FALSE MEMORY — Documentation

Overview

FALSE MEMORY is a real-time audio disintegration effect. Feed it any sound and watch it decay. A tape left in the sun. A corrupted file half-recovered from a dead drive. A memory dissolving as you try to hold onto it.

Turn up MEMORY and the destruction begins. Everything else shapes how the sound falls apart.

Core Controls

Control	Description
MEMORY	Master intensity, from 0 to 100%. At 0 the destruction stages are off. As MEMORY rises, the analog and digital chains, Fog, Ghosts and Damage all engage further. It is the most important control. The default is 0.

Character

CHARACTER decides which of two chains you hear. Both chains process the same input at the same time.

Control	Description
ANALOG / DIGITAL	A linear crossfade between the analog chain and the digital chain. At 0 you hear only the analog chain, at 100% only the digital chain, and at 50% (the default) an equal mix of both.

Analog Types

At full MEMORY the analog chain rolls the highs off with a 12 dB per octave low-pass (two one-pole filters, so the figures below are the corner of each pole and the level is about 6 dB down there), saturates the signal and wobbles the pitch. The wow and flutter depths below are the modulation depth of a delay line. The pitch deviation you hear is smaller, roughly the figure in brackets.

Type	Character
Cassette	The low-pass ends at 4.5 kHz. Soft tanh saturation, up to 3x drive. Slow wow at 0.3 Hz (depth 30 cents, about 3 cents of pitch) and flutter at 10 Hz (depth 8 cents, about 25 cents of pitch) that only starts above about 30% MEMORY.
Reel	The low-pass ends at 8 kHz, so it stays the brightest. Clean up to a knee, then compression above it, up to 5x drive. Flutter at 15 Hz (depth 12 cents, about 55 cents of pitch). Wow at 0.5 Hz (depth 20 cents, about 3 cents of pitch) starts above about 30% MEMORY.
VHS	The low-pass ends at 3 kHz, with a mid scoop of up to 6 dB at 2.5 kHz. Asymmetric saturation that adds even harmonics, up to 4x drive. Wow at 1.5 Hz (depth 45 cents, about 20 cents of pitch) and flutter at 8 Hz (depth 20 cents, about 50 cents of pitch). At full MEMORY loud signals can fold back and clip.

Digital Types

The digital chain holds samples to lower the sample rate, quantizes to fewer bits, and freezes short slices of the sound. A 4 kHz two-pole low-pass on the digital chain fades in with MEMORY squared, to take the edge off. MEMORY sets how often the freezes happen and how far the sample rate and bit depth drop. It does not change the length of a freeze. All values below are at full MEMORY.

Type	Character
Lo-fi	The sample rate drops by up to 8x and the bit depth from 16 to 8. On average 0.3 freezes per second, each replaying the last 50 to 200 ms.
Circuit	The sample rate drops by up to 24x with a random wobble in the hold length, and the bit depth from 16 to 4. On average 0.8 freezes per second, each replaying the last 5 to 50 ms.
Data-mosh	The sample rate drops by up to 4x and the bit depth from 16 to 6. On average 1 freeze per second, each replaying the last 150 to 800 ms.

Each channel decides on its freezes by itself, so left and right glitch at different moments. A freeze replays the last slice of the sound once, taken after the bitcrusher, then crossfades back to the live signal over 10 ms. Freezes end on a block boundary, so with large host buffers a freeze can run a little longer.

Degradation

Control	Description
DAMAGE	Adds tape-style dropouts: sudden dips in level, with a soft 5 ms fade in and out. The amount scales with both DAMAGE and MEMORY, so with MEMORY at 0 nothing happens. A dip lasts 50 to 500 ms on Cassette, 80 to 800 ms on Reel and 90 to 900 ms on VHS, and dips get deeper as DAMAGE and MEMORY go up. They only reach silence at the top. Left and right dip independently. At high settings the level is also pulled down between the dips, by up to about 22 dB, and the auto-gain lifts part of that back. The default is 30%. Dropouts are applied after Fog and Ghosts, to the whole signal, whatever the CHARACTER position.
DRIFT	Random pitch wandering on the whole output, whatever the CHARACTER position. At full DRIFT the delay line moves toward random targets equal to 500 cents of modulation depth, gliding to a new target every 0.2 to 0.7 seconds. The pitch you hear wanders by tens of cents, at worst around 250. This is separate from the wow and flutter of the analog types. DRIFT also adds timing jitter to the digital chain, scaled by CHARACTER, and it works with MEMORY at 0.

Control

Description

DAMAGE

Adds tape-style dropouts: sudden dips in level, with a soft 5 ms fade in and out. The amount scales with both DAMAGE and MEMORY, so with MEMORY at 0 nothing happens. A dip lasts 50 to 500 ms on Cassette, 80 to 800 ms on Reel and 90 to 900 ms on VHS, and dips get deeper as DAMAGE and MEMORY go up. They only reach silence at the top. Left and right dip independently. At high settings the level is also pulled down between the dips, by up to about 22 dB, and the auto-gain lifts part of that back. The default is 30%. Dropouts are applied after Fog and Ghosts, to the whole signal, whatever the CHARACTER position.

