Welcome to Fractal Forums

Mandelbulb3d v1.7.8


http://www.fractalforums.com/index.php?action=downloads;sa=view;down=34


I have not encountered problems with this version, but because of a function that adjusts the thread priority while rendering, it seems more safe to me that this should be tested abit before i delete the older version.

Please read the textfile in the archive for a short documentation.

CAUTION:
USE ONLY SAME PROGRAM VERSIONS TO CONTINUE ANIMATIONS OR BIG RENDERINGS DUE TO COLOR OR LIGHTING CHANGES!
ALSO FORMULA UPDATES CAN CHANGE RENDERINGS!


Thanks and have fun!

What is new in version 1.7.8:
  -  Further bugfixes.
  -  Added DEcombinate methods to mix escapetime formulas with dIFS (see the
     new example parameters containing 'mix'.
  -  Now the ini-file is stored in the programs folder and the formulas are
     only read from the initial directory for formulas, so several programs
     stored in different folders can work independent from each other.
  -  Formula and a few parameter update.
  -  Stereo calculation reworked, should give more accurate results now.
  -  The slowest thread in calculations gets a higher priority to match
     calculation times better.  If this causes trouble select 'lowest'
     thread priority in the 'Internal' tab.

Very nice thanks Jesse :D

And thanks for the plenty of inspirating formulas, the fuel for the program.  :)

Yes, thank you guys!!!  :beer:

I never had the time to download 1.7.7. due to lots of renderings going on.
Downloaded 1.7.8. yesterday. Jeeeeesus, all those new dIFS formulas, do you two work around the clock with this?  O0
Had great fun with dIFS yesterday (had a day of from work) results will be posted soon.

Again, thank you for all the work you two put into this Jesse & Luca!

Johan

Jesse, thank you again for this sublime work





ERICR

Jesse I am noticing problems with de in some cases, with some kinds of surfaces

Some are rendered perfectly normally but weird stuff appears on higher iters. Do not know how in the world I can "fix" this... Sometimes a trick works;

"If DE < a value, return DE, else return sqrt(x*x+y*y+z*z)"

Sometimes it does not work neither... :(

And note that this screws up the dynfog... :'(

Here an image, I am upping the formulas in my thread

example of the bug:


Mandelbulb3Dv17{
W.....S....O/...7....2....UIHa4TdQj/.jq2I4NFO42E8eJ/d/d.ozPi8BnSSSssz2xx2zriiQyj
................................/vLcVOUpxz1........A./..................y.2.....
................/M.0/....6EJ1...e.....E2.....UB3Xa.37CqD/..........m/dkpXm1.OaNa
y.....kD12../..........wz..................................u1....y1...sD...../..
.w1...sD9ysz5EFWXx1..........GPdL1Ws..djztZH15db2uHF2fzDG6SLzs8o.qCYfYpDGWzFu.jS
6uXBLptvEiGKzm10tpX2WrpD......oZ..............kD.2....sD..kz....................
.............oAnAr1...sD....z.qRo4..sFP.UW5h/.ISo4.stFP..e5h/.mSo4......7....k1.
..................kz.wzzz1.U..6.P....M4...EB....W....k1....F....8/...I1.....SF52
...U.qFG9yzb2zzzRYoWzz7lz16.mc..zXCc..k18XGQeGyD.sIRhrVAkz1..........2.28.kFrA0.
.Ub96aAIVz9.1se7Umvxz0........../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....U.....Hdp4QZFLG4B3.rJaQ..........................U1C....6..
..........kAnAnAnAnwz..........E.............................U.E........kz1.....
...wzcNaNaNaNazj................................................................
.....................2.....3....6.....pPgZbFjl4NhALShZYFH/......................
....4MU/0.......................................................................
.............E/EMo0FIhT67.2.....................................................
................................}

Uh, another buggy surface; CitrusIfs, has a nice aspect but on high iters it goes crazy (also the fog)

 Now M3D is even portable !! :banana: :music: :chilli: :love:

Thank a lot Jesse is a amazing software !!

