Often times Blur devices are very expensive for me (I probably use too much of them with too high radii anyway), even if i’m only using them masked off on small parts of the terrain.
Would it be a good idea to check if the mask value is zero before calculating the gaussian blur of a pixel? This could be applied to other slow devices as well.
EDIT: it just occured to me that gaussian blur seems to be a 2 pass algorithm, (cause its ordo linear with kernel size), so implementing this might not be straightforward.