Hollow wordmark

Hollow · Spectral Sidechain Ducking

Frequency-selective space, keyed by another track

Cross-signal spectral sidechain ducking that makes space for the part that matters. Hollow is a cross-signal spectral sidechain tool for dense metal mixes. It listens to a KEY signal from another track and reduces only the overlapping frequency content in the MAIN signal, so a vocal can sit inside a guitar wall or a kick can claim its low end without turning the whole track down. It is still in design, so this page describes how it will work.

Planned14 of 15Built for Windows and macOS

Hollow is designed around two inputs: a MAIN signal that gets processed, and an EXTERNAL KEY signal from a different track in the host. A detector compares the spectral content of the KEY against the MAIN and applies frequency-selective gain reduction to the MAIN only where the two overlap and only when the KEY is active. Strength, selectivity, attack and release shape how much of the spectrum is reduced and how quickly it recovers. It is a conflict-resolution tool between two sources, not a resonance suppressor inside one source and not a multiband compressor.

Not released. No build exists yet.

What is Hollow for?

You have a harsh vocal that is loud enough on the meter but still disappears inside two double-tracked rhythm guitars in the 1.5 to 4 kHz range. A full-band sidechain from the vocal to the guitar bus fixes the intelligibility but audibly pumps the guitars and dulls the pick attack every time the singer takes a breath. What you actually want is for the guitars to give up only the narrow band the voice occupies, only while the voice is present, and to leave the rest of the wall intact.

What it does

Hollow is designed around two inputs: a MAIN signal that gets processed, and an EXTERNAL KEY signal from a different track in the host. A detector compares the spectral content of the KEY against the MAIN and applies frequency-selective gain reduction to the MAIN only where the two overlap and only when the KEY is active. Strength, selectivity, attack and release shape how much of the spectrum is reduced and how quickly it recovers. It is a conflict-resolution tool between two sources, not a resonance suppressor inside one source and not a multiband compressor.

What it doesn’t do

  • It is not an automatic resonance suppressor inside a single signal; that is Exorcise territory.
  • It is not a multiband compressor with fixed crossover bands; that is Schism.
  • It does not work without a host that supports external sidechain routing, and it will not fake a KEY display when none is connected.
  • It is not a full-band sidechain ducker; reducing the entire MAIN level is the problem it exists to avoid.

The design.

Controls and displays as currently designed. They can change before any build exists.

01 / 04

Main and Key, kept separate

The design treats MAIN and KEY as two distinct signals with two distinct spectra. The MAIN spectrum is what you are shaping; the KEY spectrum is the reference that decides where reduction happens. Both are drawn in different semantic colours so you can see at a glance which content belongs to which source.

Overlap between the two is highlighted as its own layer, and the actual reduction applied to MAIN is shown as a third measurable category. If no external key is routed, the plugin should say so plainly rather than animate a fake detector.

  • MAIN and KEY spectra drawn separately, never merged into one graph
  • Overlap highlight shows where the two sources compete
  • Reduction shown as its own layer, distinct from both input spectra
02 / 04

controls

The planned controls are Key Source status, Strength or Depth, Spectral Selectivity, Frequency Focus or Range, Detection Sensitivity, Attack, Release, Mix, Output, Key Listen and Bypass, with optional Stereo Link. Strength sets how far the MAIN is allowed to drop in the overlapping region; Selectivity sets how narrow that region is; Frequency Focus limits the detector to the range where the conflict actually lives.

Attack and Release shape how fast the ducking engages and recovers. Key Listen lets you hear the KEY signal to confirm routing before you commit. None of these ranges are final, and lookahead or linear-phase behaviour is not claimed.

  • Strength and Selectivity are the two primary shaping controls
  • Frequency Focus keeps the detector out of ranges with no conflict
  • Key Listen confirms external sidechain routing before mixing
03 / 04

What the meters will show

The planned display has three real categories: the MAIN spectrum, the KEY spectrum, and the reduction applied to MAIN. A gain-reduction-over-time trace shows when ducking is active and how deep it goes. A KEY status indicator shows connected, silent, or missing.

