Welcome to Fractal Forums

Fractal Software => 3D Fractal Generation => Topic started by: lycium on October 23, 2007, 10:29:15 AM




Title: cipater
Post by: lycium on October 23, 2007, 10:29:15 AM
full 1920x1200 render: http://lyc.deviantart.com/art/cipater-68019331

preview:

(http://www.fractographer.com/wip/cipater_small.jpg)


Title: Re: cipater
Post by: twinbee on October 31, 2007, 03:55:04 PM
Love it - great work. It explores that wonderful zone between repetition of detail, and variety of shape. Also, in this one you can see slightly 'deeper' into each crater compared to your similar works.

As usual, I'd die to zoom into that, and/or fly along the surface. Do you know what would really be cool? If in Star Trek, they visited worlds like this. Not just single scenes, but their adventures would take place in these kind of surroundings. It'd make Star Trek worth watching for the gfx alone ;)


Title: Re: cipater
Post by: lycium on October 31, 2007, 06:32:12 PM
Love it - great work. It explores that wonderful zone between repetition of detail, and variety of shape. Also, in this one you can see slightly 'deeper' into each crater compared to your similar works.

thanks :) one issue with 3d fractals is that it can be difficult to find a good camera angle (it's also difficult for me because i have no realtime preview functionality, and have to edit the camera params in code). if i put the camera low, then the image becomes really flat but you get stronger depth of field effects; if i put it quite high up then the fractal becomes nearly 2d again.

after my exams are over i'll be re-writing my intersection routines to work with general linear ifs fractals. i expect it to be a lot faster, much more numerically robust, and above all a lot more interesting: the family of fractals with which i'm currently working are either completely flat or nearly spherical.

another member of fractalforums (and deviantart), david makin, has been working on this and has some really encouraging results! check out this fractal: http://makinmagic.deviantart.com/art/Julia-Rock-67422646

As usual, I'd die to zoom into that, and/or fly along the surface.

i hope you've viewed the 1920x1200 fullsize image at deviantart :) i'm currently rendering a 5120x3200 resolution image for a medium-sized print, and man is it taking forever... even with my recently upgraded computer (12ghz of cpu power) running for days, it's still not quite clean enough.

so, given how much it's punishing 12ghz just for a single image, i doubt it'll be realtime anytime soon ;) well, at least at this level of detail/quality; i've got full global illumination via metropolis light transport in there, as described by veach in his thesis: http://graphics.stanford.edu/papers/veach_thesis/. it should, however, be possible to do some realtime flying around without the fancy light transport, especially once i've worked out the new fractal algorithm.

stay tuned for cool things once my exams are finally over! :P


Title: Re: cipater
Post by: David Makin on October 31, 2007, 09:19:27 PM
I feel I should point out that the example of mine Thomas pointed you too is actually just a 2D fractal, though it is using escape-time IFS which is also how I'm doing my 3D IFS, I confess I haven't worked out how to accurately and efficiently mix non-linear transforms in with affine ones for 3D IFS yet.
Here are some of my current 3D IFS renders:

http://makinmagic.deviantart.com/art/Twin-Dragon-3D-anim-48328953 (hit the download button)
http://makinmagic.deviantart.com/art/The-Menger-Sponge-48085231
http://makinmagic.deviantart.com/art/Menger-Sponge-Zoom-48085608 (hit the download button)
http://makinmagic.deviantart.com/art/A-RIFS-reduced-cube-anim-48112383 (hit the download button)
http://makinmagic.deviantart.com/art/Sierpinski-Icosahedron-revisit-68160791

Note that only the last one of those is using my current algorithm (Hart's method of intesecting the transformed rays) - it's a lot faster and more accurate than what the others used (distance estimation).

You can find some more of my anims here:

http://www.youtube.com/profile?user=MakinMagicFractals

Though the 3D ones there as I write are using the old distance estimation method, I'm just aboiut to upload two vastly improved animations that use Hart's method :)



Title: Re: cipater
Post by: lycium on November 01, 2007, 09:09:42 AM
ah whoops, sorry about that; i thought it was 3d!


Title: Re: cipater
Post by: twinbee on November 01, 2007, 08:21:59 PM
Hehe, despite the mineral look, I had my suspicians ;-) Still, looks very nice!

For the others, like the Menger Sponge, and fractal icosa, I'd love to see them with more light sourcing, shadows, and especially with translucency materials and high indexes of refraction ;-))) Oh and some Photon mapping or true photon simluation wouldn't go amiss either >;-D


Title: Re: cipater
Post by: David Makin on November 01, 2007, 09:53:18 PM
I'm planning on shadows, reflection and trranslucency - not sure about photon mapping though :)
Actually I'm just about to add independant light sources and shadows.


Title: Re: cipater
Post by: lycium on November 01, 2007, 10:12:49 PM
For the others, like the Menger Sponge, and fractal icosa, I'd love to see them with more light sourcing, shadows

a recent render of mine using the icosahedron, full global illumination: http://lyc.deviantart.com/art/rise-64751071


Title: Re: cipater
Post by: twinbee on November 01, 2007, 11:44:53 PM
Yeah, I saw that before, and earlier today in fact. Looks so real. Love to see it with diamond-like refractivity and translucency! ;) (perhaps also try 3 different colour light sources, yellow, orange and red or something).

David, glad to hear you goin with the translucency!


Title: Re: cipater
Post by: David Makin on November 02, 2007, 12:35:07 PM
David, glad to hear you goin with the translucency!

Well I'm going to try to do it anyway. I think it's definitely possible but I'm not sure how efficient I'll get it in rendering speed terms :)