Trying to replicate a fractal

  • 23 Replies
  • 574 Views

0 Members and 1 Guest are viewing this topic.

Offline mclarekin

  • *
  • Fractal Frankfurter
  • *
  • Posts: 628
« Reply #15 on: December 07, 2019, 12:01:32 AM »
well you got symmetry and an apollonian:)

it looks like you need a lot more points to get more detail

My previous image looks  a lot difference due to an effect caused by distance estimation (the spheres), this next image is more like what you should get when using a large amount of voxels


Offline UnsoundMethods

  • *
  • Fractal Friend
  • **
  • Posts: 12
« Reply #16 on: December 07, 2019, 10:17:22 AM »
Ah, yes that's more like it. 
I used a 1 giga voxel grid for the previous image.  That should be enough for a 1k image. 

What are the amount of iterations you use, and what is the bailout set to?
I can try this in a shader, but my goal is to get geometry...

Computer is calculating a sequence with some offsets to see what it gives me, but a busy weekend ahead so I'll post something new on Monday.

Offline mclarekin

  • *
  • Fractal Frankfurter
  • *
  • Posts: 628
« Reply #17 on: December 08, 2019, 01:59:58 AM »
i am using openCL with GPU so it only takes 3 seconds, hence default settings of 250 iterations and length>bailout of 100

but at this camera to object distance i cannot see any visual difference with  13 iterations and a bailout of 10

Offline UnsoundMethods

  • *
  • Fractal Friend
  • **
  • Posts: 12
« Reply #18 on: December 09, 2019, 07:29:21 PM »
So its the number of iterations that give me the detail, but creating the SDF becomes precision work.
But I now have renders very similar.  I get most of it to generate in 48GB or RAM, though the 1.5 Gigavoxel at 10 iterations is going up to 64 GB at the moment. I'll cluster the fractal on multiple machines to make it more efficient.

Now "just" need to find the right offsets etc to get to my fractal :-)

Offline UnsoundMethods

  • *
  • Fractal Friend
  • **
  • Posts: 12
« Reply #19 on: December 09, 2019, 07:31:31 PM »
Bummer posted pic twice, and no idea how to remove the duplicate.
Meant to add the animated gif

Offline mclarekin

  • *
  • Fractal Frankfurter
  • *
  • Posts: 628
« Reply #20 on: December 10, 2019, 01:52:44 AM »
Quote
Now "just" need to find the right offsets etc to get to my fractal :-)

I suspect it is more than that. I could be wrong but it looks like there is some additional maths involved

Offline UnsoundMethods

  • *
  • Fractal Friend
  • **
  • Posts: 12
« Reply #21 on: December 10, 2019, 01:35:07 PM »
Yeah, hence my wink.

I'll need to get back in touch with the creator.

Offline 3DickUlus

  • *
  • 3f
  • ******
  • Posts: 1448
    • Digilantism
« Reply #22 on: December 11, 2019, 04:42:41 AM »
Bummer posted pic twice, and no idea how to remove the duplicate.
Meant to add the animated gif

Select Modify from the message header (first line)
Under Attached:
    Uncheck the attachments you no longer want attached:
and Save the changes
Fragmentarium is not a toy, it is a very versatile tool that can be used to make toys ;)

https://en.wikibooks.org/wiki/Fractals/fragmentarium

Offline UnsoundMethods

  • *
  • Fractal Friend
  • **
  • Posts: 12
« Reply #23 on: January 17, 2020, 03:12:52 PM »
So, after a bit of a hiatus;

Offsetting "C" in the pseudokleinian  along an axis before sphere inversion is getting it closer.