hi, welcome to the forums (there's quite an influx of new users since the 3d mandelbrot thing broke!)

those are quite interesting renders, did you precompute a volume of iterations, or use many small spheres, or ... ?

ahhh the mighty paul bourke strikes again! 

thanks for the link, interesting info

Have you ever thought about not computing a density map, but instead making it from small blobs? It would give a *huge* speedup, but it would require good finding of size & density of these blob elements.