Step 01 / 4
Tune the three bell components #
Set giStage init 1 in the full script. Use triangles at +28 st/+30 cents and +30 st/−15 cents, at −21 and −24 dB. Add a sine one octave below the played pitch at −6 dB. Those unequal high intervals against the low foundation establish the bell character. The sine sub omits the source 12% tone coloration. Use 1 ms attack, 4.3 s decay to −38 dB and 350 ms release.
| Csound control | Value |
|---|---|
| High tones | +28 st +30 cents / +30 st −15 cents |
| Levels | −21 / −24 / sub −6 dB |
| Sub | Sine, octave down |
| Amplitude | 1 ms / 4.3 s / −38 dB / 350 ms |
| Instrument / touch | −7 dB / 50% |
Sound after step 1 — WebM/Opus
Next: Round the high tones with a fixed filter →Step 02 / 4
Round the high tones with a fixed filter #
Set giStage init 2 in the full script. Use a 1650 Hz ladder lowpass at 0.34 resonance, with no note tracking or envelope movement. Higher bell components become softer as they rise beyond the fixed cutoff. This replaces PRD with a different resonant response.
| Csound control | Value |
|---|---|
| Lowpass | 1650 Hz, resonance 0.34 |
| Tracking / modulation | Off |
Sound after step 2 — WebM/Opus
Next: Surround the tone with a damped room →Step 03 / 4
Surround the tone with a damped room #
Set giStage init 3 in the full script. Feed the preceding signal through 8 ms predelay into reverbsc, with feedback 0.6271, 4500 Hz damping and 44% wet mix. Predelay leaves the initial articulation in front; feedback prolongs the tail, while damping darkens repeated circulation. The source room decay setting is 1.5 seconds, but Csound feedback is not a decay-time control: this is an interpretation of that space, with different diffusion and modulation.
| Csound control | Value |
|---|---|
| Source decay target | 1.5 s (not a measured Csound decay) |
| Predelay / feedback | 8 ms / 0.6271 |
| Damping / wet | 4500 Hz / 44% |
| Room side multiplier | 1 |
| Room input cuts | Off |
Sound after step 3 — WebM/Opus
Next: Add panning and digital grain →Step 04 / 4
Add panning and digital grain #
Set giStage init 4 in the full script. Move the left and right gains oppositely at 1 Hz by ±22%. This moves the bell and preceding room together. The simple complementary gain model interprets Auto Pan without its additional shape controls. Hold incoming samples at 16000 Hz and round their amplitudes to 12 signed bits, then blend 40% of that branch with the direct tone. The held samples add staircase texture and the rounding roughens quieter portions. This simple clocked quantizer interprets Redux without its proprietary jitter or shaping behavior. No pre- or post-filter is added.
| Csound control | Value |
|---|---|
| Panning | 1 Hz sine; ±22% opposite channel gain |
| Hold rate | 16000 Hz |
| Signed bits / wet | 12 / 40% |
Sound after step 4 — WebM/Opus
Audio & settings ↓