Yes, there is an error in the DE calculation, it would also show up with only 1 iteration if you increase the DEstop value!
(Btw, that is a good test to avoid such things in future.)
Example:
Mandelbulb3Dv17{
X.....S....O/.../....2....UlFSVTWp4J.XJRQFrp7x6E9NYvSWKN0.gnyD6i7732.7USpl3RL5.E
................................iy6Y/BLOcy1........A./..................y.2.....
................/M.0/....2Uk....8.....E2.....EqaJ/fFUTvD/..........m/dkpXm1....6
//....kD12../..........wz..................................u1....y1...sD...../..
.w1...sDXqIDScniryn5.NLVavKSzebXe20tp0sDGsGWItOHgxfwsP1ItSVezoGuLx44uDvDaVuQJGJl
1yvW.hEubvygzaGRCxT3FbuD......I0..............kD.2....sD..kz0...................
.............oAnAr1...sD....zU3Sx8.yspf.MaLj0UXSx8.qupf.shLj0U/Tx8....../....k1.
..................kz.wzzz1.U..6.P....M4...UE....W....k1....F....8/...I1.....SF52
...U.qFG9yzb2zzzRYoWzz7lz16.mc..zXCc..k18XGQeGyD/sIRhrVAkz1..........2.28.kFrA0.
.Ub96aAIVz9.1se7Umvxz0........../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....2....E.....I....U.....Hdp4QZFLG4B3.rJaQ..........................U1C....6..
..........kAnAnAnAnwz..........E........kz1..................U.E........kz1.....
...wzcNaNaNaNazj................................................................
........................}


So the DE works only on low distances, on higher it gets worse, this is because you got the errors on higher iterations, because the DEstop gets bigger, relative to the object.

What exactly must be changed in the formula, i can't say yet.  Hope you find it.

Cheers! 

I think he pretends all terms under square root; all those de work fine, others are crappy
I need to change the expressions to do this! Or it will never work

Another way you can verify the DE:
in DEcomb-mode, the Zmid 2d calculation coloring shows the distance estimate!

It should be circles in repeated, equal distances, but this is how it looks for the LimpetIFS...
(press "Zmid" to show it)

Mandelbulb3Dv17{
X.....S....O/.../....2...............1.......s1E................................
................................OaNaNaNaty1........A./..................y.2.....
................/M.0/....2k7/...8.....E2.....QzIXLua2PuD/..........m/dkpXm1....U
z.....kD12../..........wz..................................u1....y1...sD...../..
.w1...sDYsAIxzzzjz1..........UAbUdisqcljMaO9c9iBOw1v41auWPX4zU2bVezzzzyDaoHzUvMS
3t17C1JzzzzvzeCphUisqclD......2/..............kD.2....sD....0...................
.............oAnAr1...sD....zU3SE9.ys/h.Ma5o0UXSE9.qu/h.sh5o0U/TE9....../....2..
..................kz.wzzz1.U..6.P....k1...EB....L....c3....F....8/...I1....UJl52
...U.eoZgzjdqZpz8R7vzPeRN/..0c..zzzz............/wyBIC4OWz1..........2k.8.kXWF1.
.sM93P58iz9.MmnWK2zwz0........../6U0.wzzz1................................E.0c..
zzzz.................................2U.8.kzzzD.................................
/6U0.wzzz1...................................QoFx0......jpUfEe4.tANU.6i4ve9G..gS
q/UIcEDul1kzE0A.QHHcJa7......EaEkyth......EMDBeH00......yj3MAZ2......cDOISPK....
..EaqVaek0.........U./0.zz/6FG1.yzHkqs/.xzZteKA.................................
E6...6....E.....I....U.....Hdp4QZFLG4B3.rJaQ..........................U1C....6..
..........kAnAnAnAnwz..........E........................................kz1.....
...wzcNaNaNaNazj................................................................
.....................2.....3....6....kIOh/LNoZYFH/..............................
.sU1....0............AnAnAnAnAzD..........2.....................................
...wz.........zDOaNaNaNatz9.....................................................
................................}

The problem that produces those spikes in the DE might be...

