How EvoraVerb works

EvoraVerb is an algorithmic reverb plugin built around a feedback delay network. Two controls on the surface; eight delay lines, allpass diffusion and three-layer modulation underneath. This page explains what happens to your signal and how it differs from other reverbs.

Simple mode: two controls

Most reverbs give you a dozen knobs and leave the coherent combinations for you to find. EvoraVerb takes the opposite approach: Amount is a macro that rescales six internal parameters at once, along a path where every position is a usable space.

Internal parameter	At 0%	At 100%	What you hear
Size	0.20	0.95	Room dimensions
Decay	0.8 s	12.0 s	Tail length
Damping	0.55	0.30	Opens up in the highs as it grows
Pre-delay	10 ms	35 ms	Distance to the first reflection
Mod depth	0.15	0.55	Movement inside the tail
Width	0.70	1.00	Stereo spread

Turning Amount takes you from a small controlled room to a long enveloping space in one gesture. Mix sets the wet and dry balance. Both are exposed to your host for automation.

Advanced mode

Open the advanced panel and Amount steps aside: instead of driving the six parameters through a macro, you set each one directly. Nothing fights for control, because Amount only writes to them while the panel is closed.

The defaults are exactly the values the macro produced at 50%, so opening the panel never changes the sound you already had.

Control	Range	Default
Size	0 – 100%	57.5%
Decay	0.1 – 12 s	6.40 s
Damping	0 – 100%	42.5%
Pre-Delay	0 – 120 ms	22.5 ms
Modulation	0 – 100%	35%
Width	0 – 100%	85%

Reverb EQ

Two filters shape the tail: a low cut and a high cut. Both act on the reverb alone and never touch your dry signal, so you can clear space in the low end or take the edge off the top without dulling the source.

They stay out of the macro on purpose. Amount describes the space; where you cut frequencies depends on the material you are treating, not on the size of the room.

Filter	Range	Default
Low Cut	10 Hz – 1.5 kHz	Off
High Cut	100 Hz – 20 kHz	Off

Input and output gain

Two trims sit next to the meters, each ranging from −24 to +24 dB, for matching levels without reaching for another plugin.

Typed values and your own presets

Version 1.2 added two things the beta asked for. Every control accepts a value typed in directly, for when you know you want 2.4 seconds rather than somewhere near it. And you can save your own presets alongside the twenty that ship with the plugin.

User presets are stored as files on your machine, so they survive reinstalling and can be copied between computers.

The signal chain

Each stage hands the next a wavefront that is already more scattered, so the feedback network never receives a dry impulse. It sustains and blurs a cloud that already exists, which is where the metallic character of cheaper reverbs comes from.

Pre-delay

True stereo: left and right travel down independent lines instead of being summed to mono, so the width of your source survives all the way to the tank. Third-order Lagrange interpolation, with 500 ms of headroom.

Early reflections

Eight taps per channel on separate lines, with deliberately irregular timings that differ between left and right so the image does not collapse to the centre in mono. Each bounce loses highs, the way a real wall behaves.

Allpass diffuser

Four allpass stages in series per channel, again with different times left and right. This is the stage that does most of the work removing the metallic quality: it turns the wavefront into a dense cloud before it ever reaches the tank.

FDN tank

Eight delay lines with mutually prime lengths, fed back through a Householder matrix that preserves energy and spreads every line into all the others. Left and right are injected separately per line, and the matrix itself blends the channels over time.

Damping and low-end control

Every line carries its own one-pole lowpass in the loop, each with a slightly different spectral deviation so the eight do not all decay with the same timbre. A second, very slow filter removes the DC component that would otherwise accumulate inaudibly and destabilise long tails.

What makes it different

Three-layer modulation

What gives away an artificial reverb on long tails is the periodic, predictable sway of a pure sine LFO. EvoraVerb combines three independent sources on every line:

Layer	Frequency	Weight	Contributes
Main LFO	0.15 – 0.64 Hz	60%	Chorus-like movement
Slow drift	0.021 – 0.112 Hz	28%	Long-term breathing
Smoothed noise	Non-periodic	12%	Breaks audible repetition

Asymmetric by design

Reflection times, diffuser stages and output weights all differ between channels. Nothing is mirrored, so the stereo image stays open instead of pulling to the middle.

Built for real-time

No memory is allocated while audio is running — everything is reserved when the plugin is prepared. Every parameter is ramped, from 20 ms on Mix to 250 ms on Decay, so automation stays free of audible artefacts. Denormals are handled and the feedback loop carries a soft-clip safety limiter.

An interface that gets out of the way

Vector graphics, no bitmaps, at 512 × 512. It scales cleanly and stays readable next to the rest of your session.

Formats and system requirements

Formats	VST3, Audio Unit, Standalone
Windows	Windows 10 or later, 64-bit
macOS	macOS 11.0 (Big Sur) or later — Intel and Apple Silicon
Modes	Simple (2 controls) and advanced (6 controls + EQ)
Reverb EQ	Low cut and high cut, wet signal only
Channels	Stereo in, stereo out only
Latency	0 samples
Reported tail	6.0 s
Maximum decay	12 s
Processing	32-bit floating point
Sample rate	Follows the host
MIDI	Not used
Automatable parameters	12
Session state	Saved with your project
Presets	20 factory presets, plus your own
VST3 size	About 9.5 MB (Windows x64)
Version	1.2.0

AAX is not included. It requires the Avid SDK and a PACE signature, neither of which is part of this build, so EvoraVerb does not run in Pro Tools yet.

Frequently asked questions