Every one of these is intended to be driven by real signal once the audio engine exists. Until then, any concept mockup is labelled as a design preview, and no meter is animated on a timer to imitate a working detector.

  • Three separate categories: MAIN, KEY, applied reduction
  • GR-over-time trace with real units in dB
  • KEY status visible at all times, including when unconnected
04 / 04

Where it sits in a chain

Hollow belongs after the tonal shaping of the MAIN source and before shared space or bus processing. On a guitar bus, that usually means after EQ and saturation and before any bus compressor or reverb. On a vocal, it usually means after the vocal's own dynamics and before delay or reverb sends.

The KEY is normally taken pre-processing from the competing source, or from a dedicated send, so the detector reacts to the source rather than to a processed version of it. Exact placement depends on what you want the detector to hear.

  • After MAIN tonal shaping, before shared bus processing
  • KEY usually taken pre-processing or from a dedicated send
  • Not a substitute for arrangement or level decisions

Three ways to use it

Step-by-step starting points from real heavy sessions. Every mix is different; trust your ears over any recipe.

Harsh vocal inside a rhythm guitar wall
  1. Insert Hollow on the rhythm guitar bus, not on the vocal.
  2. Route the lead vocal track to Hollow's external sidechain as the KEY.
  3. Set Frequency Focus around the upper-mid range where the voice competes, roughly 1.5 to 4 kHz as a starting point.
  4. Raise Strength until the vocal reads clearly, then widen Selectivity only if the guitars start to sound hollowed out.
  5. Use Key Listen to confirm the detector is hearing the vocal and not a reverb return, then compare bypassed and engaged at matched level.
Kick and bass low-end conflict
  1. Insert Hollow on the bass bus and route the kick track as the KEY.
  2. Set Frequency Focus to the low range where kick fundamental and bass overlap.
  3. Start with moderate Strength and a short Release so the bass recovers quickly between kick hits.
  4. Check the GR trace for pumping on sustained bass notes and lengthen Release if it moves.
  5. Verify in mono that the kick still reads and the bass has not lost its body.
Snare midrange slot against guitars
  1. Insert Hollow on the guitar bus and route the snare close mic or snare bus as the KEY.
  2. Focus the detector on the snare body range rather than the full spectrum.
  3. Use a fast Attack so the ducking catches the transient and a Release short enough to clear before the next hit.
  4. Watch the overlap highlight to confirm the reduction is landing where the snare actually sits.
  5. A/B against a static EQ cut on the guitars to hear whether dynamic ducking is buying you anything.

Before you reach for it

The mistakes we see most often with this kind of tool.

Common mistakes

  • Using Hollow as a full-band ducker: pulling Strength and Selectivity wide so the whole MAIN drops instead of a band.
  • Keying from a processed or reverberant copy of the source, so the detector reacts to reverb tails instead of the performance.
  • Setting Release so long that the MAIN never recovers between hits, which reads as pumping rather than space.

Factory presets

Make space between competing instruments where and when masking happens. The factory library runs to dozens of presets, often more than a hundred, covering technical jobs, genres and instrument categories, sorted by category and searchable. Three of them are below.

Kick Through Bass

Insert on the bass with the kick on KEY. Strength sets how far the bass ducks, and Frequency Focus limits it to where the two compete, so the rest of the bass stays.

Vocal Through Guitars

Insert on the guitar bus with the vocal on KEY. High Selectivity cuts only the frequencies the vocal is using at that moment, so the guitars lose little weight.

Snare Through Guitar Wall

Insert on the guitar bus with the snare on KEY. A fast response dips the guitars on each hit, focused on the snare’s crack and body.

Preset families: Kick/Bass, Bass/Guitar, Vocal/Guitar, Snare/Guitar, Lead/Rhythm and Vocal/Instrument masking. Each preset states which signal is processed and which is the key.

