3D IFS images and animation

[doncasteel8587]

My goal is to create an application to do in 3D what Apophysis and ElectricSheep do for 2D fractals. Please take a look at my gallery and let me know what you think.

Latest 3D animation: https://fractalraytracer3d.developer.network.com/files/documents/76/77/test.mpg
This one had some serious ray tracing issues, but if you look at the areas where it did render correctly, you'll get an idea of how it's supposed to look. (Mpeg4 so you may need mplayer to view it)

This one turned out better, but still needs work: https://fractalraytracer3d.developer.network.com/files/documents/76/68/TransformRotate003.wmv

Main gallery page with a wide variety of images: http://casteeld.deviantart.com/gallery/

Here are a few 3D pieces done in Maya (click on them to see the full size images):
http://www.deviantart.com/deviation/17393714/?qo=63&q=by%3Acasteeld&qh=sort%3Atime+-in%3Ascraps


http://www.deviantart.com/deviation/17072217/?qo=69&q=by%3Acasteeld&qh=sort%3Atime+-in%3Ascraps


http://www.deviantart.com/deviation/25963367/?qo=39&q=by%3Acasteeld&qh=sort%3Atime+-in%3Ascraps


http://www.deviantart.com/deviation/26059542/?qo=38&q=by%3Acasteeld&qh=sort%3Atime+-in%3Ascraps


http://www.deviantart.com/deviation/26153544/?qo=37&q=by%3Acasteeld&qh=sort%3Atime+-in%3Ascraps

[alan2here]

You have my full support. I hope you are sucsesfull.

I like making spikey 3d looking things in Chaoscope but I also like the Apophysis flame method.

I would love to use this software you are developing.

Verry nice looking with the graphics/shaders/lighting ect... also.

Please keep us informed

[matera]

That is an awesome idea, IFS in 3D - I think you are going to have a lot of people waiting impatiently for the first release! (I'm one of them)

[Jules Ruis]

See the first IFS in 3D, made in real wax.

www.fractal.org/Fractal-Research-and-Products/Tree-3D.htm

[alan2here]

See the first IFS in 3D, made in real wax.

www.fractal.org/Fractal-Research-and-Products/Tree-3D.htm

or... http://www.bathsheba.com/crystalsci/lorenz/
not IFS/Flame, but still verry impressive, link from chaoscope website.
However, it is expensive.

[Nahee_Enterprises]

Don Casteel wrote:
>
>    My goal is to create an application to do in 3D what
>    Apophysis and ElectricSheep do for 2D fractals. Please
>    take a look at my gallery and let me know what you think.

How much would it differ from the 3-D IFS images that XenoDream has been creating since it was developed a few years ago??

What file formats are you thinking of supporting for the 3-D outputs, so that they might be imported into 3-D rendering applications??

[doncasteel8587]

>How much would it differ from the 3-D IFS images that XenoDream has been creating since it was developed a few >years ago??

>What file formats are you thinking of supporting for the 3-D outputs, so that they might be imported into 3-D >rendering applications??
 
First and foremost, I can't spend the money for Xenodream, although I'd wanted to try it out for years.

Beyond that, from what I can tell by the images I've seen from XenoDream, it renders a surface which I assume is an IsoSurface "slice" from the fractal. This represents only a small fraction of the available information that can be had when creating a 3D IFS fractal.

My goal is to render the entire dataset with transparency and refraction so you can "see inside" the fractal. What I've been able to do is create the fractal as a "volumetric density field" which acts directly on the rays bending and coloring as they pass through it. Areas where the density/intensity are greatest are opaque, areas of lower density are transparent. Additionally as the ray passes through, the opacity is cumulitive so even in a low density regon, the farther the ray goes, the more opaque it appears.

The key is there is no surface, but there is illumination and shadow based on normals which always point directly from higher density to lowest density, shadows and highlights are produced by the angle of the normal to the light sources, just like surface illumination, but also by the opacity between the point being calculated and the light source.

It all works, and I've seen glimpses of what I wanted, there's just a lot of work yet to be done before it all works correctly every time.

I hope this helps explain where the project is going. I'm getting very very close to getting results that will really show off the full effect, and I'll share them as soon as possible.

Thanks for your interist
Don
 

[Nahee_Enterprises]

Don Casteel wrote:
>
>    My goal is to render the entire dataset with transparency
>    and refraction so you can "see inside" the fractal.   .......
>    I hope this helps explain where the project is going.

Yes, this definitely gives me a clearer picture of where you are heading with your efforts.  Thank you for the clarification.

