How to Make Serum 2 Sound Analog: Warmth, Drift and Grit

What does 'analog warmth' mean in a serum patch?

For sound design it helps to think of analog warmth as a set of behaviours rather than a single effect. You can actually recreate several of those behaviours in Serum 2:

• Voice variation: individual notes have slightly different tuning, filter settings, or envelope behavior
• Pitch movement: tuning fluctuates gently instead of remaining perfectly still.
• Grit: noise interacts with the pitched sound rather than sitting on top as a separate layer.
• Nonlinear shaping: distortion changes the waveform and its character.

Some analog polysynths use separate voice circuits, each containing the components needed for a complete synthesizer voice. Calibration differences can make one voice slightly sharper, another brighter, and another slower to respond. Serum 2’s voice controls let you introduce this kind of variation deliberately.

Your goal does not have to be a perfect replica of a particular instrument. A more useful target is a patch that feels less uniform while still doing its job in your arrangement.

Choose a Starting Approach for Your Sound

Start with a simple saw based patch so you can hear what each change does. Turn on a filter and then shape the amplitude envelope for the section you're working on. For sustained chords you'll want to extend the release, for a bassline you might want to go into a lower register and select monophonic voicing.

Build one behaviour at a time. Listen to the repeated notes, the held notes and the actual musical phrase before you go and add another layer. That makes it far easier to tell the difference between useful character and excessive instability.

Recreate Voice-to-Voice Differences in Serum 2

Open Serum 2’s Global page and find the voice control panel. This gives you per-voice adjustments across up to eight modeled voice slots.

These controls are especially useful for chords because you can make the individual notes differ slightly without applying the same movement to the entire sound.

Add Small Detune Offsets

1. Select the detune parameter in the voice control panel.
2. Draw small differences across the voice slots rather than an identical value for every voice.
3. Play the same note repeatedly to hear the voice allocation cycle through the offsets.
4. Play chords and reduce the differences if the tuning becomes distracting.

This produces a different result from simply making the whole patch wobble. Each allocated voice has its own tuning offset, so the variation depends on which voice plays the note.

Vary Filter Cutoff Between Voices

Select the filter cutoff parameter and draw another restrained pattern. One voice can be slightly more open, while another is slightly darker.

Repeated single notes make this variation easy to identify. Chords reveal how those differences work together. Keep the changes small enough that the notes still belong to the same patch.

This is a useful distinction: you are changing the brightness of individual voices, not sweeping every note with one shared filter movement.

Use Six Voices for a Prophet-6-Inspired Pan Pattern

You can set the modeled voice count to six as a starting point for Prophet-6-inspired behavior. On the pan page, distribute the voice positions across the stereo field so that each voice has its own placement.

Play chords to hear the result. This creates voice-based stereo spread, rather than giving every note the same stereo treatment.

Matching the voice count and pan behavior does not turn Serum 2 into a complete Prophet-6 emulation. It recreates a particular aspect of the distribution of voices.

Choose Fixed Variation or Randomization

The patterns you draw give you a defined set of voice differences. If you want less predictable behavior, increase the associated randomization controls slightly.

• Use drawn offsets when you want control over how the voices differ.
• Add randomization when the pattern still feels too regular.
• Keep both restrained when the patch needs to remain clearly pitched and consistent.

Add Smooth Pitch Drift With Sample and Hold

For a smoother form of instability, use a sample-and-hold LFO to modulate tuning. This is a less noisy alternative to using noise as an FM source.

1. Set LFO 1 to a sample-and-hold shape.
2. Increase smoothing so the changes are less abrupt.
3. Assign the LFO to global tuning.
4. Use bipolar modulation so tuning can move above and below its starting point.
5. Start with a very small amount, around 1% in the modulation setup, then scale it down further if necessary.
6. Adjust the rate and smoothing until you hear gentle movement rather than obvious pitch jumps.

Treat that percentage as a starting point for the modulation amount, not as a specified pitch change in cents. The important test is whether the patch remains musically stable.

Per-voice detune and global tuning modulation serve different purposes. The first introduces differences between allocated voices. The second adds movement through a global tuning destination.

If LFO routing is unfamiliar, this guide to Serum LFO controls and modulation gives you useful background for working with these assignments.

Use Filtered Noise as an FM Source for Flutter

When smooth drift is not enough, noise-based frequency modulation can introduce a more fluttery, imperfect texture. At higher amounts, it can move toward an intentionally broken-synth sound.

