HOW TO: Master Class: Xfer Serum

Conquer the feature set of this powerful softsynth
Image placeholder title

In just over two years, Xfer’s Serum has gone from clever indie softsynth to desert island must-have, thanks to its extremely intuitive design, massive array of synthesis features, and creator Steve Duda’s commitment to insanely high audio quality. Originally designed as the ultimate wavetable synth with world-class filters, Serum’s ever-expanding feature set has grown to include sample playback, image import, audio processing, and eight programmable LFOs that can serve double-duty as both multi-stage envelopes and step-sequencers.

Between Serum’s features, ubiquitous user base, and the fact that you can now rent-to-own the software via the network, it’s high time we did a Master Class on making the most of this wündersynth.


Figs. 1-5, Top to bottom: Create a PPG-style sound in Serum’s wavetable editor. Synth historians know that Wolfgang Palm’s PPG Wave 2 started the wavetable phenomenon, so as an homage to Palm’s original innovation, here’s a quick way to recreate the general character of the PPG sound. The overall approach is to create a series of additive waves that become progressively more complex as the wavetable evolves.

Image placeholder title
Image placeholder title
Image placeholder title
Image placeholder title
Image placeholder title

1. Open the wavetable editor (in an initialized patch, sawtooth will be the default) and raise the volume of two lower-numbered harmonics. Here I selected the first and fourth bins (harmonics). See Figure 1.
2. You’ll immediately see the results in the wave view and in the tiny lower window for the first index. Click on the plus sign to edit the second index (Figure 2). Note that this copies the first index to that slot for further manipulation.
3. In the second index, add a few additional lower-range harmonics. Here, I added bins 6 and 9, for some organ/bell character (see Figure 3).
4. Repeat this process a few more times, adding 2-3 additional harmonics for each successive index (figure 4).
5. By the time you reach 6-8 indices, you’re ready to create your wavetable. The “morph” pull-down in the wavetable editor offers four modes, each subtly different in their technical properties (Figure 5). This type of wavetable works beautifully in the “crossfade” mode. Select that and the result will be a classic PPG-style wavetable.



Resampling your favorite digital synths for conversion to Serum wavetables creates a rich source of new material—especially when combined with Serum’s extensive modulation tools and effects. For example, creating a complex patch with an elaborate decay sweep in an FM synth is a great way to generate rich, animated tones for further processing. However, importing audio as a source for wavetables can be a tricky process with unpredictable results, unless you follow a few rules of thumb.

Unison, detuning, vibrato, and obvious noise in the source material can result in bizarre artifacts that don’t represent the original sound. When sampling another synth for use in Serum, keep your sound to a single oscillator (or carrier, when sampling an FM synth), then add detuning and chorusing later via Serum’s features.

Lower-frequency notes give Serum a bit more “headroom” for calculating the wavetable. Steve Duda recommends G#-1 (26 Hz) as an ideal note to sample, when importing your own source audio.

Fig. 6. G#-1, expressed in the formula editor. The Formula editor in Serum’s wavetable panel can be used to create unique waveshapes if math is your specialty, but it also serves as a way to give Serum’s import tools a “heads up” for indexing the audio. If your sample is G#-1, just type “G#-1” (no quotes) into the formula editor and it will split the indexes in neat little slices of 1,699 samples each (see Figure 6).

Image placeholder title

As a hands-on example, find a cool unprocessed (no effects or detuning) FM patch in Ableton Operator or Native Instruments’ FM8, then record a single G#-1 note. Type G#-1 into the formula editor and drag/drop the sample into the window. Your FM patch will then be available as a wavetable within Serum, which can also be saved to your user wavetable library for later use.

If you want to finesse the results further, try experimenting with Serum’s various morph modes.


Speaking of FM, Serum’s wave-warping modes make up a treasure chest of tools for mathematically mangling your wavetables. With these you can add “pulse-width modulation” to any waveform (along with square waves, of course), create hard sync effects, or flip/mirror your waves for more exotic results. Best of all, you can actually see the results of these processes reflected in real-time in the waveform displays.

Tucked away near the bottom of the menu is an FM option that allows you to use either oscillator as an FM source, depending on the oscillator you select. For example, the warp menu for oscillator A offers FM from oscillator B—and vice versa.


Thanks to the additive tools in Serum’s wavetable editor, DX-style two-operator patches are easy to create. Just remember that coarse tuning values in traditional FM synths correspond to the numbers of the harmonics. That is, if the carrier and modulator have a 1:3 ratio, the carrier is tuned to the first harmonic and the modulator is tuned to the third harmonic. You can accomplish exactly the same thing via Serum’s harmonic editors.

