How the reverb is tested.

First, prove the analyser.

Constructed signals have known answers for Decay accuracy, echo density and stereo behaviour. The analysers must recover those answers inside declared tolerances before they score an engine.

If the source lacks enough dynamic range or coverage, the tool returns a refusal.

Scope: this qualifies the measurement method.

One render, any buffer split.

The same input produces the same samples through Lush, Natural and the echo under irregular host buffers. Modulation uses a fixed seed. Delay timing follows sample rate.

Result: changing the DAW buffer does not change the render.

0 allocations after prepare.

Memory for the echo and both engines is reserved before playback starts. Processing stays inside that prepared storage during parameter changes and engine transitions; the same rule covers tail inspection and reset.

Checked by: the allocation harness and the audio-path source audit.

Tail reporting follows sample rate.

The declared duration is checked at 44.1 / 48 / 96 / 192 kHz. It covers ring-out below the defined floor and avoids an order-of-magnitude safety margin.

Host consequence: a DAW can keep processing after the input stops.

Artwork and hit areas share a source.

One vector build emits the panel and its control-less faceplate. The same build also emits the theme plus the interactive-control manifest.

The editor scales that vector artwork to the host window. Each sound control binds to a host parameter, and engine selection remains a two-position switch.

Gate: tests compare the shipped surface with the approved visual contract.

The engine measurements sit inside their current reference envelopes.

Qualified repeat captures establish statistical corridors for Decay and stereo behaviour; echo-density growth has its own corridor. Both engines currently sit inside those bounds. Lush’s echo-density bloom lands 1 ms from the reference band.

Limit: proximity characterises measured behaviour. It does not approve the sound, and it makes no sound-alike claim.

The Decay dial has a tolerance.

Each engine renders 5 settings between 0.4 and 12 seconds. The matrix repeats at 6 sample rates between 44.1 and 192 kHz, giving 60 cells.

Measured audible-band decay must land within 8% of the dial. All 60 cells currently pass.

Claim: the displayed seconds describe the rendered decay inside that tolerance.

The product surface reaches the engines.

The transition manager preserves a ringing tail when the engine changes. Echo Into Space and pre-delay feed the selected network, while the six reverb controls and constant-power Mix complete the path.

There are 13 host parameters. The same parameter set drives 32 factory presets, A/B comparison and the energy monitor. CLAP passes its validator; VST3 plus Audio Unit adapt that verified bundle.

Boundary: these checks cover the implemented plug-in surface.