Twistbox - spiralling -1 scale box variant

[pupukuusikko]

I found an interesting addition to the already versatile scale -1 box scene.

Formula is very simple, normal -1 box with swapping of the signs of x and z (inspired by josleys's kleinian and trafassels following work).
Still, the effect with suitable folding (fold>1) and julia coordinates is quite unique, with overwhelmingly abundant spirals
The code is below, full frag is also attached, though DE would need some tuning.

Code:

float  minibox(vec3 z)
{
	vec3 pos = z;
	float DF = 1.0;
	for (int i = 0; i < Box_Iterations ; i++) 
	{
		z.xyz = clamp(z.xyz, -fold, fold)*2. - z.xyz;
		float r =dot2(z);
		z/=-r; 
		DF/= r;
		z.xz*=-1.;
		z+=(Julia ? JuliaC : pos);
		orbitTrap = min(orbitTrap, abs(vec4(z,dot(z,z))));
	}	
	float r = length(z); //(dot2(z));
	return 0.2*sqrt(r)/DF;
}

[DarkBeam]

Tried already the dwsel crazy mod folds?
I moved the topic to the appropriate section too
http://www.fractalforums.com/amazing-box-amazing-surf-and-variations/kaliset-plus-boxfold-nice-new-2d-fractal/15/

[Crist-JRoger]

we have hyperactive fractal developer for Frag   

[pupukuusikko]

we have hyperactive fractal developer for Frag   

Nah, just some childish excitement of a noob, it will pass quickly

@luca: thanks for the relevant link, I might explore the foldings some time, but
the current formula seems nice and continuous with the default boxfold too.

[DarkBeam]

please look at the kleinian other topics. I cannot remove the separation line...
Do you know some workaround?
Should I add a divergent bailout? ...

[DarkBeam]

Code:

float  minibox(vec3 z)
{
	vec3 pos = z;
	float DF = 1.0;
	for (int i = 0; i < Box_Iterations ; i++) 
	{
		z.xyz = clamp(z.xyz, -fold, fold)*2. - z.xyz;
		float r =dot2(z);
		z/=-r; 
		DF/= r;
		z.xz*=-1.;
		z+=(Julia ? JuliaC : pos);
		orbitTrap = min(orbitTrap, abs(vec4(z,dot(z,z))));
	}	
	float r = length(z); //(dot2(z));
	return 0.2*sqrt(r)/DF;
}

Also here you definitely should check if R is > 1e20 or < 1e-20 to avoid numerical trouble...

[Sabine]

Love this formula, great fun, Pupukuusikko, thank you!

[Crist-JRoger]

Super! 

[Crist-JRoger]

very slow 

[DarkBeam]

Correct me if I am wrong. This is Buddhi's Amazing box, with a rotation of (180, 180, 0). In MB3D you should get it that way (use ABoxPlatinum).
Johan/Kraftwerk - try plsss

[pupukuusikko]

Correct me if I am wrong. This is Buddhi's Amazing box, with a rotation of (180, 180, 0). In MB3D you should get it that way (use ABoxPlatinum).
Johan/Kraftwerk - try plsss

I tried it without success. There is probably a difference in ballfold part?

my formula has just unconditional p.xyz/=r, whereas abox in mb3d has

if rr < sqr(Min_R) then
   m = Scale/sqr(Min_R)
else if rr < 1 then
   m = Scale/rr
else m = Scale
..
p.xyz*=m

[DarkBeam]

Sorry I did not notice

[mclarekin]

My two cents. It looks like buddhis mandebox with a variation to the DE calc. 

In the Mandelbox, the Fixed Radius and Minimum Radius  (now called maxR2 and minR2) are variables.  We  set  maxR2 to very big number and minR2 to zero  and it is pseudo unconditional. ( i think??)