Fig. 7. Initialized sine wave. Start with an initialized patch and use oscillator 1’s wavetable editor to create a single sine wave at the first harmonic by raising its value to maximum. This will be reflected in both the waveshape and the first index of the wavetable (Figure 7).

Image placeholder title

Fig. 8.Fig. 9.Figs. 10-11. Re-creating our 1:3 ratio. To speed up the process, select “Copy Osc A -> B” from the main menu pulldown. This will duplicate oscillator A’s sine wave to oscillator B (Figure 8).

Image placeholder title
Image placeholder title
Image placeholder title

Next, in oscillator A’s warp pull-down, select “FM (from B)” as shown in Figure 9. Now, oscillator A will be the carrier and oscillator B will be the modulator.

As mentioned above, classic FM synths relied on sine wave modulation, with specific harmonic tunings creating different FM overtone series. So, changing the “coarse tuning” of the carrier (osc A) or the modulator (osc B) is a simple matter of adjusting the volume of its harmonics. To recreate the 1:3 ratio described above, open oscillator B’s harmonic editor and raise the value of the third harmonic while lowering the value of the first harmonic. If your FM amount (warp) value is set to anything other than zero, you’ll instantly hear the tonal change (See figures 10 and 11).

PRO TIP: The Serumclavier

Now that you’ve got a handle on creating additive wavetables and FM synthesis within Serum, it’s time to dive deeper. EM subscribers may recall my October Synclavier Master Class (available at One of the hallmarks of the Synclavier sound is its ability to use time-slicing to shift the harmonic values of both the carrier and modulator.

By combining Serum’s FM tools with its approach to smoothly morphing wavetables, you can recreate the Synclavier sound by using an envelope to modulate the index position of a simple additive wavetable on oscillator B, while applying a second envelope to the FM amount on oscillator A. For added authenticity, stick to the first 24 harmonics in the series; that’s the range of the original Synclavier.


One of the easiest and most popular methods to create original wavetables is Serum’s recent addition of image importing. While the process is straightforward—just drop any PNG into the wavetable view, even without opening the editor—the results can be somewhat chaotic at times. Sure you can import old baby pictures, but do they sound good?

The trick to making the most of this feature is to view some of the wavetables that you find most useful and consider their visual properties. Often, the coolest wavetables look a lot like topographical maps, so… Google “topographical maps”.

Just glancing over the top hits, I found several maps that sounded fantastic as wavetables. And when you’ve exhausted the possibilities of topographical maps, try another obvious image type that yields results that are almost as good: Ocean waves.



A hidden gem in Serum’s feature set is the ability to drag-and-drop any audio file in .WAV format directly onto its noise generator, instantly transforming it into a sampled audio source for layering it into your sound.

Fig. 12. Noise Generator panel. While you can’t edit the sample within Serum, the noise generator’s existing parameters make this feature an incredible way to add yet more complexity to your Serum patches. Here’s an easy reference for understanding how to manipulate your samples within Serum, as shown in Figure 12.

Image placeholder title

->| icon: This toggles the sample between one-shot and looping mode. When active, the sample plays to the end and stops. When off, the sample reaches the end and starts over.

Piano icon: This toggles keyboard tracking, which is crucial when using pitched samples. When off, the sample will play whatever pitch it’s tuned to via the Pitch parameter, disregarding note information, which can be good for adding clicks and drum transients to your sounds.

Pitch: At the time of this writing, the root key for a given sample can’t be adjusted directly, so to tune your samples to the keyboard, you’ll need to use a tuner plug-in or do it by ear. That said, when using samples with an original tuning of C3, setting this parameter to 36.00% often does the trick.

Phase: While you can’t edit your samples directly, the Phase parameter allows you to adjust the sample start point, which is especially handy when working in one-shot mode.

Rand: This parameter randomly offsets the start point each time you hit a key. If you use a bright sweeping texture as your sample source, this is a cool way to get simulate sample-and-hold effects.

Pan and Volume: If you don’t already understand these parameters, put the magazine down now.


Fig. 13. Route all three sources to the filter. While Serum’s sub-oscillator tuning is limited to octaves, it offers all of the standard subtractive waveforms including sine, triangle, saw, square, and pulse. These are great for traditional uses like adding low-end or reinforcing the fundamental in bright, buzzy patches; they’re also useful for timeless techno chords.

Image placeholder title

For the classic version of this this patch, switch the sub-oscillator’s wave to sawtooth and leave the tuning at the default, which will provide the tonic. Then, for a Detroit-friendly minor triad, tune the main oscillators to +3 (minor third) and +7 (fifth). Finally, make sure all three sources are routed to the filter, so they’re processed identically (see Figure 13). From there, add some filter envelope and chorus; you're set.



