Basilica Audio

Aureate — how-to guide

Practical settings for the orchestral saturation glue stage, grounded in the factory presets.

Where it belongs

Aureate runs after the individual layers of an orchestral/choral stack (or a doubled guitar) have already been balanced — a string or brass bus, an orchestral/choir submix, or, sparingly, the full mix bus as a subtle master-glue pass. It is not a distortion or amp-sim tool: Drive is capped at a modest 24 dB, and the default Warmth/Character settings stay inside "adds richness," not "adds grit" — that role belongs elsewhere in the suite (Overture/Tenebrae).

Typical use cases: gluing a divisi string/brass section into one instrument, cohering an orchestral/choir submix before it meets the metal instrumentation, subtle master-bus glue, and (v0.3.0) mix-bus-style dynamics via the Glue section.

Quick-start settings

Subtle master-bus glue — Master Glue (Subtle)

Character Tape, Drive 3 dB, Warmth 25%, Mix 45%, Output 0 dB.

Low Drive and a Mix well under 100% is the deliberate move for a master-bus pass — this is meant to be felt more than heard, the same role a summing bus plays in a hybrid analog/ITB workflow.

Section glue — String Section Glue, Console Summing Sheen, Choir Warmth

String Section Glue: Character Tape, Drive 5 dB, Warmth 30%, Tone +5%, Mix 100%. Console Summing Sheen: Character Console, Drive 4 dB, Warmth 20%, Tone +5%. Choir Warmth: Character Tape, Drive 4 dB, Warmth 50%, Tone −15% (darkened).

Console is chosen deliberately for the "summing sheen" preset — it's the Character that stays transparent at low-to-moderate drive and only shows character once pushed harder, closer to a solid-state/transformer summing-bus archetype than Tape's more immediately audible glue.

A genuinely different circuit class — Valve Push, Brass Bloom

Valve Push: Character Valve, Drive 14 dB, Warmth 60%, Bias 15%, Output −2 dB. Brass Bloom: Character Console, Drive 8 dB, Warmth 40%, Bias 10%, Tone +10%.

Valve is the most asymmetric, even-harmonic-forward of the three Character options — a rounder, tube-like push, and the one with the highest Warmth-driven bias ceiling. Positive Bias pushes the asymmetry further still, independent of Warmth's own contribution.

Glue-section bus compression — Orchestral Bus Glue, Soft Tube Glue, Iron Bus Weight

Orchestral Bus Glue: Glue on, Glue Model VCA, Threshold −6 dB, Ratio 4:1, Attack fast, Release fast, Makeup 2.5 dB, Character Console, Drive 4 dB. Soft Tube Glue: Glue Model Vari-Mu, Threshold −10 dB, Character Valve, Tone −6%. Iron Bus Weight: Glue on (VCA), Iron 65%, Character Console, Output −1 dB.

These three show the v0.3.0 Glue section in its most useful shapes: VCA for classic, predictable console-bus behaviour; Vari-Mu for a softer knee and an intrinsic, programme-dependent release; Iron pushed hard for low-end transformer weight rather than compression character.

Parallel/New York saturation — Parallel Grit (New York)

Character Valve, Drive 20 dB, Warmth 80%, Mix 50%, Output −3 dB.

Because Mix is sample-accurately delay-compensated, a heavily driven, characterful wet signal blends back under the clean dry signal without phase smearing from the oversampling latency.

Recipes

  1. Divisi string section into one instrument. String Section Glue as a base, Tone nudged +3 to +8% if the section needs to read a little more present against guitars. Why: a little Drive and Warmth glues a section together the way tape/console summing naturally does — the goal is a section that reads as one instrument rather than a pile of close mics, not an audible saturation effect.
  1. Orchestral submix before it meets the metal instrumentation. Console character (stays transparent until pushed), Drive 4–6 dB, Mix 100%, sitting after individual sections are already balanced and before the material reaches the full mix bus. Why: Console's "least characterful until pushed" behaviour means it can sit on a whole submix without audibly coloring quieter passages, only adding character where the material is actually loud enough to reach it.
  1. Choosing VCA vs. Vari-Mu for bus glue. Reach for VCA (Orchestral Bus Glue) when you want classic, predictable console-bus behaviour and want the Attack switch to actually do something; reach for Vari-Mu (Soft Tube Glue) when you want a softer knee and a release that changes shape with the programme rather than a fixed time constant. Why: see the theory box below — the two aren't just different presets of one law, they're different detector/gain-cell topologies entirely.
  1. Iron for weight, not for compression. Iron Bus Weight as a base, Iron between 40–70%, Drive kept moderate (6–8 dB) since Iron's own low-end saturation adds plenty on its own. Why: because transformer core flux is the integral of the applied signal, the core saturates far harder at low frequencies for the same level — third-harmonic content rises steeply toward the bottom end on its own, which is what reads as "weight" rather than "fizz," without needing to push Drive into audible distortion territory.
  1. Auditioning Drive without it becoming a loudness contest. Enable Auto Gain while you dial in Character/Drive/Warmth, so what you're comparing is character, not level. Why: Auto Gain is a one-point calibration per Character (measured against equal-RMS pink noise), a deliberate listening aid rather than a mastering tool — turn it back off once you've settled on a Drive setting and gain-stage manually from there.

Theory — why a vari-mu-style law glues differently than a VCA. Both of Aureate's Glue laws share the same feedback topology — a detector reading the signal after the gain cell, one sample old — which is what produces a soft knee and a program-dependent release on both, rather than a separately curve-fitted knee shape. Where they genuinely diverge is what sits inside that loop. The VCA law solves a straightforward dB-domain timing network: its attack time is a real, settable time constant, and doubling a transient's size roughly doubles how fast the gain reduction changes — an exponential, predictable response. The Vari-Mu law swaps in a current-limited rectifier and a three-capacitor release network modeling a tube-limiter-style gain cell, and that changes the physics: attack is no longer a dial-able number because it's intrinsic to how fast the rectifier can respond, a bigger overshoot takes proportionally longer to reach its own gain reduction (not less), and the four Release positions' hardware-class markings (0.1/0.3/0.6/1.2 s) don't match their measured recovery times (roughly 0.3/0.8/2/5 s) — because the reservoir capacitor's recovery genuinely depends on how long it was charged, which is exactly the "glue" behavior a fixed time constant can't produce.

Pitfalls