Inverse tree generation (escape-time like)

  • 33 Replies
  • 1143 Views

0 Members and 1 Guest are viewing this topic.

Offline FractalDave

  • *
  • Uploader
  • *
  • Posts: 171
    • Makin Magic Fractals
« Reply #15 on: October 03, 2018, 02:01:18 PM »
A fix? for the single sequence method:
If you accumulate a divergent matrix instead of transforming the point i.e. transform the original point at each stage using the accumulated divergent matrix and once you've got d+1 transforms combined (where d is the required depth in the original tree) undo the first by applying the inverse of the first in the accumulation to the accumulation then just the single point, the accumulated matrix and the original transforms and their inverses are required (plus temporary stuff) - only the accumulated matrix is required from loop to loop all the rest can be essentially hard coded.

Edit:
Also a simple (untested) algorithm for creating a reasonable sequence is at each stage check which transform added next gives the highest number of possible (not already covered) sequences of the required depth with itself last, another afterwards last another afterwards last etc. - this algorithm is independent of the required depth and number of transforms as long as all possibilities are counted.
More: The algorithm for creating the shortest sequence could also be modified to allow for different branches having different depths (i.e. the required sequences to include not necessarily all having the same number of transforms) - of course the render loop would also need to be modified to allow for this, I think it could still be managed without adding arrays or more dynamic matrices.
« Last Edit: October 04, 2018, 05:35:35 PM by FractalDave »
The meaning and purpose of life is to give life purpose and meaning.

Offline msltoe

  • *
  • Fractal Fanatic
  • ***
  • Posts: 27
    • My cheesy music
« Reply #16 on: October 05, 2018, 03:40:30 AM »
Inverse tree generation in 3D is slow. Distance estimation might help.

Offline gerrit

  • *
  • 3f
  • ******
  • Posts: 1666
« Reply #17 on: October 05, 2018, 03:53:24 AM »
Nice one. You can almost feel the fuzz.

Offline Sabine62

  • *
  • Fractal Frankfurter
  • *
  • Posts: 624
  • It's just a jump to the left...
    • sabine62.deviantart.com
« Reply #18 on: October 05, 2018, 10:41:38 AM »
Ohh, this is Nice! :yes: :thumbs:
To thine own self be true

Offline FractalDave

  • *
  • Uploader
  • *
  • Posts: 171
    • Makin Magic Fractals
« Reply #19 on: October 05, 2018, 03:02:18 PM »
Timings for recursive escape-time method in UF - (entirely on CPU):
Images 715*954 straight from UF except addition of Coloursync.

Full view: Magnification *3.5, one thread: 17s, 23 threads 7s
Detail: Magnification *227.5, one thread 1m2s, 23 threads 20s

As implemented the basemodel hexagonal tube could be any primitive, in fact as written it's nowhere near optimum especially with respect to this.
Edit: That's because I was too lazy to code it up properly and did it just for proof of principle so the "basemodel" tube is actually an IFS itself - hence the rather slow render times...
« Last Edit: October 08, 2018, 01:39:44 PM by FractalDave »

Offline hobold

  • *
  • Fractal Fruit Salad
  • *****
  • Posts: 76
« Reply #20 on: October 05, 2018, 04:52:08 PM »
Inverse tree generation in 3D is slow. Distance estimation might help.
Thank you for doing the 3d experiment! The fuzziness is really cool; never before have I seen a 3D fractal carried out to such a high iteration depth. IMHO the overall impression of the object is significantly changed by that fuzz. Time to re-explore many old classics?

Offline Alef

  • *
  • Fractal Phenom
  • ****
  • Posts: 53
  • catalisator of fractals
    • My deviant art page
« Reply #21 on: October 05, 2018, 05:22:54 PM »
Personaly the most I like rotated tree. Bu I think the best would be a spruce/fir tree. In generation it is very simmilar to fern, once I got christmas tree by modifying fern ifs. It have branches curved in just opposite direction. This is darwinian adaptation to snow. On curved branches of spruces or birches snow don't stayes, so tree don't need to carry so mutch snow.
 