Fig. 14. Serum’s filters. While most softsynths include classic options like lowpass, highpass, bandpass, notch (and sometimes comb and formant) with different roll-off slopes, Serum’s filter includes ninety modes. While many of these are variations on those mentioned above, as well as 21 combination modes that are reminiscent of the original Oberheim Xpander, the “miscellaneous” group is of special note.

Image placeholder title

Here, you’ll find a few filter types that aren’t immediately obvious in their uses (See Figure 14). Granted, the EQs are relatively self-explanatory, as are the ring mod options. But it’s worth exploring the more esoteric modes below:

SampHold: Works a lot like a bit-crusher, with low “cutoff” values dissolving the sound into a mass of digital grunge. Try applying this filter type to one of the white noise options and the result will be quite similar to vintage video game explosions. Bottom line: It’s great for fans of the chiptune genre.

Reverb: At first listen, this mode doesn’t seem to add a traditional reverb tail to the sound, unless you use extremely low cutoff values, resulting in a tight room reverb. At its core, the Reverb filter is actually a feedback delay network that can be used for hard-to-reproduce physical modeling effects when the cutoff is static. When the cutoff is modulated, all bets are off, as it sounds a bit like a flanger through a wormhole. For squeaky transients, modulate the cutoff quickly with an envelope. For full-on mayhem, use an LFO.

Fig. 15. Algorithms are available in the Filter effect module.PRO TIP: All of Serum’s filter types are also available in its end-of-chain Filter effect, so if you want to hear the Reverb mode (or any other) in a different context, audition the options in the effect device (see Figure 15).

Image placeholder title

Formant I, II and III: Each of these formant filters is optimized for different vowel transitions when adjusting the cutoff parameter. I’ve used these modes extensively for analog choir effects.

German LP and French LP: These lowpass filter modes are amazingly warm, with an analog-like depth, though being located in the “Misc” category means they’re often overlooked. For the technically inclined, German is a “zero delay feedback” LPF, while French is a pair of LPFs cascading in series.

Fig. 16.Watch those parameters. Many of the filter’s parameters actually change, depending on the filter mode. The top row of parameters is always cutoff, resonance, and pan (which offsets the cutoff in stereo, with increasing widths). The bottom row includes drive and mix, which controls the wet/dry balance of the filtered signal. The center parameter on the bottom row is the wild card that you should always note when switching between types. For example, in EQ modes, it’s a boost/cut parameter. In formant modes, it allows further customization of vowel characteristics. In the multi modes, it governs filter morphing between the various slopes (see Figure 16).

Image placeholder title

PRO TIP: While you can individually modulate the volume of each oscillator via velocity, it can be a bit of a chore to assign and scale velocity to all four sources separately. My shortcut? Assign velocity to filter drive instead. Here, a little goes a long way, but it’s a more organic effect and is easier to manage as you finesse your sounds.



Serum’s eight LFOs are so complex and flexible, they’re worthy of their own guide for making the most of their features. Here’s a handy cheat sheet for shortcuts and hacks.

Folder Icon: The first thing many users notice is that there are no LFO waveforms. Why? Because almost every aspect of the LFO is customizable. Instead of a waveform selector, there’s a folder icon that leads to a menu of common waveshapes, sidechain bouncing effects, and a few more exotic shapes to use as starting points. Best of all, you can save your own creations to the user folder for repeated use.

PRO TIP: You can alt- (PC) or option- (Mac) click on the folder icon to advance through the preset shapes (or shift-alt-click for reverse direction).

Fig. 17. LFO with breakpoints.Env: This little button switches any of the LFOs to one-shot mode, thereby transforming it into a multi-stage envelope with Bezier curves and up to 64 breakpoints (see Figure 17). So if you want to whip up a patch with 7 envelopes (there are also three dedicated ADSR generators) and 4 LFOs, go for it. You can even use Shift+Command+Click (Mac)/shift+Ctrl+Click (PC) to create nested loops within your one-shot LFOs, blurring the lines even further.

Image placeholder title

Tempo-Synced Everything: Switching any of Serum’s LFOs to tempo-sync mode does more than just lock the rate to standard note divisions, including triplets and dotted values. It also allows you to sync the LFO fade-in (aka “rise”) as well as sync the delay segment before the onset of the fade.

Fig. 18.Copying LFOs: If you want to copy an LFO to another LFO slot to modify it for complementary modulation, just alt- (PC) or option- (Mac) drag its tab to the destination LFO. Everything is instantly duplicated and ready for further tweaks.

Image placeholder title

Step Sequencing: If you want to use an LFO as a step sequencer, hold the shift key and the drawing tool will allow you to quickly draw rectangular shapes that are quantized to the LFO grid size, which can be up to 16x16.