UGen Changelog | SuperCollider 3.14.0 Help

This file documents changes affecting UGens in each SuperCollider version, starting from version 3.14.

Version 3.14

Summary of changes:

Delay1 and Delay2 accept new additional arguments: predelay sample values can be set explicitly
Added LevelComp pseudo-UGen, providing level compensation logic for multichannel mixing and panning
GVerb roomsize modulation was fixed;
Sweep rate modulation now affects next sample instead of current;
ToggleFF doesn't toggle randomly anymore when running at audio rate with a control-rate input;
EnvGen rounding error when calculating segment duration was fixed;
PanAz fixed server crash and numeric blow-up with negative width
Most of the other UGens listed here had their initialization sample calculation fixed, affecting their first samples of output;

Fixed initialization sample. No changes in output. Previously, calling AmpComp_next when arg rates were anything but all scalar rate caused state vars to be uninitialized and init sample would always be zero. Now init sample is always equal to first output sample.

Decay

Fixed initialization sample. First sample of input was incorrectly added twice, fixing it causes a change in the first samples of output:

Decay2

Fixed initialization sample. First sample of input was incorrectly added twice, fixing it causes a change in the first samples of output:

Delay1

Added new argument x1 for pre-delay (the value of first output sample). New behavior: when running at audio rate, x1 defaults to 0.0, i.e. "silence" precedes the delayed input. At control rate, the default value is set to the first sample of in, i.e. the first input sample is held during the delay.
Fixed initialization samples, changes first sample of output: a trigger on first sample wasn't delayed correctly, now it is.
Initialization sample used to be the input's value, now is fixed to reflect first sample of output (0)

Note: first sample can now be set by user

Delay2

Added new arguments x1 and x2 for explicitly setting predelay values. x2 is second sample predelay (first output sample value); x1 is first sample predelay (second output sample value). When running at audio rate, x1 and x2 default to 0.0, i.e. "silence" precedes the delayed input. At control rate, the default values are set to the first sample of in, i.e. the first input sample is held during the delay.
Initialization samples. Changes first sample of output: a trigger on first sample wasn't delayed correctly, now it is.

Note: first samples can now be set by user

EnvGen

EnvGen.kr: fixed accumulation of rounding errors causing segment of shorter durations than specified.
Fixed init sample calculation. Fixes initial phase output: it used to miss the first sample.

GVerb

GVerb seems more glitchy now than before! Are there any other reproducers?
maxroomsize has now a minimum of 1.

Example by @nhthn (https://github.com/supercollider/supercollider/issues/2302)

InRect

sclang now only accepts audio-rate xy inputs for InRect.ar

LFUGens

Fixed init sample calculation. Fixes initial phase output: it used to miss the first sample.

{ LFPulse.ar(SampleRate.ir/4) }.loadToFloatArray(2e-4, action:_.postln)
// 3.13: [1, 0, 0, 1, 1, 0, 0, 1]
// 3.14: [1, 1, 0, 0, 1, 1, 0, 0]
{ LFSaw.ar(SampleRate.ir/4) }.loadToFloatArray(2e-4, action:_.postln)
// 3.13: [0.5, -1.0, -0.5, 0.0, -1.0, -0.5, 0.0]
// 3.14: [0.0, 0.5, -1.0, -0.5, 0.0, -1.0, -0.5]
{ LFPar.ar(SampleRate.ir/5) }.loadToFloatArray(2e-4, action:_.postln)
// 3.13: [0.35, -0.84, -0.83, 0.36, 1.0, 0.35, -0.84]
// 3.14: [1.0, 0.35, -0.84, -0.83, 0.36, 1.0, 0.35,]
{ LFCub.ar(SampleRate.ir/5) }.loadToFloatArray(2e-4, action:_.postln)
// 3.13: [0.94, 0.56, -0.56, -0.94, 0.0, 0.94, 0.56, -0.56]
// 3.14: [0.0, 0.94, 0.56, -0.56, -0.94, 0.0, 0.94, 0.56]
{ LFTri.ar(SampleRate.ir/5) }.loadToFloatArray(2e-4, action:_.postln)
// 3.13: [0.80, 0.39, -0.40, -0.79, 5.96e-08, 0.80, 0.39, -0.40]
// 3.14: [0.0, 0.80, 0.39, -0.40, -0.79, 5.96e-08, 0.80, 0.39]
{ Line.ar(0, 1, 10/SampleRate.ir) }.loadToFloatArray(2e-4, action:_.postln)
// 3.13: [0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8]
// 3.14: [0.0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7]
{ XLine.ar(1, 2, 7/SampleRate.ir) }.loadToFloatArray(2e-4, action:_.postln)
// 3.13: [1.10, 1.21, 1.34, 1.48, 1.64, 1.81, 2.0, 2.0]
// 3.14: [1.0, 1.10, 1.21, 1.34, 1.48, 1.64, 1.81, 2.0]
{ Vibrato.kr(1, ControlRate.ir/4, 1,0,0,0,0) }.loadToFloatArray(0.01, action:_.postln)
// 3.13: [2, 1, 0, 1, 2, 1]
// 3.14: [1, 2, 1, 0, 1, 2]

PanAz

Fixed server crash and numeric blow-up with negative width (now negative width behaves exactly like positive).

Peak, PeakFollower

Fixed init sample when input is negative

Phasor

Fixed ignoring trigger on first sample, now Phasor starts from resetPos if triggered on first input sample.

PulseDivider

Fixed init sample calculation. First sample of output changed for trigger on first sample and div=1

SpecPcile

In case of a memory allocation fail, now the UGen outputs all zeros, instead of potentially garbage values.

Sweep

Fixed outputting 0 when triggered on first sample (it used to immediately increment to 1)

Fixed increment behavior: now modulating rate affects the next sample calculation (instead of the current, for which the previous value of rate is used).

T2A

Fixed init sample calculation, affecting first sample of output. Used to ignore triggers on first sample, now it outputs them correctly

TDelay

Fix delay of trigger on first sample. Output changes only if TDelay has a trigger on its first sample of input. If delayTime > dt (where dt is the time between first and second trigger), the output re-aligns with the previous version at the second trigger.

Timer

Fix init sample calculation, fixes the first measured duration.

ToggleFF

Fixed init sample to be the same as first output sample
Fixed random toggling when running at audio-rate with a control-rate input trigger

Trig, Trig1

Fixed init sample calculation, fixes initial phase output: it used to miss the first sample.

VarSaw

Fixed behavior for out-of-range initial phase: changes output if iphase is not within 0 and 1.

Fixed init sample calculation: was hardcoded to 0, now it's equal to first output sample.

helpfile source: /usr/local/share/SuperCollider/HelpSource/Guides/UGen-Changelog.schelp
link::Guides/UGen-Changelog::