How to Use Noise as a Modulation Source for a Bubbling Water Texture
To create a bubbling water texture, sample a noise signal once per bubble and use the held value to vary a short tone’s pitch. Shape each tone with a quick amplitude envelope, add a small pitch sweep, and vary the spacing between events. This gives noise a clear job: controlling the differences between bubbles.
This tutorial is for sound designers working in a modular synth or a software instrument with flexible modulation routing. The concrete task is to build a continuous texture from individual bubble-like sounds. The settings below are illustrative patch choices, not measured properties of water or a tested preset.
What should noise control in a bubbling patch?
Start with three separately adjustable features: the tone of each bubble, its duration, and the interval before the next event. Tsugi’s original water-synthesis walkthrough uses a dedicated Bubble generator with pitch-related radius and rise controls, per-event randomization, and adjustable trigger timing. That provides a useful design reference for constructing a simpler patch from general synthesis tools. Patch Anatomy #7 – Water Swish
For this patch, a sine oscillator supplies the audible tone, an envelope creates each short event, and noise supplies changing control values. VCV’s introductory manual explains the relevant distinction: audio signals provide sound, while control voltages can modulate parameters such as oscillator pitch or amplifier level. Getting Started — VCV Rack Manual
Give each modulation route a specific purpose. Let event-level randomness choose a bubble’s starting pitch. Let an envelope create its internal pitch movement. Add slower variation to the event rate only after the individual bubble sounds useful. This separation makes adjustments easier to hear and diagnose.
What tools do you need?
Use an instrument or modular environment with the following functions. If a conventional synth cannot sample its noise source into a modulation signal, use a modular patching environment for this method.
- A noise generator and sample-and-hold processor.
- A sine oscillator with a pitch modulation input.
- A triggerable attack-decay envelope and a voltage-controlled amplifier, or equivalent amplitude control.
- A clock or pulse generator.
- Modulation scaling and offset controls.
- A mixer and audio output.
- Optionally, a second envelope, another sample-and-hold channel, and a slew limiter.
VCV Rack is one suitable implementation. Its official setup guide describes adding modules and configuring the Audio module’s driver and output device. Establish a working audio output before building the modulation network. Getting Started — VCV Rack Manual
For explicit noise sampling, Bogaudio’s S&H accepts an external input and samples on a rising trigger or gate edge. Its AD envelope runs an attack-decay cycle when triggered. These functions provide the event-level control and amplitude shaping needed here. BogaudioModules — maker documentation
How do you build one clear bubble first?
Begin without noise modulation. Patch the sine oscillator into the amplifier, then connect the amplifier to the output mixer. Route the attack-decay envelope to the amplifier’s level control. Keep the amplifier’s manual gain closed so the oscillator becomes audible only when the envelope opens it.
Trigger the envelope manually or with a slow clock. Use these starting settings, adapting them to the available controls:
| Parameter | Illustrative starting setting | What to listen for |
|---|---|---|
| Oscillator frequency | 600 Hz | A clear, compact tone |
| Amplitude attack | 2–5 ms | A defined onset without a dominant click |
| Amplitude decay | 60–120 ms | A short event with an audible tail |
| Event rate | 3–5 triggers per second | Enough space to judge individual bubbles |
| Pitch-envelope duration | 30–70 ms | A brief internal pitch movement |
Add a second envelope to pitch, beginning with a small amount. Try an upward sweep, then invert the route and compare a downward sweep. Choose the direction that gives the desired character in your patch. Tsugi’s Bubble implementation specifically uses pitch movement to produce its characteristic liquid articulation; the envelope shape and direction suggested here remain sound-design choices. Patch Anatomy #7 – Water Swish
If only one envelope is available, split its output between amplitude and pitch, with separate depth controls. This ties the two contours together, but it is enough to establish the basic sound. Keep the pitch movement modest while adjusting the amplitude decay; a long, exposed sweep may draw attention away from the short bubble event.
How do you turn noise into a different pitch for every bubble?
Create this control route: noise → sample-and-hold → range scaling → oscillator pitch. Split the event clock so it triggers both the sample-and-hold and the amplitude envelope. Each event then receives a fresh sampled value that remains stable between triggers. The sample-and-hold behavior is documented in both VCV’s module manual and Bogaudio’s maker documentation. VCV Free Modules, BogaudioModules
Keep the noise out of the audio mixer during this step. Listen to how its control values change the oscillator. For a straightforward VCV implementation, connect a noise output to Bogaudio S&H’s input, send the clock to its gate input, and pass the held output through a scaling utility before the oscillator’s V/OCT input. Leave track-and-hold mode disabled.
