Welcome to Fractal Forums

Fractal Software => Mandelbulb 3d => Topic started by: Jesse on February 25, 2011, 10:03:06 PM




Title: InvSphericalC formula
Post by: Jesse on February 25, 2011, 10:03:06 PM
Just a simple one that also inverts the adding constant for a full coordinate inversion.
It shows up in the '3D' folder, because that disables the analytical function of '3Da' formulas, that would not work anymore.



Title: Re: InvSphericalC formula
Post by: DarkBeam on February 26, 2011, 08:15:57 AM
Jesse it is not needed to make this transform. I already did great images with inversion of IFS formulas, with the pre transform trick (invspherical 1 iter and ifs 999 iters)
Instead you can do the pre and post transform option :)

Another question!

It would be useful if you store the 2 old values of xyzw vector.
I may use it for Phoenix formula, function of those variables
It is easy make this; initially Sold = C, Sbak = 0.
After all formulas cycle, update Sbak with Sold values, and Sold with S. S is xyzw vector :)


Title: Re: InvSphericalC formula
Post by: Jesse on February 26, 2011, 09:40:25 AM
Jesse it is not needed to make this transform. I already did great images with inversion of IFS formulas, with the pre transform trick (invspherical 1 iter and ifs 999 iters)
Instead you can do the pre and post transform option :)

It only makes no difference on formulas that are not using the C component, like the IFS's.   :dink:

Use a bulb or the amazing box for a comparison.

Quote
Another question!

It would be useful if you store the 2 old values of xyzw vector.
I may use it for Phoenix formula, function of those variables
It is easy make this; initially Sold = C, Sbak = 0.
After all formulas cycle, update Sbak with Sold values, and Sold with S. S is xyzw vector :)

It is not that easy because i have 13 iteration loop functions that must be changed  :dink:


Title: Re: InvSphericalC formula
Post by: DarkBeam on February 26, 2011, 09:50:58 AM
Jesse it is not needed to make this transform. I already did great images with inversion of IFS formulas, with the pre transform trick (invspherical 1 iter and ifs 999 iters)
Instead you can do the pre and post transform option :)

It only makes no difference on formulas that are not using the C component, like the IFS's.   :dink:

Use a bulb or the amazing box for a comparison.

Well, the existing function seemed to work fine with everything but not tested every combination ^-^

Quote
Another question!

It would be useful if you store the 2 old values of xyzw vector.
I may use it for Phoenix formula, function of those variables
It is easy make this; initially Sold = C, Sbak = 0.
After all formulas cycle, update Sbak with Sold values, and Sold with S. S is xyzw vector :)

It is not that easy because i have 13 iteration loop functions that must be changed  :dink:
:o
But Phoenix fractals are so damn pretty :embarrass:


Title: Re: InvSphericalC formula
Post by: DarkBeam on February 26, 2011, 06:59:50 PM
I fear there is a trouble! :-\ :-\ :-\

Mandelbulb3Dv16{
M.....S....O/...m....2....Ut31n3Ajx/.bc1Q359/X2E7D/UBQDTbw1GvVr1rnr7zsxKU7t0gmzD
................................fUWHY3moez1........Y./..................y.2.....
................/ME//....6k8....9/....E2.....kThOYXTuKrD/..........c./...w1/...U
z.....kD..../2.dkpX0LD8wz.............................................sD...../..
.w1...sDQV/Grsbipx1..........qGqaSpmA5ejiAQ9CcR8Br9LMUoBydPRzMAacMpmA5ej6X78KhAn
Vu1ANPuJ9nQcygfSpJXTuKrD......o.......................sD..kz....................
.............................UJRR4.wppN.6OLb/.mRR4.irpN.EVLb/UCSR4..............
...........................U..6.P....M4...EB....W....61....F....8/...I1....UJl12
...U.iVFwxDE./ozPM2Tzz7lz16.mc..zXCc.El18XGQeGyDjvIRhrVAkz1............28.kFrA0.
FWb96aAIVzvh1se7Umvxz0........../6U0.wzzz1................................E.0c..
zzzz.................................2U.8.kzzzD.................................
/6U0.wzzz1...................................2CcN/UvNP6.eeWCNq0.yRii.EJJUk1f..XR
S1.mx3CcN/UvNP6.QsLsUa3.ibhV..bTV1OK.sSq40.ly3CcN/UvNP6.MwLsUa3.ibhV.kqTV1OK.sSq
40.kz3CcN/UvNP6...EsUa3.eeWCNq0.IJ36wk8.wyLsUa3.................................
E....6....E.....I.........kLH/5OZ7LN7taR1/UQ.........................c..........
...................6./........zj................................................
................................................................................
.....................E4...E/....3....6IRg7qPs/.B2B4.Z75.........................
.Qk07Y......................../E........Uz1........wz.........zD..........2.....
................................................................................
................................}


Title: Re: InvSphericalC formula
Post by: Jesse on February 26, 2011, 07:37:03 PM
That is a good point:  this formula does not work in Julia mode!!!!

Because the Julia vals are not restored on iteration start.


Title: Re: InvSphericalC formula
Post by: Jesse on February 26, 2011, 08:34:36 PM
Quote
Another question!

It would be useful if you store the 2 old values of xyzw vector.
I may use it for Phoenix formula, function of those variables
It is easy make this; initially Sold = C, Sbak = 0.
After all formulas cycle, update Sbak with Sold values, and Sold with S. S is xyzw vector :)

It is not that easy because i have 13 iteration loop functions that must be changed  :dink:
:o
But Phoenix fractals are so damn pretty :embarrass:

It is not impossible to do it with the current options, you have 4 double values to store one vec: dfree1+2 and Deriv2+3,
set them to 0 at iteration=0 (use bFirstIt to detect + increase it), then you have in the iteration:
the Sbak as stated above, the current vec + you calculate the next vec.
And set Sbak to the entry-vec.
 ^-^