Step 01 / 5
Build the sine body and touch response #
Begin with instrument 1 and the shared output instrument 90 from the full script. Set giStage init 1 for the carrier alone. giSine contains one sine cycle; the carrier starts at phase zero for a repeatable attack.
aAttack linseg 0, 0.001, 1, 1, 1
aBody expsegr 1, 1.10, 0.0001, 0.350, 0.000001
aAmplitude = aAttack * aBody * ampdb(-3) * ampdb(-10)
aAmplitude = aAmplitude * (0.4 + 0.6*iVelocity)
The attack gate reaches full level in 1 ms. The exponential body falls to a near-silent floor over 1.10 seconds; releasing a held note starts a 350 ms fade from its current value. The floors are positive because exponential segments cannot end at zero. These curves interpret Operator's envelope rather than duplicate its internal shape.
Carrier gain is −3 dB before a −10 dB instrument gain. Velocity scales level by 0.4 + 0.6 × velocity, so gentler strikes remain audible. The source is centered; stereo space arrives later.
Sound after step 1 — WebM/Opus
Next: Add a short FM edge and tracked lowpass →Step 02 / 5
Add a short FM edge and tracked lowpass #
Set giStage init 2. Feed a second sine oscillator into the carrier's instantaneous frequency. This is a two-oscillator serial path: the modulator is not mixed directly into the output.
iRatio = 3.051
iIndex = 8 * ampdb(-12) * (0.3 + 0.7*iVelocity)
aModEnvelope expsegr 1, 0.172, 0.0001, 0.100, 0.000001
aModulator oscili iHz*iRatio*iIndex*aModEnvelope, iHz*iRatio, giSine, 0
aCarrier oscili aAmplitude, iHz+aModulator, giSine, 0
A ratio just above 3:1 gives the strike a slightly inharmonic edge. The 172 ms modulator decay removes the bright sidebands much sooner than the body fades. Modulator velocity response uses 0.3 + 0.7 × velocity: stronger strikes raise brightness as well as volume. The factor 8 translates the source's −12 dB modulator level into a Csound FM index; it is an interpretation, not a universal conversion for Operator levels. The explicit ratio approximates the source's Coarse 3 / Fine 17 setting.
Pass the result through butterlp at 5330 × 2^(0.6 × (MIDI note − 60) / 12) Hz, limited to 30–18000 Hz. This gives 60% pitch tracking around MIDI 60. It is a clean 12 dB/octave lowpass; the source's small resonance and Tone control are not modeled separately. No pitch sweep or LFO is added.
Sound after step 2 — WebM/Opus
Next: Round the transient with soft saturation →Step 03 / 5
Round the transient with soft saturation #
Set giStage init 3. Process the combined note bus through DC removal and a smooth saturation curve before the room. The shared bus lets overlapping notes drive the curve together.
aClean dcblock2 aDry
aDriven = tanh(aClean*3*ampdb(1.2))/3
aDry = (aClean*0.60 + aDriven*0.40)*ampdb(-1)
Drive is +1.2 dB and the processed share is 40%. The factor 3 sets the curve's sensitivity; division by 3 preserves small-signal gain before the drive boost. The −1 dB output trim balances the result. Harder strikes round more strongly while soft strikes stay close to a sine. This tanh curve interprets the gentle Soft Sine color; it does not implement Ableton's proprietary transfer curve.
Sound after step 3 — WebM/Opus
Next: Surround the strike with a short stereo room →Step 04 / 5
Surround the strike with a short stereo room #
Set giStage init 4. Delay the room input by 8 ms, then use the shared reverbsc output at 15% wet.
aPre delay aDry, 0.008
aRoomL, aRoomR reverbsc aPre, aPre, 0.68, 4500, sr, 0.35
aLeft = aDry*0.85 + aRoomL*0.15
aRight = aDry*0.85 + aRoomR*0.15
The predelay lets the strike arrive before the room. Feedback 0.68 produces a short room-like tail, while 4500 Hz damping darkens its upper frequencies. Delay modulation depth 0.35 keeps the tail from sounding completely static. The centered source feeds both reverb inputs; the reverb generates the stereo difference.
The Ableton room's 1.29-second decay, size and diffusion have no direct numerical equivalent here. Feedback and damping approximate its role around the short FM tone; they are not a claim of identical decay time.
Sound after step 4 — WebM/Opus
Next: Apply the final output trim and DC removal →Step 05 / 5
Apply the final output trim and DC removal #
Set giStage init 5 for the finished sound. Keep the room's stereo difference and trim both channels by −2.77 dB.
aLeft dcblock2 aLeft*ampdb(-2.77)
aRight dcblock2 aRight*ampdb(-2.77)
outs aLeft, aRight
clear gaVoice
DC removal follows the nonlinear stage and room. There is no bass-mono or extra width process in this sound. Clear the shared note bus after each control block so only the current note samples feed the next pass. The result has an initial bright knock, a quieter pitched body and a short stereo tail.
Sound after step 5 — WebM/Opus
Audio & settings ↓