A Gauss kernel.
https://github.com/Jam3/glsl-fast-gaussian-blur
Constructs a new Gauss kernel.
The kernel size. Should be an odd number in the range [3, 1020].
Determines how many edge coefficients should be cut off for increased accuracy.
The offsets for linear sampling.
The weights for linear sampling.
The offsets for discrete sampling.
The weights for discrete sampling.
The number of steps for linear sampling.
The number of steps for discrete sampling.
A Gauss kernel.
See
https://github.com/Jam3/glsl-fast-gaussian-blur