When to Keep a Sub Oscillator Before Distortion in a High-Passed Synth Layer
Keep the sub oscillator before distortion when it produces a useful change in the layer’s surviving mids: stronger harmonics, a preferred texture, or a more suitable envelope. Make that decision with the final high-pass filter active and the alternatives matched for loudness. Removing the sub’s fundamental afterward does not undo its influence on the distortion.
For a producer building a distorted synth layer above a separate bass part, the concrete task is to decide whether that internal sub earns its place. The following routing checks and listening test separate a useful sound-design ingredient from unnecessary input level.
Why can a sub matter after its fundamental is filtered out?
Distortion changes the signal before the high-pass filter receives it. FabFilter’s Saturn 2 documentation explicitly describes harmonics generated by its distortion algorithms and explains that Drive controls how strongly the input drives the clipping stage. Those new harmonics can occupy frequencies above the sub’s original pitch. FabFilter Saturn 2 Help — Band controls
Consider an illustrative 50 Hz sine sub feeding a distortion stage, followed by a high-pass filter set to 150 Hz. Its third harmonic is 150 Hz, fifth harmonic 250 Hz, and seventh harmonic 350 Hz. If the distortion generates those harmonics, filtering the 50 Hz fundamental can still leave upper-frequency content. These are frequency calculations, not measured output levels; the distortion curve and drive determine which components become substantial.
When the sub and main oscillators enter the same nonlinear stage, their interaction also matters. Rod Elliott’s original technical demonstrations show that nonlinear processing of multiple tones creates intermodulation products, with the particular products depending on waveform and distortion symmetry. Simple sum and difference tones should not be assumed for every algorithm. Elliott Sound Products — Intermodulation: Something ‘New’ To Ponder
The practical inference is that the sub can alter existing midrange content as well as contribute its own harmonics. A post-distortion high-pass filter attenuates the resulting spectrum; it does not reconstruct the waveform that would have emerged without the sub.
Which signal path actually lets the sub shape the distortion?
Check where the oscillators combine, where distortion happens, and where the high-pass filter sits. A patch’s visible module names are insufficient if its routing sends the sub around the processor.
| Signal path | What the sub can contribute | Practical use |
|---|---|---|
| Main + sub → shared distortion → high-pass | Sub harmonics and interaction with the main signal | Test when the sub changes the desired midrange character |
| Main + sub → high-pass → distortion | Contribution from whatever sub content survives the first filter | Test when you want less low-frequency influence on distortion |
| Main → distortion → high-pass; sub on a separate path | Independent low-frequency support | Use when the bass should remain separately controllable |
| Main and sub → separate distortion stages → sum → high-pass | Separately generated distortion spectra | Test when shared distortion sounds too congested |
These are different processing arrangements, not interchangeable ways to remove bass. In particular, separate distortion stages omit the interaction created by summing both oscillators into one nonlinear processor.
Official synth documentation provides concrete routing examples. Phase Plant’s generator-area distortion processes the signal arriving from modules above it. In Serum 2, the Direct routing target bypasses the filter and effects section, while Main passes through effects. A sub assigned to Direct therefore cannot drive distortion in that bypassed effects path. Kilohearts Docs — Phase Plant, Serum 2 Manual — Routing an Oscillator or Filter
Before comparing sounds, trace the complete path through the DAW as well. A high-pass inside the synth may precede an external distortion insert. For this test, place the intended final high-pass after the distortion stage whose response you are evaluating.
When is keeping the sub worth testing?
Keep it as a candidate when muting it removes a quality you need from the filtered layer: a particular rasp, a fuller lower midrange, or an attack-to-sustain movement that fits the phrase. Those are listening criteria, not automatic benefits of enabling a sub.
A useful starting case is a monophonic sound with a sub tuned one octave below the main oscillator. That gives you a simple relationship to evaluate before introducing detuning, chords, or more complex modulation. Start with a sine so you can investigate its effect without adding the richer input spectrum of another waveform.
Also test whether the sub’s envelope is doing useful work. Compare a short sub decay with a sustained sub while leaving the main oscillator unchanged. If the filtered sound changes mainly at the beginning of the note, you may prefer the short version. If it changes throughout the sustain, decide whether that continuous texture belongs in the arrangement.
Keep the sub only if the improvement survives the musical context. A texture that sounds satisfying in isolation may occupy space needed by another part. Listen with the separate bass, drums, and the busiest accompanying instrument before committing.
