Why does my imported table sound out of tune?
The slicing didn’t match the source’s cycle length. Re-import with pitch-based analysis, and make sure the source is monophonic — chords confuse pitch detection. If it’s still a few cents off, fix it with the oscillator’s fine tune and move on.
Do custom wavetables use more CPU than stock ones?
No. A table is a table; Serum doesn’t care who made it. CPU in the Wavetable engine comes from unison count and quality settings, so keep unison at 3 to 7 instead of 16 and you’re fine. The Granular and Spectral engines are the expensive ones, not your custom tables.
Start With One Table
Don’t try to build a bank this weekend. Record one note of your own voice, drag it onto oscillator A, delete the junk frames, name it properly, and use it in a real patch today. One table you actually use beats fifty you forgot.
And if you want a shortcut while you learn, grab our 250+ free Serum 2 presets — no strings attached. Tear the patches apart, swap our tables for yours, and see how much of the sound was the table all along.
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 →
