Welcome to Fractal Forums

Mandelbuld 3D v1.8.6 beta test



http://www.fractalforums.com/index.php?action=downloads;sa=view;down=45

Nice to see you back, Jesse!  :joy: And with an immediate update too... Woohoo!  :beer: :beer:

The new "soft" hard shadow is a great addition for realistic renders. Thanks so much Jesse!

And combined with the MC method, WOW, I have no idea about all these new MC options, but the quality is.....WOW...and even without Anti Aliasing!! Is that how Mandelbulber works?

great to have you back jesse, thanks for the update!

@bib what do you mean with MC method ???

i see, montecarlo is meant by it ;) shame on me!

Thanks you for missing me  :)

just did 2 weeks of intensive programming and this is not what i will do very often...
but i keep on lurking this forum regulary.

I updated the beta because of some issues i found already, check the beta-2 please!

@Bib:
Mandelbulber uses a different lighting with a better ambient occlusion than default m3d lighting, but of course not a that slow rendering method like monte carlo or something.

Does this mean we can use 16 bit grayscale PNG maps in dIFS heightfield formulas?

Does anyone know of a (bitmap/vector) graphics program that exports 48 bit PNG?

Yes, this should work now, if you have 16 or 48 bit png files that does not work, please report.


Was searching myself for a free program for my tests, but only found cinepaint, a follow up version of a filmgimp? or so... forgot.
But the windows version is a bit crappy, couldnt use it anyhow.

So i am also interested in further suggestions...

Hi Jesse,

When saving as PNG with v1.8.2 I get this error message:

Access violation at address 020313A3 in module MANDELBULB3D.EXE. Read of Adress FFFFFFFF.

After I click OK on the error message box, the PNG file saves OK.  ;D

When saving as PNG with v1.86 I get this error message:

Access violation at address 005E3CC9 in module MANDELBULB3D.EXE. Read of Adress FFFFFFFF.

After I click OK on the error message box, the PNG doesn't save, nothing happens.  :hurt:

All other filetypes save OK without error messages in any version.

Cheers, glad to see you around again.

I would have assumed that. When I'm not terribly wrong, the monte carlo method is similar to the way, some of the "big" raytracers like cinema 4d generate global illumination. I guess this goes beyond mandelbulber's lighting model...
looking forward to test it. ;D

Dunno what exact model those pretty renderers use, i only know that m3d makes some simplifications, without them it would be to slow.
But maybe i extend the possibilities furthermore...

Ok, thx for this report, hope i find the reason...

Dunno as well, but the stochastic mode is the slowest in c4d too. And it also creates some grainieness...

Amazingly realistic, cheers Jesse! :o

Here a test parameter with a reflective ball in a coloured room. The walls, completely white in the default engine, become fancily colored in MC mode :beer:

Mandelbulb3Dv18{
d.....S....O/...0....26....8AuCD1m/vz0TQNKBHhk1E3meZSOKlpy1L3fVd.JQizMeL.EJ.SEvD
................................b6B89ZViE.2........A.J0........E........C/2...wD
...UzM....E7..../w.0/......y....A.....Ej.....AzUMbFDg3pD/w.........m/dkpXm1....U
z.....kD92../..........wz.................................U5.....y1...sD...../..
.w1...sDuhg27TOw6xnS8UiSmNrFz8DiSAt24bmD5.RkUYV6YwP8euIXgK16zUujCC/sS2pDticmwXVn
5xPNRf3Id2LGzaoRRS8Xywij......o/....7.........kD/6....sD..E.....................
.............oAnAt1...sD....zw1.........................................0....k1.
.....i53iz1.......kz.wzzz1.U..6......M4...EB....W....A4....F....8/...In4....SN52
...U.GPiBzTJYVszRYoWzz7lz16/3.lzcXf7..UGsza1HGyj5cilpNHjLz1.2LyBHClozG..LwD.z1..
.EUWETAHnz1.XhVpgiM1.1UDR80mwSwD/6U0.wzzz1...........s/...................E.0c..
zzzz.................................2U.8.kzzzD............8....................
/6U0.wzzz1....................................6U.0....Uvqp.U.06.....STE2zzzz....
.sytH.6U.0....Uv620U.06.....iDL7..kz.caOzvi7e.6U.0....Uv3A5.....zzzzi1AS.06U....
.sSovxD.......kt..EsUa3feeWCNqGQIJ36wk8EwyLsUa3f................................
E....6....E.....I....o....UQjxKPXlaQ7NoI.IaQ..........................U1C....6..
..........U........Yz........SyD................................................
..............zDBnAnAnAn8.InAnAnAnAsz..........E...................wz...........
.....................2.....3....B....A5QcJaQZZYFH/..............................
.sU1.sk1DwU.02E......MaNaNaNadyD..........YNaNaNaNatz0..........................
...wz.........zD........kz1........wz...........nAnAnAnAbz1........wz.........zj
................................}
{Titel: sphe35*}

Nice to see you back jesse!
Great new version, that Monte carlo renderer is too cool!