DRIFT

Random pitch wandering on the whole output, whatever the CHARACTER position. At full DRIFT the delay line moves toward random targets equal to 500 cents of modulation depth, gliding to a new target every 0.2 to 0.7 seconds. The pitch you hear wanders by tens of cents, at worst around 250. This is separate from the wow and flutter of the analog types. DRIFT also adds timing jitter to the digital chain, scaled by CHARACTER, and it works with MEMORY at 0.

Atmosphere

Control	Description
FOG	Blends in a frozen, smeared version of the spectrum, made with a 2048-point FFT. At 0 the signal is untouched. At 100% you hear only the fogged signal. MEMORY sets how hard the spectrum freezes, so with MEMORY at 0 FOG has no effect.
GHOSTS	Mixes in audio from earlier in the sound, from a 3-second history of the input. The level is GHOSTS times MEMORY, so with MEMORY at 0 nothing is heard. The lookback is fixed and does not change with MEMORY.

Fog Types

Type	Character
Tonal	A gentle freeze. The spectrum of the last few hundred milliseconds lingers and the harmonic structure stays intact.
Inharmonic	The energy is smeared across neighbouring frequencies, from about 3 bins either side at the low end to about 30 at the top, which destroys the harmonic relationships and gives a metallic cloud. It freezes a little less than Tonal until MEMORY is very high.
Evolving	An almost static freeze once MEMORY is above about 90%. Three bands sweep across the spectrum at changing speeds and widths, boosting up to about 3 dB, while everything outside them sits about 9 dB down.

Ghost Types

Type	Character
Echoes	A copy of earlier audio from the 3-second history, played about an octave higher, with a delay that sweeps over time. The delay starts at 1.5 seconds, sweeps down to nearly zero, jumps back to about 3 seconds and sweeps again.
Imprints	Audio from about 1.5 seconds back, run through a one-pole low-pass (corner about 2.5 kHz) so it turns into a soft, smeared tone that no longer sounds like the original.
Granular	Up to 8 grains at a time, each 20 to 80 ms long with a cosine window, taken from random places in the last 3 seconds. Higher GHOSTS and MEMORY make new grains start more often and louder.

Texture

Control	Description
HISS	Pink noise. It works without MEMORY. The level depends on the analog type: VHS is the loudest, Cassette comes next, Reel is the quietest.
CRACKLE	Random pops of random polarity and size. The knob sets both how often they happen and how loud they are. It works without MEMORY. Cassette has the most pops, Reel the fewest but the loudest, VHS is in between.

HISS and CRACKLE are part of the analog chain, so the CHARACTER slider fades them out as you move toward Digital. At full Digital you hear neither.

Evolve

Control	Description
EVOLVE	Off, or a ramp that runs for 1, 4, 8, 16, 32 or 64 minutes.

While a ramp is running, it raises a floor from 0 to 100% over the chosen time. MEMORY, DAMAGE, FOG and GHOSTS never go below that floor. Each of them still follows its own slider when the slider is higher than the floor, so you can push them further, but you cannot pull them under the ramp. When the ramp finishes, the four sliders are set to 100%. Choosing another EVOLVE time restarts the ramp, and turning EVOLVE off removes the floor but leaves the four sliders where the ramp put them. The ramp counts processed audio, so it runs faster in an offline bounce, and its position is not saved in presets. DRIFT, CHARACTER, HISS, CRACKLE and MIX are not affected by EVOLVE.

Randomize

The RANDOMIZE button gives every control a random value, except MIX, HISS, CRACKLE and EVOLVE. That includes MEMORY, so a random roll can land on a clean sound.

Output

Control	Description
MIX	A linear dry/wet blend, smoothed over 20 ms. At 0 you hear only the original signal, at 100% only the processed one. The default is 100%.

There is no output level control. Anything above full scale is clipped hard.

Signal Flow

Input ─┬─ Analog chain:  HF low-pass → Saturation → Wow/Flutter → Hiss/Crackle
       │                 (Cassette / Reel / VHS)
       │
       └─ Digital chain: Bitcrusher → Buffer repeat → 4 kHz low-pass
                         (Lo-fi / Circuit / Data-mosh)

       Analog / Digital crossfade
             ↓
       Drift (pitch wander) → Fog → Ghosts → Damage (dropouts)
             ↓
       Auto-gain → Mix → Hard clip → Output

The Ghosts take their audio from the dry input, and their output is added to the processed signal.

Formats

AUv3 (iOS): runs inside AUM, GarageBand, Logic, or any AUv3 host. Multiple instances supported.
Standalone: runs as its own iOS app with built-in audio routing. iPhone and iPad.
VST3 + AU + CLAP (Mac): universal binary, Apple Silicon and Intel.
VST3 + CLAP (Windows): 64-bit.

Requires: iOS 14.0+ or macOS 11+ or Windows 10+.