PAA Audio Plugins Bundle [MAC]

• Products: Legend Channel, Not Another Glue-Comp, Vintage Delay, Tape56K
• Developer: PAA Audio
• Version: 2026.07
• Format: VST3, AU
• Requirements: macOS 12 or later
• Source: paa-audio.de

Legend Channel is a channel-strip plugin modeling a British console channel at the circuit level — the transformer, transistor, and VCA stages solved from the original schematics rather than sampled or curve-fit. It runs preamp, high/low-pass filtering, a 4-band EQ, and full dynamics (compressor plus gate/expander) in one signal path. The transformer’s magnetic hysteresis and the NE5534-based gain stages are computed in real time, so the saturation character shifts with input level and gain setting instead of following one fixed curve. It ships native — VST3, AU, AAX — for macOS, Windows, and Linux, no iLok or subscription.

Key Takeaway

Legend Channel’s real distinction is the PCCN circuit model: it solves the actual transformer and transistor equations in real time rather than applying a fixed saturation curve, so tone changes with how hard it’s driven. That real-time modeling costs more CPU per instance than a curve-based “analog” plugin, so stacking it on every channel of a large session adds up fast.

Solved in Real Time, Not Curve-Fit

Most plugins claiming analog character use one of three shortcuts: a fixed saturation curve, an impulse-response snapshot of the hardware at one setting, or a neural model trained to imitate its output. Legend Channel’s PCCN engine instead solves the transformer’s magnetic hysteresis and the amplifier/VCA stages as the actual NE5534 and transistor circuit equations, computed per sample rather than looked up from a table.

Because the model runs the real equations, its behavior changes with input level and gain the way the physical hardware’s would — transformer bloom at higher drive, VCA grab under gain reduction — instead of scaling a single fixed curve up and down. A static IR snapshot can’t reproduce that: it captures one operating point, and everything off that point is interpolation.

Running physical equations per sample is heavier than evaluating a lookup curve, so Legend Channel costs noticeably more CPU per instance than a tanh-based “warmth” plugin. A mix leaning on one detailed channel strip for two or three key tracks gets the most out of the modeling; loading it on all 40+ channels of a dense session will show up in the CPU meter.

Preamp Drive Without Preamp Gain

The mic/line preamp carries an input-transformer stage with a dedicated “Preamp Drive” control that pushes the transformer’s iron toward saturation independently of the main gain — so a track can pick up harmonic density without also getting louder, rather than drive and level being tied to one knob.

High- and low-pass filters run at 12 dB/octave, steep enough to clear low rumble or high hiss off a track without reshaping the midrange the way a shallower 6 dB slope would. Both sit ahead of the EQ in the signal path, so filtering decisions get made before tonal shaping rather than after.

A vocal with plosive rumble below 80 Hz can get filtered clean before any EQ move touches the vocal’s actual tone; a track that just needs brightness without any low-end cut doesn’t need the high-pass engaged at all.

Four Bands, Proportional Q

The EQ runs four bands — HF, HMF, LMF, LF — each switchable between shelf and bell, with proportional Q that narrows automatically as gain increases, the way passive and console EQs typically behave rather than holding a fixed bandwidth at every gain setting.

That proportional behavior means a wide, gentle boost and a narrow, aggressive boost on the same band come from turning one knob further, not from switching a separate Q control — fewer parameters to manage per band, at the cost of not being able to dial an unusually wide Q at high gain if the situation calls for it.

Engineers used to console-style EQ get a fast, musical result without hunting for Q; someone who wants surgical, independently-set Q and gain on every band will find this EQ’s coupling of the two a constraint rather than a shortcut.

Compression and Gating in One Path

The compressor runs ratio up to a limiting setting, with auto-release and auto-gain options, so a fast setup can lean on the automatic behavior while a slower one can dial release and makeup by hand. The gate/expander section adds range, hysteresis, duck mode, and an external key input — hysteresis specifically keeps a gate from chattering on a signal hovering near the threshold.

Both dynamics stages share the same circuit-modeled VCA behavior as the rest of the strip, so gain reduction carries the same level-dependent character described in the PCCN section rather than switching to a clean digital VCA model for just the dynamics section.

A vocal that needs both de-essing-adjacent gating and glue compression can do both without a second plugin in the chain; a source that only needs a light limiter and nothing else is running two dynamics stages’ worth of processing for a one-stage job.

One Path, One Character

The signal path ends at output trim and a fader with a peak meter, and the whole strip resizes from 100% to 150% for different screen setups. Presets store the full chain — preamp, filters, EQ, dynamics — as one state rather than per-section.

Because every stage in the strip shares the same PCCN-modeled circuit behavior, moving a signal through preamp, EQ, and dynamics compounds that one character at each stage rather than mixing a modeled preamp with a separate, differently-behaved EQ plugin downstream. A session built entirely around this one channel character gets a more consistent throughline than swapping between different vendors’ plugins stage to stage; a mix that wants a completely clean EQ and only wants coloration from the preamp will get coloration at every stage whether it’s wanted there or not.

FAQs