It seems .with_seed() has the surprisingly low period of 8 for most noise with d <= 2.
The below equalities holds true for (at least) all n < 10000 and dim.width < 1000:
I did not find any such periods in cellular2_2d.
I haven't spent much time looking at d >= 3 noise, but also there I didn't encounter any obvious periods.
It seems
.with_seed()
has the surprisingly low period of 8 for most noise withd <= 2
. The below equalities holds true for (at least) alln < 10000
anddim.width < 1000
:gradient_1d.with_seed(n) == -gradient_1d.with_seed(n+8)
gradient_2d.with_seed(n) == gradient_2d.with_seed(n+8)
fbm_1d.with_seed(n) == -fbm_1d.with_seed(n+8)
fbm_2d.with_seed(n) == fbm_2d.with_seed(n+8)
ridge_1d.with_seed(n) == ridge_1d.with_seed(n+8)
ridge_2d.with_seed(n) == ridge_2d.with_seed(n+8)
turbulence_1d.with_seed(n) == turbulence_1d.with_seed(n+8)
turbulence_2d.with_seed(n) == turbulence_2d.with_seed(n+8)
I did not find any such periods in
cellular2_2d
. I haven't spent much time looking atd >= 3
noise, but also there I didn't encounter any obvious periods.