Are Crackle and Dust repetitive?

Questions
1

I’m using Crackle + Dust in multichannel expansion. Today after several minutes a repetitive patterns appeared, with a loop length of approx. 1 or 2 seconds. I can’t track down the issue, and can’t reproduce the phenomenon.

I can’t say if it was one channel after another that acted like a multitap delay or if the repetition happened on each individual output.

Is it possible that the seed is distributed over the iterations in a way that each iteration does the same thing with a short delay.

I suspect Crackle, it’s the dominant part in this code.

I had no drugs this morning, I slept well and had breakfast and coffee. Maybe my brain fouled me anyway, am I getting nuts?

What are Crackle and Dust doing under the hood, how do they get their seeds?

(
play {{
HPF.ar(
		
		Crackle.ar(2,LFNoise1.kr(0.3,0.2,0.3))

				+RLPF.ar(Dust2.ar(LFNoise2.kr(0.2,0.9,1).cubed*20,LFNoise1.kr(1,0.4,0.6)),
					LFNoise1.kr(0.03,2000,3000),LFNoise1.kr(0.01,0.2,0.9)),
                    400,0.2)

}!8
};
)
2

Crackle is a chaotic function: https://github.com/supercollider/supercollider/blob/608bb981162c2c26f0a32c09d82557b29774a32e/server/plugins/NoiseUGens.cpp#L379 – it’s seeded by a random number once, but after that, the random number generator isn’t used at all in the main calculation loop:

LOOP1(inNumSamples, ZXP(out) = y0 = fabs(y1 * paramf - y2 - 0.05f); y2 = y1; y1 = y0;);

Chaotic functions do often show some self-similarity.

Dust2 applies a threshold to a random number generator. That “should” be less likely to give an impression of repeating.

What does it sound like if there’s only one instance instead of 8? (Also, are you sure you’re hearing 8 channels? I don’t see that they’re mixed down in any way.)

hjh