Graphics and fractal programmer here

[Efoia]

Hi everyone, I'm a C++ software engineer with a focus in computer graphics.  I've been developing my own 3D fractal software for about 2 years now called Oak Fractal Sandbox.  I've gotten really useful information from these forums over the years so thank you all .  I currently don't have plans for a public release but I hope you all still find it interesting.  Hopefully I can at least be of some help to people attempting to write the same kind of software.

Here are a few renders:







Current features:

Appearance
    - lighting with multiple light sources and colors
    - shadows, soft and hard
    - ambient occlusion
    - fog
    - configurable colors
    - Hosek-Wilkie sky model
    - global illumination
    - image-based lighting
    - depth of field
    - geometric orbit traps
User interaction
    - realtime geometry manipulation
    - orbit camera and flight-sim camera controls, automatic distance-based camera speed adjustment
    - GUI which lets users modify geometry and appearance in realtime
    - highly data-driven and configurable (fractals, lights, colors, etc. all loaded from human-readable config files)
    - boolean operations on multiple solids
Rendering
    - super/sub-sampling (render resolution can vary independently from screen resolution)
    - high-resolution image and video rendering
    
Thanks for reading, I have more info on my website at http://www.davidbyrd.io/oak/.

[tit_toinou]

Great renders ! 

[quaz0r]

looks very nice, and it is also always nice to see C++ people endeavor to make new programs in modern C++.  i guess you intend for your program to be payware though?  unfortunately that lands it squarely in the "irrelevant" category.   

[Efoia]

looks very nice, and it is also always nice to see C++ people endeavor to make new programs in modern C++.  i guess you intend for your program to be payware though?  unfortunately that lands it squarely in the "irrelevant" category.   

No I'm not intending to sell it, I just don't consider it mature enough to release yet.  Though I suppose that still leaves it in the irrelevant category for most people! 

Also sorry to pull a switcheroo but I replaced the renders on my main post because I just saw how old they were, eep

[quaz0r]

No I'm not intending to sell it, I just don't consider it mature enough to release yet.

excellent!

Though I suppose that still leaves it in the irrelevant category for most people!

you and me both then 

Also sorry to pull a switcheroo but I replaced the renders on my main post because I just saw how old they were

very nice indeed 

[DarkBeam]

Oh fantastic the renders look really similar to those of Mb3d but you already added further interesting render features! Fantastic job please continue it!

[thargor6]

Super-awesome! I especially like the DOF-effect in the presented images.
And, the mixing of shadows with real backgrounds, I'm now thinking about to implement that feature in my software, thanks for the inspiration!

Cheers!

[lycium]

Nice renders, and always interesting to see more "rendering people" getting into fractals

i guess you intend for your program to be payware though?  unfortunately that lands it squarely in the "irrelevant" category.  

Friendly as always...

[mclarekin]

I like the shadow on to the background image.  I have been wanting one of them

[ciric50]

Love that first image in particular. Can you tell me what kind of fractal that is and how you did it?

[Efoia]

Love that first image in particular. Can you tell me what kind of fractal that is and how you did it?

That's the Pseudo-Kleinian as described by knighty in this post (his fragment shader with the distance equation is attached to the post).  If I remember correctly, all I did was increase the TThickness parameter which turns the default perfect spheres into the shape you see in the image.  The fractal normally repeats in all three axes, but I added subtractive half-space volumes to cut it down to a single repetition.  As for the lighting, it's globally illuminated via monte-carlo path tracing.  Quite a cool fractal .

[ciric50]

Thanks! The lighting in your image is really nice.

[SCORPION]

I want to try it!

[M Benesi]

Nice shadows.  Are you using 3d images, or just adjusting where the shadow falls with a floor function?

  Looks 3d in the Julia in this thread, but on your page the Mandelbulb shadow looks smooth (like the shadow is applied to a texture on the "floor" of the scene).  

  I've only had half a coffee so far this morning, so I don't even know if my eyes are working correctly yet.


   I'm thinking about smoothly varying the toroidal and prismatic orbit traps (make (1-a%) toroidal + a% prismatic)....

  Are you using the traps for DE (distance estimation) for those images on your page, or are you doing brute force raymarching (boolean- a pixel is either in or out)? 


  Ohh... Hi!   Cool software! 
  

[Efoia]

Nice shadows.  Are you using 3d images, or just adjusting where the shadow falls with a floor function?

  Looks 3d in the Julia in this thread, but on your page the Mandelbulb shadow looks smooth (like the shadow is applied to a texture on the "floor" of the scene).

The images are 2D equirectangular projections, so they show up as a full panorama when I map rd (ray direction) to uv coordinate.  But the shadows are cast onto a perfectly flat floor plane so they aren't accurate to the actual image (but accurate enough if the ground in the image is flat!).

  Are you using the traps for DE (distance estimation) for those images on your page, or are you doing brute force raymarching (boolean- a pixel is either in or out)?

The geometric orbit traps are calculated as part of the DE, at each iteration I keep track like so:

Code:

distance = min(distance, trap(pos) + <more math>)

  Ohh... Hi!   Cool software! 

Thanks!