Title: Mandelbulber 0.60 rendering problem Post by: MarkJayBee on June 02, 2010, 01:14:51 PM Hi guys,
I hate to bring this up - considering that I find Mandelbulber the best application for exploring the 3D fractal 'realm' - but I've begun to encounter a wee problem with some renderings. This takes the form of what look like dark 'ink blot' patterns in images rendered at various sizes, and using different settings; ambient occlusion on or off, varied perspective and iteration count raised/lowered. So far this only occurs when rendering deep zooms of Icosahedral K-IFS, but I will check out other settings. Obviously this is a little worrying when preparing an image which could take several hours to render! :sad1: Any thoughts, suggestions or - even better! - solutions would be gratefully recieved! Included 2 example images - 'blots' are on right of images: Title: Re: Mandelbulber 0.60 rendering problem Post by: Buddhi on June 02, 2010, 03:20:33 PM These "ink blots" are caused by limited accuracy of distance estimation (used for calculation ray-marching step and finding fractal surface).Because this istance is only estimated value there are possible some oversteppings. This could appear also for other fractal formulas. Solution is very simple. Please reduce a little "DE step factor" value. Standard value is 1.0 (raymarching step = DE_step_factor * estimated distance). Maybe better will be 0.8 or lower. Lower value reduces overstepping problem but increases rendering time.
Title: Re: Mandelbulber 0.60 rendering problem Post by: MarkJayBee on June 02, 2010, 09:55:53 PM Hi Buddhi, Many thanks for the advice. I did as you suggested, and changed the DE step factor to 0.8, and that seems to have done the trick! :D Now I just need to relocate the spot I was at when this started!! :dink: |