Ambi Decoder Adaptive

s3g Ambi Decoder Adaptive 64 renders first- through seventh-order ACN/SN3D to a curated speaker layout while changing high-frequency focus from the field's current directional confidence and transient activity. Low frequencies retain a restrained diffuse decode, so the rendering can sharpen clear events without making the whole spectrum chase momentary direction estimates.

The adaptation is signal analysis inside the decoder. It does not identify sources, classify content, or use object metadata.

Ambi Decoder Adaptive interface showing its Icosahedron 20 speaker field, EPAD and MaxRE sharp decoder settings, focus, diffuse, confidence, transient, crossover, and smoothing controls — Adaptive keeps the layout and sharp decoder explicit while Focus, Diffuse, Confidence, Transient, Crossover, and Smooth define its content-reactive balance.

Workflow

Insert s3g Ambi Decoder Adaptive 64 after the encoded source and any Ambisonic processing on a 64-channel track. Match ORDER to the source, then choose LAYOUT, sharp-path MODE, and WGT.
Route each active output lane to the corresponding physical speaker in the selected layout.
Set XOVR to separate the stable low foundation from the adaptive high band, then raise FOCUS while audio is playing.
Balance directional CONF and onset-sensitive TRAN, retain low diffuse energy with DIFF, and set SMTH for the desired response time.

Signal Flow

Every active ambisonic input channel is divided into low and high components at XOVR.
The low band is decoded through a restrained BASIC, INPHASE diffuse matrix.
The high band is decoded twice: once through that diffuse matrix and once through the selected sharp MODE and WGT.
First-order directional confidence and positive W-envelope change produce the two adaptation cues.
A smoothed focus value crossfades the two high-band decodes; the low diffuse branch is added, OUT is applied, and unused output lanes are silenced.

Diffuse and Sharp Paths

The sharp path uses the selected BASIC, EPAD, or MMD matrix and the selected NONE, MAXRE, or INPHASE weighting. The diffuse path is deliberately fixed to BASIC plus INPHASE, with decoder width limited to a narrower internal range. This keeps the comparison defined: the adaptive value changes the high-band balance between a stable diffuse reference and the user-selected focused decode.

DIFF controls retention of the low diffuse foundation. Its range trims that branch from 75% to 100% rather than muting it completely. It does not set the high-band crossfade; the analyzed focus value does.

Confidence, Transients, and Focus

Directional confidence compares the magnitude of first-order X/Y/Z energy with the omnidirectional W component above XOVR. A coherent directional event produces a stronger confidence cue than a diffuse or cancelling field. The transient cue follows positive changes in a short W amplitude envelope.

CONF and TRAN are cue weights, not thresholds. Internally, the focus target follows FOCUS × (CONF × directional confidence + TRAN × transient cue), limited to the valid blend range. SMTH then controls how quickly the high-band decode moves toward that target. With both cue weights at zero, the high band stays on the diffuse path regardless of FOCUS.

Layouts and Field View

The layout menu shares the curated Speaker decoder geometries: Quad, Cube 8, Cube 17, Cube 41, Dodeca 12, Icosahedron 20, Octo Ring, Dome 24, Dome 25, LPAC 41, SRST 25, Quad+OH, and Sphere 24. The selected layout sets the active leading output lanes.

CUSTOM generates a full-sphere field using the active count inherited from the previously selected layout. The Adaptive editor does not provide the Speaker decoder's count control, custom-position editor, or per-speaker mixer. Choose a curated layout first when a particular count is required, and use Ambi Decoder Speaker for a measured custom array. Top, Side, and three-quarter field views, view rotation, and zoom affect only the display and are saved with plugin state.

Controls

Control	Meaning
LAYOUT	Selects a curated speaker geometry and its active output count, or generates `CUSTOM` at the current count.
MODE	Selects `BASIC`, `EPAD`, or `MMD` for the sharp high-frequency path.
ORDER	Matches first- through seventh-order input and activates the corresponding leading ACN channels.
WGT	Selects `NONE`, `MAXRE`, or `INPHASE` weighting for the sharp path.
FOCUS	Sets the maximum analyzed contribution of the sharp high-band decode.
DIFF	Retains between 75% and 100% of the low diffuse branch.
CONF	Weights the high-band directional-confidence cue.
TRAN	Weights the positive-onset transient cue.
XOVR	Sets the low/high analysis and decode split from 20 Hz to 5 kHz.
SMTH	Sets focus smoothing from 0 to 500 ms.
OUT	Sets final output gain from -60 to +12 dB.

INIT restores Sphere 24, EPAD, 3OA, MAXRE, 75% Focus, 35% Diffuse, 70% Confidence weighting, 35% Transient weighting, a 650 Hz crossover, and 45 ms smoothing while preserving the current OUT trim.

Order and Routing

Order	Active input channels	Output bus
1OA	4	64 fixed; the selected layout uses its leading speaker lanes and the remainder are silent.
2OA	9
3OA	16
4OA	25
5OA	36
6OA	49
7OA	64

Both decode paths use the complete active order. Only the confidence analysis is first-order. Match output lanes to the selected layout before connecting physical outputs.

Limits

The split uses a compact one-pole low/high decomposition rather than a linear-phase mastering crossover. Strong adaptation can make localization and brightness move with the program, especially at a low XOVR or short SMTH. Begin with moderate focus and compare against FOCUS 0% as the stable diffuse reference.