Set a bounded pitch range before increasing the depth. For example, with a control signal spanning 0–10 V, multiplying by 0.1 and subtracting 0.5 V produces a range of −0.5 to +0.5 V. If the source has a different range, adjust the scaling accordingly; inspect it with a scope or voltage display instead of assuming its limits.
On a 1 V/octave pitch input, that illustrative range places a 600 Hz base frequency between approximately 424 and 849 Hz. The calculation is frequency = 600 × 2^control_voltage, where control_voltage is the scaled voltage in volts. VCV documents the octave-per-volt relationship used in this calculation. Getting Started — VCV Rack Manual
Use this range as a decision aid. If successive events sound too similar, widen it slightly. If they resemble a wandering melody, narrow it and shorten the amplitude decay. Leave a pitch quantizer out of the route while developing this continuous texture.
How do you vary timing without losing the bubbles?
A fixed clock is useful while tuning the sound. Once the pitch variation works, add gentle variation to the clock rate. Keep pitch changes tied to individual events, while allowing the timing control to move more slowly.
One approach is to sample noise with a second, slower clock, smooth that held signal, and scale it into the main clock’s rate input. Begin with small deviations around four events per second, then try a wider range such as two to seven. These are illustrative settings: the useful amount depends on the clock’s input response and the envelope duration.
Slew limiting makes the timing control move gradually between its targets. Bogaudio’s S&H offers an optional glide setting, while VCV Process provides a separate slew-limited output. Consult the chosen module’s timing units before entering a value. BogaudioModules, VCV Free Modules
Listen for the relationship between event spacing and decay. At four events per second, the average interval is 250 ms; an 80 ms bubble has room to finish. At twelve events per second, the interval is about 83 ms, leaving little separation. Raising the rate may therefore require shorter decays or additional independent voices rather than more modulation depth.
How do you make the texture fuller?
Duplicate the completed bubble voice before adding extensive processing. Give the second voice a different base pitch, decay, and clock rate. For example, try a lower voice around 400 Hz with a 110 ms decay beside a higher voice around 900 Hz with a 50 ms decay. Use separate sampled values so the two voices do not repeat identical pitch gestures.
A useful optional relationship is to make higher-pitched events shorter. Split the held pitch control, invert and rescale the second branch, and send it to decay time if the envelope supports that input. Treat this as an artistic mapping. Check the minimum and maximum decay values so an extreme control value cannot stretch one event into an unintended sustained tone.
Add a quiet background layer only after the bubbles read clearly. Filtered noise can provide a continuous bed, but increase it from silence and stop before it obscures individual events. Tsugi’s water patch likewise combines distinct generators for textural variety and uses restrained processing to preserve their character. Patch Anatomy #7 – Water Swish
How do you troubleshoot the result?
Change one route at a time and compare at a similar output level. The following sequence helps identify which part of the patch needs attention.
| What you hear | First adjustment | Follow-up check |
|---|---|---|
| Mostly hiss | Mute any direct noise-to-audio connection | Confirm noise still reaches the sample-and-hold |
| A continuous whistle | Close the amplifier’s manual gain | Confirm the amplitude envelope returns to zero |
| A repeating plip pattern | Add modest timing variation | Give additional voices separate clocks |
| A wandering tune | Narrow the sampled pitch range | Shorten exposed tonal tails |
| Clicks dominate | Lengthen the amplitude attack slightly | Check whether pitch jumps occur during an existing tail |
| A smeared texture | Shorten decay or reduce event rate | Reduce overlapping voices before adding effects |
| Little audible variation | Inspect the held control signal | Confirm the sample-and-hold receives triggers |
If pitch changes disturb an unfinished bubble, slow the clock until the envelope closes before the next sample is taken. Alternatively, use separate voices so one bubble can finish while another begins. Where trigger routing updates pitch late, a small delay on the envelope trigger can give the pitch value time to settle; assess that adjustment in the chosen environment.
Finish by muting the background layer and any spatial effects. The dry patch should still contain distinct short events, controlled pitch differences, and useful gaps or clusters. Restore the supporting layers gradually, keeping the noise modulation ranges narrow enough that each event retains the character you established in the first bubble.
