trafassel
Fractal Bachius
Posts: 531
|
|
« on: February 15, 2010, 09:48:51 PM » |
|
Pow 8 Mandelbulb in 1000 Iterations and in 10000 Iterations.
|
|
|
Logged
|
|
|
|
hobold
Fractal Bachius
Posts: 573
|
|
« Reply #1 on: February 16, 2010, 01:52:32 AM » |
|
What kind of rendering method was used? I mean, did the renderer in question actually iterate to the limit? Because I am unable to see any difference.
|
|
|
Logged
|
|
|
|
kram1032
|
|
« Reply #2 on: February 16, 2010, 02:34:53 PM » |
|
There are differences: Some rather random-looking white spots (not single pixels, roughtly 10*10 pixel groups) are brighter in one than in the other. Hardly any diference, still.
|
|
|
Logged
|
|
|
|
KRAFTWERK
|
|
« Reply #3 on: February 16, 2010, 02:47:22 PM » |
|
The difference: THERE IS AN IMAGE BELOW!!!
|
|
|
|
trafassel
Fractal Bachius
Posts: 531
|
|
« Reply #4 on: February 16, 2010, 07:54:46 PM » |
|
Sorry, wrong pictures. Both pictures are rendered with 1000 Iterations. The first one uses a better Deph-Estimation (the resolution of the deph ray) and is also rendered with 1000 Iterations.
The following pictures are renderings of the power 5 mandelbulb with 300 and 1000 iterations.
|
|
|
Logged
|
|
|
|
kram1032
|
|
« Reply #5 on: February 16, 2010, 08:00:26 PM » |
|
well, the difference is greater now but it seems that most details (in both cases) are lost in jpg-artifacts...
|
|
|
Logged
|
|
|
|
hobold
Fractal Bachius
Posts: 573
|
|
« Reply #6 on: February 16, 2010, 08:36:33 PM » |
|
The good news is, I probably don't need to waste so much time iterating all the way to 1000. Hm ... how to quantify how many iterations are "needed"? Did anybody ever make histograms of how many voxels iterated how often before bailout?
|
|
|
Logged
|
|
|
|
trafassel
Fractal Bachius
Posts: 531
|
|
« Reply #7 on: February 16, 2010, 10:33:29 PM » |
|
|
|
|
Logged
|
|
|
|
trafassel
Fractal Bachius
Posts: 531
|
|
« Reply #8 on: February 17, 2010, 09:08:02 PM » |
|
And here ist the power 8 Mandelbulb in 10000 iterations. I was changing my rendering algorithm, so you cannot compare it directly with the first pictures.
|
|
|
Logged
|
|
|
|
trafassel
Fractal Bachius
Posts: 531
|
|
« Reply #9 on: February 18, 2010, 08:51:09 PM » |
|
Mandelbulb zoom of the upper right part with 1000 and 10000 iterations.
|
|
|
Logged
|
|
|
|
hobold
Fractal Bachius
Posts: 573
|
|
« Reply #10 on: February 18, 2010, 11:40:59 PM » |
|
Now I can make out a difference in the lower right corner of the images. There are hints of an extra "leaf" being uncovered as the dust recedes. I would not have expected such distinguished structures to be revealed that late in the iteration process. After all, the solid interior seems to have well defined boundaries after just 100 iterations.
This could mean that the available rendering algorithms so far have not represented the "true" fractal shape as well as we think they have (including my brute force volume sampler). Or it could mean that all this dust everywhere isn't really dust, but does occupy a nonzero volume somehow.
|
|
|
Logged
|
|
|
|
KRAFTWERK
|
|
« Reply #11 on: February 19, 2010, 08:57:22 AM » |
|
|
|
|
Logged
|
|
|
|
|