It seems that lights have slightly different positions in main render and MC.

Thank you so much for this amazing update, Jesse!  The Monte Carlo renderer is fantastic but I have a question. Does it end when it reaches a special raycount or do I have to stop the rendering when I think it looks OK?

ty...


Hmm, it shouldnt... can you give me the params?



It doesnt stop, you can alway stop and continue later on, but dont forget to save the m3c file.

Usually there should also be a noise filter, but i have not the time to do it in this release.
I recommend for noise filtering something like a nlm filter what can be found in fitswork for example.

Thanks for your quick reply! Using MC takes quite long but the fractals look so damn good. I will most likely re-render all my fractals. You're a genius.

i have not found the MC button yet, how to activate it ?!?! ???

I also had to ask a friend :DDD






This is in the Utilities tab

Ok, should be ok if you uncheck the 'png par:' option in the 'Save pic' tab, i fix it in the final release!

thx

I've just done another test and realize it was me who inadvertly modified the light position. Sorry. :hurt:

lol

Yes indeed!


Cool, thanks!

Probably I have too outdated PC to use it for fractal generation, and I'm thinking about new PC. But probably problem with win2K is compiler with no intention of code working in win2K;)

:D :gum:

For 16 and 32 bit (no 48 ) you may use GMIC, not the gimp plugin that is limited by the gimp 8bit bottleneck , but the command line that anyway include a nice viewer
for all OS http://gmic.sourceforge.net/

Ty PhotoComix, i am a bit nerdy when it comes to commandline usage of programs  :)
but i give it a try!

Btw, with 48 bit i meant 3x16bit RGB, for saving RGB png's...

Oh no for 16 bit i intend 4x16bit for RGBA (A for the transparency), so no problem for  3x16  RGB , :dink:

Humm

  GraphicsMagick http://www.graphicsmagick.org/index.html is more straightforward for similar task but is again for command line

Oh Boy, doin the Happy Dance now

Test MC render in 600x450 on a corei5 laptop. Left image: normal render, 17 seconds. Right image: MC, 45 minutes.

(http://fc00.deviantart.net/fs71/f/2013/045/7/9/mc_test_by_bib993-d5uxykz.jpg)

that mc rendering combined with the 16 bit heightmaps will keep us busy for all over the year :D
woot, i wonder what amazing piccies will be created ... !!!! crazy stuff!

Impressive demo scene bib! really looks like antialiased ;D

Thanks taurus. Even without MC, the softer HS is a fantastic addition.

ck: I agree that the potential for creativity and ultra realistic pictures will be huge by combining MC with soft HS, reflections/transparency, heightmaps,...

Seeing these super slow renders reminds me of the brute force monte carlo GI renders I did of the original mandelbulb (http://lyc.deviantart.com/art/siebenfach-139038934), on a dualcore AMD back in 2009.

I really need to get some distance estimators into Indigo, hopefully with Knighty's help!

Just wanted to mention that i have updated to beta-3 with most fixes...
please still report all issues with it so i have not to change the final release that often.  O0
 :beer: :beer:

Jesse: Just curious, what random number generator are you using?


I'll try my best. :) That would be cool.

The fastest i found so far, dunno if its the best one:


Seed is initialized for each thread independent with its Thread ID and the built in random generator.
.. with Seed as Integer.

Absolutely great render comparison, bib!

Jesse, I can't change the term 'Author' to my name. I started a new fractal from scratch just to test this new option. I clicked the Author button and replaced 'Author' with my forum name. When I clicked the Insert button it inserted 'Author' instead of my name. Am I doing something wrong or is it a bug?

Hmm, seems there is a bug, maybe you must restart the program when changed, i will fix it...

i think the only way by now is to edit the Mandelbulb3D.ini textfile and add 2 spaces plus your name behind "Author"

That did the trick. Thank you very much! However, when you've fixed it I will check it again to confirm it.

You maybe need to try a better RNG In order to make the rendering converge / reduce noise level. The RNG quality (periodicity, uniformity and such) can have a huge impact. I'm not an expert though, maybe Lycium could give us better insight.

I will give it a test in comparison with the built in RNG, rendering something up to same average raycount.

Btw, the code can be simplified further by omitting the shr edx, 8 and adjusting the norm factor.
Guess this was done because the lower bits are not that uniform... but 23 bits or so would be more than enough i guess.

Of course the magicians mathematicians decisions should be preferred...

