Creative Effects

VALVEFRAME

Four original amplifier models, a synthesised multi-microphone cabinet, and the studio chain around them.

£55 £69Launch price

See the measurements

Formats: VST3 · AU · Standalone
Operating system: macOS · Universal
Version: v1.0.0

What it does

A cabinet made of physics, not captures.

No sampled impulse responses ship with VALVEFRAME. Every cabinet response is synthesised from a description of the driver, the box, the microphone and where it is pointing — driver resonance, cone breakup modes, voice-coil inductance, baffle step, microphone response, proximity effect, then early reflections.

That is what makes microphone position genuinely continuous rather than a set of discrete captures, and it is why the whole cabinet section is a few hundred bytes of coefficients instead of tens of megabytes of audio.

The chain is mono up to the cabinet, as a real amplifier is. Stereo begins at the microphone panning.

Key features

Valves solved properly, offline

Plate current from Koren's empirical triode model. The load line is solved by bisection at load time over a 4096-point grid, then looked up per sample — bisection because plate current is monotonic in the unknown, so it cannot diverge; offline because we can then afford to converge properly instead of stopping after two Newton steps.

The dynamic parts stay dynamic

Grid conduction charging the coupling capacitor and dragging the bias toward cutoff — the splutter-and-recover of a hard-driven amp. A partly bypassed cathode resistor reducing gain at low frequencies only. Miller capacitance. Supply sag.

A power amp that changes class under drive

A push-pull pair whose conduction overlap narrows as it is pushed, so a crossover notch appears as the stage moves from class AB toward class B. Transformer saturation makes the core a signal-dependent gain, so low notes compress while the top end stays open.

Continuous microphone position

Three microphone slots, five microphone characters, and position that moves smoothly rather than stepping between captures. Each slot holds two convolvers so responses swap by crossfade rather than by a discontinuity.

Five reverbs from one network

One feedback delay network mixed each pass by an orthonormal Hadamard matrix, so the mixing step neither creates nor destroys energy. Spring gets a dispersive all-pass ladder; shimmer gets an octave-up shifter inside the feedback path, so the tail climbs rather than merely having a bright layer added.

Verified against published values

A high-mu twin triode with a 100k plate load and a bypassed 1k5 cathode resistor is the most documented gain stage in guitar electronics, and its voltage gain is about 60. The model returns 62.9.

The interface

The VALVEFRAME effects page: pre-EQ and post-EQ banks, tape and modulation sections, and delay, reverb and output stages, each a coloured panel. — VALVEFRAME · 1 of 2

The VALVEFRAME cabinet page: a two-by-twelve speaker cabinet drawn in perspective with a microphone beside it, above three mic slots each with its own model, position, distance, level and pan controls. — VALVEFRAME · 2 of 2

How it works

The signal path, as the plugin’s own documentation states it.

input -> gain staging / impedance -> compressor -> pre EQ
      -> [ 8x oversampled: drive -> preamp valves -> tone stack -> power amp ]
      -> cabinet + 3 mic blend -> post EQ
      -> tape -> modulation -> delay -> reverb -> widener -> limiter

Audio

Audio demos for VALVEFRAME are not published yet.

We would rather show you nothing than show you a demo that was not recorded through this plugin.

Measured

99 test cases and the bundled benchmark · Apple Silicon · 48 kHz

Latency: 194 samples · 4.04 ms
CPU, full chain: 3.3 % of one core
Aliasing below 16 kHz, tanh drive, oversampled: −153 dB
Aliasing without oversampling: −16 dB
Convolution vs direct convolution: 2.5e-7 relative
Oversampler passband ripple to 16 kHz: < 0.1 dB
Allocations in process(), across 400 blocks: 0, verified by replacing operator new
auval: passes with zero warnings and zero failures

Figures are produced by this plugin’s own test suite and benchmark, on the hardware and sample rate named above. They are not estimates.

Controls

Amplifier

Amp: Downtown Clean, Garage 30, Dream, Lo-Fi. All four are built up front and kept resident, so selecting one is a pointer change plus an 8 ms fade.
Drive: Transparent, Mid Focus, Soft Clip, Tape Drive.
Tone stack: Plus a presence control that works the way the real one does — by rolling the feedback off at the top rather than by boosting treble.

Cabinet

Cabinet: Chime 2x12, American 2x12, Stack 4x12, Alnico 2x12, Vintage 1x12.
Microphones: Close Dynamic, Vintage Ribbon, Large Dynamic, Studio Condenser, Valve Condenser — across three continuously positionable slots.

Studio chain

Compressor: Studio VCA, Optical or FET.
Modulation: Analog Chorus, Dimensional, Clean Chorus.
Space: Delay and five reverb characters including spring and shimmer.
Output: Widener and limiter.

Technical details

Formats: VST3 · AU · Standalone
Operating system: macOS 10.13 or later
Architecture: Universal — Apple Silicon and Intel
Plugin code: Icv1
Version: 1.0.0
Latency: 194 samples · 4.04 ms
CPU: 3.3 % of one core, full chain
Factory presets: 6

All reachable as host programs as well as from the browser.

Worth knowing

The four amplifiers are voiced toward the character of the American clean, British class-A chime, high-headroom clean and small-cheap-amp archetypes. None is modelled from measurements of a specific product and none claims any relationship to one.
The amplifier voicings are physically motivated and internally consistent, but have not been matched against measurements of reference hardware. The figures that are checked against published values are the valve operating point and gain.