Though I can see how there will be a definite difference between the saved image that has had light passing through it from various directions, and presumedly using different colored lights, and the actual 3-D object which could then be imported into another rendering application.

[doncasteel8587]

I'm sorry I forgot to answer the export question in my earlier response.

Since I'm not generating any kind of surface for the fractal itself, the only exports that can be done are pointclouds or 3D bitmaps (like MRI and CAT-scan data sets).

Some of my earlier work did output the volume as a series of 2D png files that when stacked produced a volume that you could load into Vtk and create the isosurface and export into a mesh file. The last three images in my original post were done that way.

Beyond that I doubt I'll try to provide other types of exporting.

[alan2here]

good luck nahee making anything that has lighting this good
http://tn3-1.deviantart.com/fs8/300W/i/2005/344/5/6/Itterated_Function_System_001_by_casteeld.jpg
or has internal structure like this
http://tn3-2.deviantart.com/fs6/300W/i/2005/109/8/f/Tinted_Glass_Fractal_Sculpture_by_casteeld.jpg
or has specular reflectiveness like this
http://tn3-1.deviantart.com/fs8/300W/i/2005/344/0/d/3D_fractal_polished_burl_wood_by_casteeld.jpg
or acuretly represents a flame fractal animation like this one
<a href="http://video.google.co.uk/googleplayer.swf?docId=3216455705264352366" target="_blank">http://video.google.co.uk/googleplayer.swf?docId=3216455705264352366</a>
up to the nerest pixel in 3 dimentions like this in xodream

Im not saying xodream is bad, for what ive seen it can do its extrodenary, but I think this porgram has potental to be much better for 3d flame fractal's

[Nahee_Enterprises]

Alan Tennant (alan2here) wrote:
>
>    good luck nahee making anything that has lighting this good
>    or has internal structure like this
>    or has specular reflectiveness like this

Thank you for wishing me "luck", but I have never really needed it.  I do quite well by relying on my knowledge, experience, and talent.     

Want to see some internal structure??
    http://www.Nahee.com/Fractals/XenoDream/Images/Untitled-34.jpg
    http://www.Nahee.com/Fractals/XenoDream/Images/Skewered_On_Spit-c.jpg

How about lighting??
    http://www.Nahee.com/Fractals/XenoDream/Images/Shapes-c.jpg
    http://www.Nahee.com/Fractals/XenoDream/Images/GT_Bristlecone1.jpg
    http://www.Nahee.com/Fractals/XenoDream/Images/Skewered_On_Spit-c.jpg
    http://www.Nahee.com/Fractals/XenoDream/Images/Octal-02b.jpg

And reflectiveness?? (Look real closely at the first one here.)
    http://www.Nahee.com/Fractals/XenoDream/Images/Spirals-Four-b.jpg
    http://www.Nahee.com/Fractals/XenoDream/Images/Quad-00c-2.jpg
    http://www.Nahee.com/Fractals/XenoDream/Images/Electric-Snail.jpg

>
>    Im not saying xodream is bad, for what ive seen it can do
>    its extrodenary, but I think this porgram has potental to be
>    much better for 3d flame fractal's

Personally, I hope each fractal software application strives to acheive results in its area better than any of the others.  In that way there is not only growth and new ideas coming forward, but competiton and multiple choices for the User.

[lycium]

Im not saying xodream is bad, for what ive seen it can do its extrodenary, but I think this porgram has potental to be much better for 3d flame fractal's

some keywords with small notes:

1. speed (cf. java vs c++, and gpus can do this volume tracing in realtime...)
2. scale (n^3 storage is costly, even with octrees, and fixes the resolution)
3. honesty (the best rendering features seen are due to maya, reproducing this standalone is difficult)
4. flexibility (xeno's holon system looks quite powerful)

And reflectiveness?? (Look real closely at the first one here.)

gotta say though, xeno's reflections are sooooooooo fake ;) don's work is infinitely more accurate in this regard (reflection / refraction) than xeno, simply because it's impossible with xeno's z-buffering based rendering. from that first image you showed, it seems as though xeno is just mapping 2 of the reflection vector's components directly to (u,v) of the rendered image - ie, it's not making any kind of spatial considerations and is strictly an image based technique.

[GFWorld]

Don Casteel - I took time today following your different links - let me answer here at last :
All my compliments for such great work ! 

Margit

[doncasteel8587]

Thank you very much, I'm glad you liked my work.
I'm really looking forward to getting all of the issues fixed so I can concentrate on the art side of it.

[titia]

....well, I'm very impressed too, Don. It is amazing work in my eyes.

Titia