Title: optimal de-step factor Post by: Ike1970 on May 22, 2011, 06:39:33 PM Good evening everyone,
Does anyone have an explanation why the program is returning such strange de-values when using a mandelbox? When I use the new (great) function in mandelbulber to detect the best de-step factor, the results are allways extremely low when using a mandelbox (1e-03 and similar). But using such low values completely change the appearance of the image. thanks a lot, Ike Title: Re: optimal de-step factor Post by: Buddhi on May 22, 2011, 09:11:13 PM Could you attach settings file for which you get such a low DE step factors? I'm asking for it because in standard situations it gives values between 0.8 - 1.5
Title: Re: optimal de-step factor Post by: Ike1970 on May 23, 2011, 12:30:13 AM hi buddhi,
i´ve attached a .fract file (hopefully). On this one i have already let the software do the work - it came up with 0,0008. What I changed in the parameters where the alpha rotation (to 199) and folding limit (to 1,5). All else remained the same. take care and good night, Ike Title: Re: optimal de-step factor Post by: Buddhi on May 23, 2011, 06:17:37 PM hi buddhi, i´ve attached a .fract file (hopefully). On this one i have already let the software do the work - it came up with 0,0008. What I changed in the parameters where the alpha rotation (to 199) and folding limit (to 1,5). All else remained the same. take care and good night, Ike This problem with very low DE factor is because of high rotation in Mandelbox formula. This rotation causes that distance estimation (and ray-marching) is completely not accurate, not continues and not linear. So for achieve good quality of image without oversteppings there is needed very low DE factor. Title: Re: optimal de-step factor Post by: MarkJayBee on May 23, 2011, 07:38:33 PM On a related matter, buddhi; there may be a wee bug associated with the optimal DE
feature. On using it the resulting small test render was white. I discovered this was because the program had also input the - very high - DE figure into the 'glow' field?! This happened twice, so looks like there could be a glitch in there somewhere? :surprised: Title: Re: optimal de-step factor Post by: Buddhi on May 23, 2011, 09:55:10 PM On a related matter, buddhi; there may be a wee bug associated with the optimal DE feature. On using it the resulting small test render was white. I discovered this was because the program had also input the - very high - DE figure into the 'glow' field?! This happened twice, so looks like there could be a glitch in there somewhere? :surprised: It is my fault. It was not so good idea to connect glow intensity with DE factor. I did this because when DE factor is very low, then the intensity of glow sometimes is too high. I have to find some smarter way to connect these parameters. |