|
![](/i/fill.gif) |
Nathan Kopp wrote:
>
> The mapping is (if I did the math correctly):
>
> old_value = (new_value / 0.48985582) - 1.05242 + 0.5
>
> -Nathan Kopp
I wonder, whether these values are correct too? I followed this
"Wondering about Noise()" thread in programming group and as I
understand, they come from Xander's noise test, where he took several
million samples over a box. But there is no guarantee, that there does
not exist points, where noise value could exceed this range <-1.05242,
0.988997>. So I guess that better would be to just add 1, divide by 2
and clamp into range 0-1. It is easier to understand, as there is no
"Tambov constants" (as we call here constants, which should be added to
value in order to get correct result) and more logical. I guess that
little difference between this approach and current implementation in
unnoticeable.
Post a reply to this message
|
![](/i/fill.gif) |