prog.txt

below is the render function of the original, pre-optimised, version.
only main change is the option to do a two-pass which first applies the filter normally, then applies the filter rotated through 90 degrees, then adds the two results.

#define data(xx,yy) framebuffer[c[yy]+min(max(i+xx,0),lastcol)]
#define scale(out) min(max((long)((out+farray[25])/farray[26]),0),255)
int C_FILTER::render(char visdata[2][2][576], int isBeat, int *framebuffer, int *fbout, int w, int h)
{
	int c[5], k, l = 0;
	int lastcol = w-1;
	int lastrow = w*(h-1);
	long outR, outG, outB, dta;
	if (!enabled) return 0;
	c[0]=0;c[1]=0;c[2]=0;c[3]=w;c[4]=2*w;
	for(int j=0;j<h;j++)
	{
		for(int i=0;i<w;i++)
		{
			k = 0;
			outR = 0;
			outG = 0;
			outB = 0;
			for(int y=0;y<5;y++)
			{
				for(int x=-2;x<3;x++)
				{
					dta = data(x,y);
					outR += farray[k]*R(dta);
					outG += farray[k]*G(dta);
					outB += farray[k]*B(dta);
					k++;
				}
			}
			fbout[l] = RGB(scale(outR),scale(outG),scale(outB));
			l++;
		}
		c[0]=c[1];c[1]=c[2];c[2]=c[3];c[3]=c[4];c[4]+=w;c[4]=min(c[4],lastrow);
	}
	return 1;
}