Why KRIG Audio

Before KRIG we spent years mixing with other makers’ plugins, and we learned what a heavy session needs from a tool. So every plugin starts with the sound: DSP written to a measured target and tested until it behaves the same way every time.

  • Built from years at the deskWe mixed and mastered with plugins from many makers for years, on our own records and on other people’s, and the same problems kept costing us time: gates that chatter, saturation that smears a riff, limiters that flatten the kick. Each KRIG plugin starts from one of those problems.
  • DSP written to a measured targetWe write our own signal processing. Before the first line of code we decide what the plugin must do to the signal: the curve it draws, the harmonics it adds, the noise it must not. The DSP is measured against that target, and when the two disagree, the code goes back.
  • Tested again and againSine tones, sweeps, impulses and noise at the levels people mix at. Every control swept while audio plays, to catch clicks and zipper noise. Every sample rate and buffer size a session can run at. Then real multitracks, at matched level, judged by ear. Every new build runs it all again.
  • Meters that tell the truthEvery meter and display is fed signals whose values we already know, and has to show exactly that. A VU reads like a VU, gain reduction reads gain reduction, and nothing on the screen moves for show, so what you see is what the plugin is doing to your track.
  • Presets for every jobDozens of factory presets per plugin, many with more than a hundred: technical fixes, genre starting points and instrument categories, sorted and searchable. A beginner starts from the job in front of them; an engineer gets a tested reference to build on.
  • Buy once, keep itEvery KRIG plugin is a one-time purchase with a perpetual licence, never a subscription. A 14-day full-feature trial is planned, so you can judge each plugin in your own sessions before you pay.

Not sure it’s the right tool?

How Hollow differs from its closest KRIG neighbours.

  • Exorcise · Exorcise suppresses moving resonances inside one signal; Hollow uses a second signal as the key to duck the first.
  • Schism · Schism compresses fixed frequency bands of one signal; Hollow reduces only where two separate sources overlap.
  • Sigil · Sigil is a manual parametric and dynamic EQ; Hollow's reduction is driven by an external KEY rather than by hand-set bands.

Where Hollow stands

Hollow is planned. No build exists yet, and the design can still change. There are no release dates until a plugin is ready. Follow the roadmap

  1. Research & specificationPlanned
  2. Interface designPlanned
  3. DSP implementationPlanned
  4. Validation & QAPlanned
  5. Release preparationPlanned
  6. ReleasedPlanned

Details

What we can state today. System requirements, tested hosts and test results are published here before release.

Type
Spectral Sidechain Ducking
Status
Planned
Platforms
Built for Windows and macOS
Formats (planned)
VST3 for Windows and macOS, AU for macOS. No AAX
Presets
Factory library, see above
Made for
Beginners to mastering engineers
Licence (planned)
Perpetual, one-time purchase. Computers and updates confirmed when the plugins are finished
Trial (planned)
14-day full-feature trial

Hollow questions

It is not an automatic resonance suppressor, not a multiband compressor, and not a full-band sidechain ducker. It is a cross-signal spectral tool that reduces the MAIN only where it overlaps with an external KEY. If you want to tame resonances inside one source, that is Exorcise; if you want band-split compression of one source, that is Schism.

A beginner can load it, route the KEY, and get a usable starting point from the factory presets, while an engineer can go deep on Selectivity, Frequency Focus, Attack and Release.

KRIG plugins are being built for Windows and macOS.

Exorcise works inside a single signal, finding and suppressing moving resonances without any second input. Hollow requires a second signal as the KEY and reduces the MAIN only where the two spectra overlap. One is self-contained resonance control; the other is conflict resolution between two sources.

Yes. Hollow's whole function depends on routing a second track into its KEY input, so the host must support external sidechain. If no key is connected, the plugin should report that state plainly rather than simulate a detector.

Next on the roadmap

Thrall

Automatic Gain Rider

Thrall wordmark