Set Up Oscillator C as the Noise Modulator

1. Load a noise source into oscillator C.
2. Low-pass-filter that source.
3. Disable pitch tracking for oscillator C so the noise source does not follow the played note.
4. On the oscillator producing your pitched sound, select FM from oscillator C.
5. Raise the FM amount slowly while holding a note.
6. Back it down until the effect is barely noticeable, then refine the noise filtering.

Using oscillator C for this job leaves the dedicated noise oscillator available for a separate audible noise layer.

The low-pass filtering helps you shape the modulation source before it affects the pitched oscillator. You can temporarily open the pitched sound’s filter to hear the flutter more clearly, then return to the darker tone you want.

For subtle warmth, aim for movement that becomes most obvious when you remove it. For a worn-out texture, deliberately increase the modulation depth. Those are two different sound-design goals, not better and worse versions of the same setting.

Make Bass Noise Interact With the Oscillator

Simply increasing the noise oscillator’s level adds noise alongside the pitched sound. For a more integrated, gritty bass texture, use the pitched oscillator itself to control the noise level.

Route Oscillator A to Noise Level

1. Move your patch into the bass register and enable monophonic voicing.
2. Enable the noise oscillator at a restrained level.
3. Choose oscillator A as a modulation source.
4. Route it to the noise oscillator’s level.
5. Reduce the modulation depth immediately if the result is abrasive.
6. Open the modulation mapping curve and experiment with its shape.

This is audio-rate modulation: the rapid fluctuations of oscillator A control the noise level in real time. The result differs from a slow LFO or an envelope that simply fades noise in and out.

Shape the Mapping Curve, Not Just the Amount

Serum 2 lets you change how the modulation source maps to its destination. By emphasizing different portions of the mapping curve, you can make the noise interaction smoother or buzzier.

Compare several curve shapes at a low modulation amount. If you only adjust depth, you may miss a more suitable texture available through the mapping itself.

You can also make the noise layer stereo. That is a separate choice from setting the patch to monophonic voicing: one concerns stereo presentation, while the other concerns how many notes you play at once.

Finish the Character With Distortion and Dynamics

Once the oscillator behavior feels right, use distortion to move the patch away from a clean saw-wave sound. Then decide whether compression and external processing add anything useful.

Try Tube Distortion First

Tube distortion is a useful starting point when you want a rounder character. Raise the amount gradually and listen to the interaction between the oscillator, noise and filter.

Do not assume that more drive means more warmth. Stop when the tone has the character you need.

Explore Asymmetric Distortion

Asymmetric distortion lets you shape the positive and negative portions of the waveform differently. This gives you another way to adjust the patch’s character rather than relying only on a drive control.

Make small curve changes and compare them. Once the distortion is working, revisit filter resonance. The combination can produce a smoother, more characterful response than the untreated patch.

Add a Fast Compressor and Check Your Levels

A fast compressor can be another finishing step, but check the output as you add processing. Distortion and compression can cause the patch to clip.

If you want additional clipping as part of the sound, use a deliberate downstream stage. In Ableton Live, an analog-style clipping stage is one option. Do not confuse intentional clipping with an overloaded output you have not checked.

Keep Reverb and Tape Processing Optional

For a spacious finish, you can add reverb such as Blackhole Immersive. Airwindows ToTape, a free tape-processing plugin, is another optional external stage.

These effects are finishing choices, not prerequisites for analog-style behavior. Establish the voice variation, drift, and grit first. Then add space or tape character if the part calls for it.

Listen while moving the filter cutoff, too. The noise texture can become noticeably crunchier as you close the filter, giving you another expressive variation to use.

Common Mistakes When Making Serum 2 Sound Analog

• Adding every technique at once: start with the behavior your patch actually lacks, then reassess.
• Using too much tuning modulation: excessive depth changes subtle drift into audible detuning.
• Expecting randomization to replace voice offsets: fixed differences and random changes give you different kinds of variation.
• Confusing audible noise with a noise modulator: noise mixed into the output is not the same as noise driving FM.
• Ignoring the modulation curve: the mapping shape can be as important as modulation depth for audio-rate noise control.
• When using effects to hide a weak patch, reverb and tape processing should complement the core sound.
• Overlooking clipping: check levels after adding distortion, resonance and compression.

If the finished sound needs tonal correction in the mix, use these practical EQ workflows rather than adding more sound-design processing without a clear purpose.

A Practical Next Step

Make two variations of the same simple patch. For the chord version, use small per-voice detune and cutoff differences, then add pan spread. For the bass version, focus on oscillator-driven noise and distortion. Compare each version with the untreated patch in your arrangement.

For more sound-design practice, explore our Serum LFO tutorial.

The useful test is not whether your patch proves digital synthesis can replace every analog instrument. It is whether a few controlled imperfections make the sound work better in your music.