How do you compare sub-on and sub-off fairly?
Use a three-way comparison to distinguish the sub’s spectral contribution from the simpler effect of driving the processor harder. Kilohearts documents that its distortion Drive boosts the input signal, producing heavier distortion; raising another input can therefore be an important alternative to adding a sub. Kilohearts Docs — Phase Plant
- Prepare a representative phrase. Include a low note, a high note, short articulations, and a sustained note from the actual part.
- Set the final high-pass first. Use the cutoff and slope intended for the arrangement, then hold them constant during the initial comparison.
- Create version A with the sub muted. Preserve the main oscillator, distortion, and envelopes.
- Create version B with the sub feeding the shared distortion. Raise its level gradually from silence.
- Create version C without the sub but with more pre-distortion gain or Drive. Try to approach B’s distortion intensity.
- Match playback loudness after the high-pass. Use a final output trim for each version, then alternate them in the mix.
Version C is the crucial control. If it delivers the same useful body and articulation more simply, the sub may be unnecessary. If B retains a distinct texture you prefer after matching playback level, you have identified a reason to keep it beyond extra drive.
Do not use input gain to match the final listening level: that changes the distortion you are comparing. Also account for automatic compensation. Saturn 2, for example, adjusts output level as Drive increases, but that does not establish equal loudness between different patches after your final filtering. FabFilter Saturn 2 Help — Band controls
Repeat the comparison several times without watching which version is active. If oscillator randomness makes repeated notes substantially different, temporarily reduce that randomness. Phase Plant’s manual distinguishes fixed starting phase from phase randomness applied to successive notes, providing one example of the controls to inspect. Restore the intended variation before the final decision. Kilohearts Docs — Phase Plant
How should you tune the sub level and high-pass together?
Treat the controls as two separate decisions. Sub level determines what enters the nonlinear stage; the final high-pass determines which parts of its output remain. Begin with a modest sub contribution, choose the useful distortion character, and then refine the filter in the arrangement.
For an illustrative setup, use a main oscillator at 100 Hz, a sine sub at 50 Hz, one shared distortion stage, and a final high-pass at 150 Hz. Sweep only the sub level first. Save the lowest setting that provides the texture you want, then compare nearby filter settings. These values describe a repeatable experiment, not a recommended crossover for every track.
Do not interpret the cutoff as a boundary below which nothing exists. Filter slope and resonance affect the transition. Phase Plant’s filter documentation, for example, provides different slopes and explains that higher Q introduces resonance near cutoff. A resonant filter can become another contributor to the sound you are judging. Kilohearts Docs — Phase Plant
After choosing a combination, play the entire phrase. A fixed high-pass acts on different harmonics as pitch changes. Check that low notes retain enough definition and high notes do not become disproportionately dense. If one setting works only on a single audition note, try less sub, a different envelope, or note-dependent filtering before keeping it.
What signs suggest reducing or removing the sub?
Use this decision aid after matching loudness and checking the routing:
| Observation | Next test | Decision criterion |
|---|---|---|
| Sub-on sounds better only before the final filter | Compare again with the final filter active | Keep it only if the delivered layer benefits |
| Sub-on sounds thicker but loses articulation | Reduce sub level or shorten its decay | Prefer the setting that preserves the phrase’s rhythm |
| More Drive without the sub gives an equally useful result | Compare versions B and C in context | Choose the simpler configuration if neither has a meaningful advantage |
| Repeated notes vary more than intended | Check starting phase and randomness | Keep the sub if the intended variation remains usable |
| Chords become congested through shared distortion | Compare separate processing or less sub | Judge the actual chord voicings, not a single note |
| The layer overlaps the separate bass excessively | Adjust the final filter and compare the combined parts | Keep the texture only if the full arrangement improves |
Distinguish unwanted digital artifacts from the character you are trying to create. FabFilter documents that strong distortion can introduce aliasing and that Saturn 2’s High Quality modes reduce it through oversampling, with additional CPU cost. If your processor offers a comparable option, audition it at the intended rendering setting before deciding that a brittle texture comes from the sub itself. FabFilter Saturn 2 Help — Input and output controls
Oversampling does not settle whether you like the interaction between oscillators. Evaluate that separately with the sub-on, sub-off, and extra-drive versions. Keep the smallest sub contribution that gives the filtered layer a useful, repeatable character across the part. If the advantage disappears after filtering, loudness matching, or playback in the arrangement, remove it and continue shaping the layer with the controls that remain.
