Every compressor reduces gain when a signal crosses a threshold. But whether that gain reduction happens through a JFET, an electroluminescent panel, a vacuum tube, or a VCA chip determines not just the sound character but the fundamental physics of how and when the circuit responds. The four dominant circuit topologies — FET, optical, vari-mu, and VCA — each impose different constraints on attack speed, release behavior, and distortion profile that no amount of parameter tweaking can circumvent.
FET: The 20-Microsecond Attack
The 1176 Peak Limiter, introduced by UREI in 1967, replaced transformer-coupled tube designs with an N-channel JFET as the signal attenuator. The JFET (junction field-effect transistor) acts as a voltage-controlled resistor in the signal path. As the gate voltage becomes more negative, channel resistance increases, reducing the amount of signal that passes. As gate voltage approaches zero, resistance drops and gain rises.
This architecture enabled attack times between 20 and 800 microseconds — faster than any preceding design. At 20 microseconds, the 1176 can grab transients that resolve in under one millisecond: kick drum attacks, snare peaks, consonant onsets in vocals. The JFET's response is limited only by the RC time constants of the gain control circuit, not by mechanical or chemical processes.
The 1176's distortion character comes from the JFET's non-linear transfer function. When driven hard, the device generates second and third-order harmonics — a subtle edge that engineers describe as punchy or aggressive. This harmonic profile explains why the 1176 became standard on drums and vocals rather than mastering chains.
The so-called "all-buttons-in" mode — engaging all four ratio buttons simultaneously — was not a designed feature. Pressing all buttons simultaneously shifts the bias point of the gain control amplifier, pushing the compression ratio dynamically between 12:1 and 20:1 depending on input level. Engineers discovered this mode in studios after the unit's release. The result is heavy, harmonically rich compression that saturates transient peaks without fully clamping them, which is why the mode became associated with drum room mics and bus saturation.
Optical: When Physics Replaces the Knob
The Teletronix LA-2A, introduced in 1962, has no attack knob and no release knob. That is not an oversight. It is the defining feature of how the circuit works.
The LA-2A's gain element is the T4B opto-attenuator: an electroluminescent panel illuminates a pair of light-dependent resistors (LDRs). The sidechain signal drives the EL panel — brighter light means lower LDR resistance means more attenuation. The fundamental constraint is that you cannot set a fixed time constant for a photoresistor responding to light. The LDR's resistance follows the intensity of the light source, and that intensity follows the audio signal's envelope through the EL panel's own frequency response.
The result is program-dependent behavior. The LA-2A's release follows a two-stage curve: roughly the first 50% of recovery happens in approximately 60 milliseconds, while full recovery from heavy gain reduction can take several seconds. The LDR retains a kind of memory — if the cell has been driven hard for an extended period, it releases more slowly. No fixed-time-constant circuit replicates this because it does not emerge from designed time constants but from material physics.
This behavior makes the LA-2A well-suited to leveling vocals and bass without audible pumping. The slow tail of the release curve keeps gain reduction applied gently as phrases end, while the fast initial recovery prevents dropout after short peaks. The interaction between these two phases tracks phrasing naturally, which is why the unit has no controls for time constants: the physics of the opto-cell are already doing the work.
Vari-mu: Variable Ratio as a Feature, Not a Bug
The Fairchild 670, introduced in 1960 and hand-built in approximately 800 total units, uses a different mechanism: the variable-mu triode. In a tube amplifier, the gain of the tube depends on its grid bias voltage. By feeding a rectified sidechain signal back to the grid, the circuit continuously re-biases the tube in proportion to input level. More input means more negative bias means less gain.
The critical difference from VCA and FET designs is that there is no fixed threshold where compression begins. The tube is always being re-biased, always partially compressing. The compression ratio rises gradually as input increases, starting below 2:1 at low input levels and climbing toward 10:1 or beyond at high levels. This inherently soft-knee behavior is not a design choice selected at specification time. It is a consequence of how triode gain varies with grid voltage.
Tubes operating in their non-linear region generate predominantly even-order harmonics, particularly second harmonic. Even-order distortion is perceived as warmth rather than edge — the characteristic "glow" of vari-mu units comes from the same mechanism that distinguishes tube guitar amplifiers from solid-state designs. The combination of program-dependent release, variable ratio, and even-order harmonic saturation is what engineers mean when they describe vari-mu compression as "gluing" a mix.
The Manley Variable Mu, introduced in 1994, modernized the Fairchild architecture with consistent production tolerances and a switchable high-pass sidechain filter. The filter prevents low-frequency content from driving the gain reduction circuit, which is particularly useful on mix bus applications where kick drum transients would otherwise cause the entire mix to duck.
VCA: Precision at the Cost of Character
The dbx 160, introduced in 1976 and designed by David Blackmer, brought a fundamentally different architecture: a dedicated voltage-controlled amplifier IC with the gain reduction element isolated from the signal path. The VCA chip receives a control voltage from the detector circuit and adjusts its gain accordingly. This lineage continues in THAT Corporation's 2181 series — current-in, current-out devices used in both hardware and the emulation models built on circuit simulation.
The VCA's advantage is predictability. Attack and release times are set by RC time constants with no program-dependent variation. THD in a quality VCA design runs below 0.05%, with predominantly odd-order harmonics. This translates to what engineers call transparency: the circuit shapes dynamics without coloring the signal.
This same predictability is also a constraint. VCA compressors do not add warmth, do not introduce LDR memory effects, and do not vary their ratio continuously with input level. On a mix bus, this can make a VCA compressor feel clean to the point of sterility compared to vari-mu alternatives. On a parallel drum bus or a mastering chain requiring precision gain control, that same cleanliness is an asset.
Choosing by Circuit, Not by Name
The topology directly determines what the compressor is and is not capable of. The 1176's JFET cannot produce the LA-2A's program-dependent release — the physics do not exist in the circuit. The LA-2A cannot achieve a 20-microsecond attack — the EL panel and LDR have no mechanism to respond that quickly. A VCA compressor cannot produce even-order harmonic saturation at nominal operating levels because the gain cell does not operate non-linearly under normal conditions.
When selecting a compressor for a specific task, the relevant question is not which topology is better but which physical mechanism matches the source material and treatment goal. Transient shaping on a snare benefits from the 1176's microsecond-scale attack response. Vocal leveling benefits from the LA-2A's program-dependent release that tracks phrasing without a fixed time constant. Mix bus cohesion benefits from a vari-mu's continuous, soft ratio behavior. Mastering chains requiring linear, predictable behavior benefit from a VCA's separation of detection and gain reduction.
The circuit topology is not metadata about the compressor. It is the compressor.