OK, now this 3D is far from artistic. But I think this thing is quite a perspective. People prefare natural patterns they see in life.  :thumbs:

Realy the last one seems to be close to christmas tree.
« Last Edit: October 05, 2018, 05:35:10 PM by Alef »
catalisator of fractals

Offline msltoe

  • *
  • Fractal Fanatic
  • ***
  • Posts: 27
    • My cheesy music
« Reply #22 on: October 06, 2018, 02:10:00 AM »
Thanks for the compliments, everyone!

Hobold: Since I have never used accurate distance estimation in my codes, my renderings tend to be more fuzzy. My renderer normally marches through space at roughly 0.1 voxels/step looking for the viewable plane. It then backtracks a few times to an accuracy of 0.025 voxels. For the most recent picture, I set the initial step size to 0.25 voxels, to save time.

I think fuzziness is found a lot in tree fractals. I was also able early on to get fuzziness using an increasing power Mandelbulb, i.e. first iteration is power 2, next power 4, next power 6, etc. Once again, having a very fine step size helps. 

One other point is that with tree fractals, it would be very challenging to get a lot of fuzz/hair with the forward method, because the number of primitives grows exponentially and would use up a lot of memory. The inverse variant is much more memory efficient only keeping track of the current recursion stack of visible search pointer. Maybe a hybrid of the two methods would strike the right balance of memory/speed?

Offline mclarekin

  • *
  • Fractal Frogurt
  • ******
  • Posts: 493
« Reply #23 on: October 06, 2018, 03:10:52 AM »
I like the fuzziness tree.

 the C code is at first glance, a  bit advanced for me.

 On my wish list is trees that can be animated to music, and also  trees that grow


Offline msltoe

  • *
  • Fractal Fanatic
  • ***
  • Posts: 27
    • My cheesy music
« Reply #24 on: October 06, 2018, 05:39:15 AM »
I made a "fern" with a little more color.

Offline gerrit

  • *
  • 3f
  • ******
  • Posts: 1666
« Reply #25 on: October 06, 2018, 05:42:05 AM »
I made a "fern" with a little more color.
Really nice. I'd like to have it in higher resolution though.

Offline Alef

  • *
  • Fractal Phenom
  • ****
  • Posts: 53
  • catalisator of fractals
    • My deviant art page
« Reply #26 on: October 06, 2018, 02:59:46 PM »
I made a "fern" with a little more color.
It's beautifull. It could be some trophical palms often grown in buildings (I don't know the name).

I think it could be extended to spruce trees just by curving branches in the opposite directions:

Offline msltoe

  • *
  • Fractal Fanatic
  • ***
  • Posts: 27
    • My cheesy music
« Reply #27 on: October 06, 2018, 05:16:09 PM »
Really nice. I'd like to have it in higher resolution though.
I'd like to get to higher res, too. It's actually a 1600x1600 render downsampled to 800x800 (+ crop) to get rid of some of the noise when the features get too small. The render took 2 hours on an 8-thread machine. So, a 3200x3200 with my current technique would take 16 hours.

Offline msltoe

  • *
  • Fractal Fanatic
  • ***
  • Posts: 27
    • My cheesy music
« Reply #28 on: October 07, 2018, 02:37:38 AM »
Fuzzy tree.

Offline gerrit

  • *
  • 3f
  • ******
  • Posts: 1666
« Reply #29 on: October 07, 2018, 06:01:01 AM »
Very nice. I want to eat it!


clip
Escape Time Blur

Started by AlexH on Share a fractal

0 Replies
298 Views
Last post August 15, 2018, 05:42:06 AM
by AlexH
xx
escape-time fractals (et) - instalation problem

Started by Adam Majewski on Fractal Programs Discussion, Help, & Support

2 Replies
396 Views
Last post May 31, 2018, 03:08:41 PM
by Adam Majewski
clip
Fourier series in escape time fractals

Started by v on Fractal Mathematics And New Theories

0 Replies
188 Views
Last post December 19, 2018, 07:32:09 PM
by v
xx
Julia sets: True shape and escape time

Started by marcm200 on Fractal Mathematics And New Theories

29 Replies
966 Views
Last post June 24, 2019, 10:24:25 PM
by marcm200
clip
3D M-set for tricomplex numbers: Escape time and shape change?

Started by marcm200 on Fractal Mathematics And New Theories

2 Replies
125 Views
Last post May 04, 2019, 09:58:37 PM
by marcm200