R = sqrt(x*x+y*y)
DE = sqrt4(sqr(R2 + z*R + Rmul*R) + z*z) - R2

if z*R becomes negative so the term inside sqr() becomes zero, the DE is calculated only by sqrt4(z*z) ...
what makes such low values on bigger distances, does ...) + z*z*z*z) - R2 work?

Uhm, no ;D sqr not sqrt :) ...

sqr(a) = a*a

Thanks for the help. The problem is that I really don't know how to improve that formula because all terms are already under sqrt and there are no other ways to fix. When distance grows the field imho works as expected (or no?)

Now I am beginning to understand... But the expression is that, how can I think another working expression that is = 0 on the surface? :-\

Edit - don't worry about that, I'm trying to rewrite the expression now

The problem is easily solvable! Like I did for other troubleful surfaces, I simply cut off  :D

1. I calculate the de
2. I calcuate R - 3*R1
3. I calcuate abs(z) - 3*R1

DE = Min(de1,de2,de3)

(That's a cylindrical scissor baby ;D )

Now it's 3x faster and more stable :beer:

  :thumbsup1:

Hello
Concerns Mandelbulb general and also V1.7.8
First sorry for my poor english …and my low knowledge level in fractal’s matter.
I have some problems of understanding differents points of the Jesse’s  « Read me » and I do not know where I may show these problems on the Fractal Forums, so I try here.

1 – Formulas
I do not understand what is exactly done :
1°) under the option « Alternate »(used in turns ?) what exactly happens ? I understand the sequence of formulas but no the aggregation of each partial result.
2°) under the option « Interpolate », which variable is interpolate ? DE ? and how ?
3°) under the option « DE Combinate », for me, there is no making head or tail for it.
Replaced the average DE combinate method with an inverting and max comb.
What is precisely a combination ? How works a « inverting and max comb » ?
Options for inside rendering (beta) and "min" DEcombinate with overlap ???
S1 = minimum function with a linear overlap function ?
S2 = like S1, nonlinear function, more smooth overlap ?
4°) dIFs
Before V1.7.7 , mixing of dIFs with other formulas doesn’t work (OK) but after
  « Added DEcombinate methods to mix escapetime formulas with dIFS »
how can I know which formulas are « escapetime » in the lot of proposed formulas ?
2 - Lighting /Background Picture »
« Fit left+right borders, with . . . » : Independently of the relative sizes of the main image and the background picture, I have no effect . . .  or I don’t understand the wording.
3 – Save Image/png.par
« Save text parameters in PNG.file, if selected »
Where is the resulting file ?
Thanks in advance

Jesse : Congratulations for such a detailed and useful job but why don’t you use a  no black color for the new text in you « Read me » ?

The best answer is; try and try... It's too long explain all in detail for Jesse :beer:

I agree with Darkbeam for the most part.  It is often more an intuition for many settings more than any set procedure  :embarrass:

But I can help with One of your questions!


Try varying the Weight Parameter.

Here's a thread discussing it, with an example m3a file to show how to animate the weights and morph one formula into another  (along with many interesting shapes in-between).

http://www.fractalforums.com/mandelbulb-3d/the-interpolations-in-mandelbulb3d/

And may as well post the gif just because...

(http://www.sockrateaze.com/stuff/proof.gif)

Other advice:  When you see an interesting image posted in the Gallery or Forum or elsewhere, and it has the parameters posted, copy them and open them in M3D - then look at how it was done, change stuff to see what happens, and learn from others.

Hope that helps..

Wow - a new version? Sweet!

Looks like I have to figure out how to upgrade, thanks a bunch Jesse!  :beer: