Step 01 / 5
Build the detuned saw voices and slow amplitude contour #
Define the CirrusVoice opcode and three voice calls from the complete orchestra. Begin with the saw branch only: use aBasic as the opcode output. Keep the bandpass and lowpass bypassed, and listen directly to the amplitude-shaped voice sum before effects.
kAmp linsegr 0, 0.85, 1, 2.5, ampdb(-2), 2.4, 0
kAmp = kAmp^1.6
The 850 ms attack softens the saw edge; the power curve makes the middle of the rise quieter. Use detuning of −7, 0 and +7 cents, with ±2.2 cents of slow pitch drift. Weight the outside voices toward opposite stereo channels and keep the center voice shared. The −8 dB instrument gain and 0.65 multiplier set the level before effects. Longer held notes reveal beating; shorter notes may release before the attack reaches its peak.
Sound after step 1 — WebM/Opus
Next: Add pulse width movement and serial filters →Step 02 / 5
Add pulse width movement and serial filters #
Restore aBasic + aPulse in the voice opcode. The pulse is −16 dB, 5 cents above each saw voice. Its width moves around 0.29 by ±0.025 under the shared 0.09 Hz sine LFO. Keep this LFO in instrument 90 so overlapping notes share the same motion.
kFilterEnv linsegr 0, 1.8, 1, 3, 0.6, 2.4, 0
kCutoff = 850 * pow(2, 0.8*kFilterEnv + 0.15*(p4-60)/12)
aBpL butterbp aL, kCutoff, kCutoff/0.85
aLpL butterlp aBpL, 6200
Apply the same filters to the right channel. The bandpass focuses the middle frequencies while its slow envelope opens the tone. The 6200 Hz lowpass softens the upper edge. Multiply both filtered channels by kAmp * (0.55+0.45*iVelocity) * ampdb(-8) * 0.65 into the shared buses. Higher pitches open the bandpass slightly; velocity changes level without changing the envelope timing.
Sound after step 2 — WebM/Opus
Next: Widen the pad with a three-tap ensemble →Step 03 / 5
Widen the pad with a three-tap ensemble #
In instrument 90, highpass the effect inputs at 220 Hz, then use three vdelay3 taps centered on 20 ms. Modulate each delay by ±3.6 ms at 0.24 Hz, with sine phases 0, 1/3 and 2/3. The staggered movement thickens the sustained tone.
kMod1 oscili 3.6, 0.24, giSine, 0
aTap1 vdelay3 aHpL, 20+kMod1, 40
Use the complete orchestra's tap routing: the center tap receives the mean of both highpassed channels. Mix 78% dry and 22% wet, with the outer tap weighted 75% and the center tap 25% in each wet channel. Convert to mid/side and multiply the side by 1.05. Increasing delay depth makes the pitch movement more obvious; increasing wet level brings the moving image forward.
Sound after step 3 — WebM/Opus
Next: Add a damped stereo hall →Step 04 / 5
Add a damped stereo hall #
Feed the ensemble output through 28 ms of predelay into reverbsc.
aPreL delay aChL, 0.028
aPreR delay aChR, 0.028
aRevL, aRevR reverbsc aPreL, aPreR, 0.84, 3300
aMixL = aChL*0.75 + aRevL*0.25
aMixR = aChR*0.75 + aRevR*0.25
Feedback 0.84 gives a lingering hall, while the 3300 Hz feedback damping darkens repeated circulation. The 25% wet mix leaves the filtered attack audible. Longer predelay separates the pad from its room; lower damping cutoff makes the tail darker. These controls approximate the character of the Ableton room rather than matching its decay control numerically.
Sound after step 4 — WebM/Opus
Next: Focus the lows and finish the output →Step 05 / 5
Focus the lows and finish the output #
Convert the reverberated signal to mid/side. Highpass only the side channel at 180 Hz, then reconstruct left and right. This reduces low-frequency spread while leaving upper frequencies wide.
aMid = (aMixL+aMixR)*0.5
aSide = (aMixL-aMixR)*0.5
aSide butterhp aSide, 180
aOutL dcblock2 (aMid+aSide)*ampdb(10.9)
aOutR dcblock2 (aMid-aSide)*ampdb(10.9)
outs aOutL, aOutR
clear gaL, gaR
The +10.9 dB output gain restores level after the filters and effect blends. dcblock2 removes DC offset. Clearing both buses prevents old note samples from accumulating on the next control block. The finished sound has a slow, focused center with moving upper harmonics and a diffuse stereo tail. See the complete orchestra for the connected implementation.
Sound after step 5 — WebM/Opus
Audio & settings ↓