question, mostly for twinbee

[gussetCrimp]

Congratulations to everyone for these beautiful pictures, and to twinbee in particular for the attention this has gotten around the web and in New Scientist. I'm sure there are lots of lurkers, like me, who come back to the forums every day hungry for a new look at these amazing objects. It's a really exciting discovery - the 3D-ness of these beasts makes it all the easier to imagine actually encountering them, floating in space or hovering above the landscape or buried under water for millions of years, like in some of the interpretations we've seen so far. I'm a mathy person but unfortunately I've never done anything sophisticated with computer graphics or I'd be spending all my time exploring them myself. As it is I just keep coming back here to see more pictures, and it's gotten to the point where if I see especially ornate textures in real life - on an old building for example, or yesterday in my pasta and broccoli - I immediately think Mandelbulb. Favourites so far: twinbee's renderings at the skytopia site, iq's video on the old thread (and wow I saw the new one just before posting, and Cyril Crassin's engine looks extremely exciting too), David Makin's Garden of Many Hues, cbuchner1's golden bots, and enforcer's metallic renderings.          
                                                                                
So this post is mostly to offer kudos and to let you all know there are surely many more people reading than there are contributing.
                                                                          
Two questions for twinbee: I think of the image with the caption "And here is the beast itself" as the "iconic" image of the mandelbulb. It looks radially symmetrical and also at least roughly symmetrical around its equator. It's hard to see how the object that is captioned "Here's the whole thing, with some perspective this time" could be the same object! It is not radially symmetrical or north-south symmetrical, and it has "flaps" extending outwards at the top and bottom that I see no evidence of in the iconic image. Could you make a couple of elevations from other angles so we could get a better overall feel for the shape of the thing? The ideal would be to make a little applet with pan controls over the view of a low-detail rendition, so we could roll the object along different axes. Another great thing along these lines would be to make a quicktime VR spherical panorama of the inside as rendered by cbuchner1 (I'm referring to the images "in the bowels" and "chocolate tunnel", whose titles don't quite sound fit for a family forum).

And the "big picture" question for Daniel: what is going on now? From the intensity I see in your writings about this discovery, and the promise that the best is yet to come, I can't believe you're sitting back right now. Are you feverishly rendering an entire atlas of the mandelbulb, to be ready by Christmas time? Are you in high-level talks with Pixar to produce a gorgeously textured animated exploration for IMAX screens? I want to complete a few orbits around a slowly tumbling mandelbulb then arc in gracefully, getting closer and closer until our point of view is dwarfed by surface structures rising like canyon walls around us... There is surprisingly little activity on the forum, especially from you. Could you give us a clue about what is going on and when the world can expect to see new wonders?

And just so this post is not an entirely wordy waste of everyone's time, I've made a parallel-view version of enforcer's stereo view, because I can't do the cross-eyed version and I suspect others can't either.

[cbuchner1]

And just so this post is not an entirely wordy waste of everyone's time, I've made a parallel-view version of cbuchner1's stereo view, because I can't do the cross-eyed version and I suspect others can't either.

I am the guy with the golden bots, not the one with the silver-metallic T1000's So above image is miscredited, it belongs to Enforcer who has just posted a binary also.

The above image was created with a HLSL shader, which I find pretty exciting - because it does not depend on a particular hardware vendor's API (but Microsoft's DirectX10 unfortunately). I wonder how the normals were computed. The HLSL code posted in the forum was only for the raymarching part.

[Enforcer]

The above image was created with a HLSL shader, which I find pretty exciting - because it does not depend on a particular hardware vendor's API (but Microsoft's DirectX10 unfortunately). I wonder how the normals were computed. The HLSL code posted in the forum was only for the raymarching part.

Thanks. Actually i played with your OptiX version, which i found exciting too, (unbelievable, but my patch appeared here  , i didnt search for it!)  when i decided to make my own.

Normals are computed using DE of 4 points, fast version with ddx/ddy also included:

Code:

inline float3 CalcNorm(float3 t, float c)
{
   float delta=4.0/25600.0;
   float3 tx1 = t;
   tx1.x+=delta;
   float cx1 = Tex( tx1 );
   float3 ty1 = t;
   ty1.y+=delta;
   float cy1 = Tex( ty1 );
   float3 tz1 = t;
   tz1.z+=delta;
   float cz1 = Tex( tz1 );
   float3 d1 = float3(c-cx1,c-cy1,c-cz1);
   return normalize(d1);//*25600;
}

inline float3 CalcNormDD(float3 t, float c)
{
   float3 n1=ddx(t);
   float3 n2=ddy(t);
   return normalize(cross(n1,n2));
}

[gussetCrimp]

I am the guy with the golden bots, not the one with the silver-metallic T1000's So above image is miscredited, it belongs to Enforcer who has just posted a binary also.

thanks, I corrected the attribution in my post.

[twinbee]

gussetCrimp, hi! Indeed, I have been slaving over these past few days. At the moment, I'm rendering a deepish zoom into the object. But it's not just any old zoom - I'm actually moving the camera inside, so everything goes past in amazing perspective with parallax. It's an utter pig to render, and the resolution leaves a lot to be desired, but should still be worth it

I want to complete a few orbits around a slowly tumbling mandelbulb then arc in gracefully, getting closer and closer until our point of view is dwarfed by surface structures rising like canyon walls around us

Yes, that was next on the agenda (yes, weirdly enough I thought of that too). To see giant towering structures whiz by at speed when the camera is in close would surely be too spectacular to contemplate

The other thing I'm working on is creating a 2000x2000x2000 voxel representation (had about 4 (!) contacts from people who want to build a real-life object of this thing, mostly for non-profit afaik). I intend to make an animation of that too, in the style of the mountain cross sections video on Youtube. But giant 2000x2000 PNGs instead.

Finally, I've spent a lot of time fixing bugs in the renderer, and tidying up the code. I want to add some colouring techniques which I think will look really nice, but polishing up the code takes priority unfortunately.

Contacted by Pixar? Nope. I had a few new sources contact me and a call from Nature about using one of my pics. Oh it was on the Discovery science news site as well.

It's hard to see how the object that is captioned "Here's the whole thing, with some perspective this time" could be the same object!

Ah but it is That's what perspective does - it distorts the shape, but our senses are accustomed to this distortion with familiar objects (particularly with perpendicular angles). Here's 3 renders. Each look the same size, but in fact, the camera is fairly close to the object on the last one, and very far in the first: