Golden-Chaos Vibrato — User Guide
Deterministic multi-rate delay vibrato in which one sinusoidal delay trajectory is phase-modulated by two additional low-frequency oscillators. The presets use π, e, and φ to produce periodic or quasi-periodic modulation patterns with no random process.
What this does
Golden-Chaos Vibrato creates continuously varying delay modulation from three deterministic low-frequency rates. A primary sine controls the delay excursion, while two additional sine waves bend the phase of that primary trajectory. The result ranges from regular cyclic vibrato to slowly evolving quasi-periodic motion, depending on the frequency relationships chosen.
- Deterministic modulation: the same settings applied to the same Sound produce the same modulation trajectory.
- Three-rate phase modulation: one primary rate plus two secondary phase modulators.
- Fractional-delay interpolation: the source is read between samples using linear interpolation instead of rounded sample positions.
- Periodic and quasi-periodic presets: the preset family deliberately includes both behaviors.
- All channels preserved: every channel is processed with the same delay trajectory.
- Exact dry bypass: 0% Wet returns an untouched copy and skips the safety stage.
Quick start
- Select exactly one Sound in Praat.
- Run
Golden-chaos_vibrato.praat. - Choose a preset. Golden Shimmer is a useful moderate starting point; Pi Cycle gives a periodic trajectory; Mathematical Quasiperiodic gives denser multi-rate motion.
- Use Modulation depth to control how far the delay moves around the base value.
- Use Dry/Wet to choose between pure variable-delay vibrato and a dry+delayed blend.
- Leave Safety peak enabled if you want attenuation only when the mixed result exceeds the chosen ceiling.
Processing model
1. Local modulation time
The LFO uses local Sound time, t = x - sourceStart. Moving the Sound to a different Praat xmin therefore does not change the modulation pattern.
2. Phase-modulated delay trajectory
Rate1 is the primary modulation rate. Rate2 and Rate3 do not create independent delay taps; they phase-modulate the primary LFO by the amounts Rate2 mix and Rate3 mix.
3. Fractional-delay read
The delay in milliseconds is converted to a non-integer sample offset. For each output sample, the script reads the two neighboring source samples and linearly interpolates between them:
At the beginning of the Sound, source indices are clamped to the valid sample range. The operation is applied separately to every channel using the same delay trajectory.
4. Dry/Wet mix
At 100% Wet, the output is the interpolated variable-delay signal. At 0% Wet, the script takes an exact bypass path and does not execute the vibrato or safety attenuation.
5. Safety ceiling
If Safety peak is above zero, the script measures the finished mixed Sound. It applies gain reduction only when the measured peak exceeds that value:
This is an attenuation-only ceiling, not target normalization. A quiet result is not raised.
Periodic and quasi-periodic modulation
The behavior depends on the relationships among the three modulation rates.
Quasi-periodic set: when at least one active rate is incommensurate with the others, the combined phase trajectory does not close into one finite repeating cycle.
Version 0.4 therefore does not treat every π/e/φ preset as equivalent. Pi Cycle uses π, 2π, and π/2, which are rationally related and therefore periodic. The other mathematical presets use at least one incommensurate relationship and are presented as quasi-periodic.
Presets
| Preset | Base | Depth | Rates (Hz) | PM amounts | Behavior |
|---|---|---|---|---|---|
| Custom | form values | form value | form values | form values | User-defined |
| Golden Shimmer | 5 ms | 0.08 | φ / 2φ / 1/φ | 0.3 / 0.5 | Quasi-periodic |
| Euler Wobble | 7 ms | 0.15 | e / 2e / 1 | 0.8 / 0.2 | Quasi-periodic |
| Pi Cycle | 6 ms | 0.12 | π / 2π / π/2 | 0.2 / 0.1 | Periodic |
| Mathematical Quasiperiodic | 8 ms | 0.20 | π / e / φ | 1.0 / 1.0 | Quasi-periodic |
| Subtle Irregularity | 4 ms | 0.05 | π / e / φ | 0.5 / 0.5 | Quasi-periodic |
| Deep Math Texture | 12 ms | 0.25 | π/10 / e/10 / φ/10 | 0.7 / 0.7 | Quasi-periodic |
The presets overwrite the delay, depth, rate, and phase-modulation values shown in the form. Dry/Wet, Safety peak, visualization, and playback remain user-controlled.
Parameters
| Parameter | Default | Implemented behavior |
|---|---|---|
| Preset | Custom | Selects one of six predefined rate/delay sets or leaves the manual values unchanged. |
| Base delay | 6 ms | Centre of the variable delay. If necessary it is automatically reduced so the maximum modulated delay stays below about 45% of the Sound duration. |
| Modulation depth | 0.14 | Fractional delay excursion around the base; clamped to 0…0.95. |
| Rate1 | π Hz | Primary LFO rate. |
| Rate2 | e Hz | First phase-modulation rate. |
| Rate3 | φ Hz | Second phase-modulation rate. |
| Rate2 mix | 0.6 | Phase-modulation amount for Rate2; clamped to -4…4. |
| Rate3 mix | 0.4 | Phase-modulation amount for Rate3; clamped to -4…4. |
| Dry/Wet | 100% | Linear dry/delayed mix, clamped to 0…100%. 0% is exact dry bypass. |
| Safety peak | 0.99 | 0 disables safety. Otherwise only attenuates when the result exceeds the selected peak. |
| Draw visualization | On | Draws input, output, measured control trajectory, and summary. |
| Play result | On | Plays the created Sound after rendering. |
B(1-D) to B(1+D).Audio and object behavior
- Offline processing: the complete Sound is rendered in Praat; this is not a real-time processor.
- Channel count: preserved. All channels receive the same modulation trajectory.
- Sample rate: preserved.
- Duration: preserved.
- Start time: preserved.
- Randomness: none. There is no seed because no random process is used.
- Output level: unchanged unless the optional safety ceiling has to attenuate it.
Vibrato versus chorus-like mixing
A moving delay changes the local read speed and therefore creates pitch modulation on the wet path. Mixing that wet path with the original introduces frequency-dependent interference in addition to the pitch motion. For the clearest delay-based vibrato, use a high Wet value; for a more chorus-like color, use an intermediate Dry/Wet value.
Visualization
The figure is a process view of the actual render settings rather than a decorative mathematical diagram.
Input
A mono display copy of the source waveform. For multichannel Sounds this conversion is used only for drawing; the DSP itself remains multichannel.
Output
A mono display copy of the rendered result, again used only for visualization.
Delay trajectory
The next panel draws up to the first three seconds of the exact control equation used by the processor. The horizontal reference is the base delay; the plotted curve shows the phase-modulated movement between the minimum and maximum delay limits.
Summary
The final strip reports Base delay, Depth, Wet percentage, Safety setting, all three rates, both phase-modulation amounts, and channel count.