...ähh, did some testing and can't see any differences (beside the speed advantage) in the results of both RNG's.
Tested with a 6000x3000 pixel background map with a one-pixel white spot on dark background and performed a DoF on it.
The result is a noisy disc bokeh that resamples the randomness quite well...
but as the raycount is limited to 65535, i don't think that a very high quality for the RNG is neccessary at all.
Those special cases where this would have been taken into account are not realistic for a CPU renderer, not if you want to wait a month for the result...

When trying to do buddhabrots I searched number of articles about random numbers and found some article claiming that best algorithm is quasi random number generator, cos it don't pretends of being random but instead concentrates on the most equal distribution possible, and someone proved, that quasi random number generator could be expanded to 360 dimensions for doing some sort of financial analysis.
Throught I newer looked deeper.


Doing google search I found few links. Knighty's quasi-random buddhabrots indeed looks smoother than random.

Here is thread by Knighty:
http://www.fractalforums.com/programming/buddhabrot-montecarlo-and-quasi-montecarlo/?PHPSESSID=b140a5bcfb93304fed816b4a0de8d24b (http://www.fractalforums.com/programming/buddhabrot-montecarlo-and-quasi-montecarlo/?PHPSESSID=b140a5bcfb93304fed816b4a0de8d24b)
And then this article:
http://en.wikipedia.org/wiki/Quasi-Monte_Carlo_method (http://en.wikipedia.org/wiki/Quasi-Monte_Carlo_method)


Still, I wount be soon, when I 'll see monte carlo in action, cos programm don't works on Windows 2000, and probably a compiler is a reason why programm don't works on Win2K :hurt:

What is the "No ipol" check box on the Object tab of the Lighting panel?

Thx Alef for the infos,

so i see that pure randomness is not the best for montecarlo, but to implement for example the halton method i must store at least the iteration number for the sequence to continue at a certain position.
This might be possible for one or two situations like ambient rays at depth1... might give it a try.
Just if you must continue at iteration 10000 or more, it might be unproductive to always iterate the RNG 10000 times for the next nn rays.

Nevertheless there are plenty of optimizations left, that is nice and a curse to know, lol!

To turn of the linear object-color interpolation, but it is a little buggy in the beta3 because it is not considered yet in the color presets (those with the preview image at the bottom of the lighting panel)!

Tried the halton pseudo random generator on ambient rays, the result is quite promising!

Attached as little example with an average of 13 rays per pixel...

Now i would like to have also a 2d pseudo random generator for disc distribution instead of a rect!

You could generate a 2D halton sequence (for instance by using base 2 for X and base 3 for Y), and then transform to a uniform disk distribution using formula 19 in the GI compendium: http://people.cs.kuleuven.be/~philip.dutre/GI/TotalCompendium.pdf

Nice article, thank you very much!

Because all values will be precalculated, i can take formula 19b   ^-^

MC render changed Depth and Hard Shadows.
Any suggestions on settings to change in MC to get closer to original?

Why do a new kind of render just to try to get close to the old kind of render???  O0
I think Jesse said somewhere that MC is no good together with  dynamic fog, that might be the problem...
Try to adjust contrast and gamma (but it will not change the lost fog of course...)

To achieve predictable results.
It is more important with the MC method since Lighting/Postprocess can not be adjusted after the render.

The dynamic fog is equal in the background of both images. (Difference mode comparison in Photoshop)

Is there a way to disable Hard Shadows?
It was not enabled in the parameters that were imported.

Pretty visible difference.
(http://www.fractalforums.com/index.php?action=dlattach;topic=14975.0;attach=8592;image)

Click the "Postprocess" button and disable the automatic calculation of hard shadows.

EDIT: Oops! I totally misunderstood your question. I didn't realize that you were talking about disabling the hard shadow in the MC renderer.

A quote from the Readme (hard to find in the whole text i guess):

    -  Hard shadows will always be calculated, switch off all lights that would
       give no big light amount or lies outside a room, use positional lights
       to light a room inside.  The only way to switch off the HS calc would be
       to decrease the max length calcuation to 0.


You can also light only by ambient, use a background image for example, but
i want to get closer to reality with this renderer, so i did the opposite:
send the parans back to main and the params are changed to fit more the
mc output!  This way you can judge the result better.

Here some more results from the montecarlo pseudo RNG approach:
2 images attached, first the halton numbers and all pixels get the same ones for ambient rays, it is more like many straight reflects...
so i did a way to shift the halton values (also pseudo random) based on the pixel position, what makes it comparable to the random noise generator with the ability of noise filtering!
The benfit is that it seems to converge linear with raycount and not quadratic!
(Or better: noise~1/raycount  versus noise~1/Sqrt(raycount) with standard RNG)...
Correction: this can't be really true, must be a bit different...  :hmh:

Of course one needs for every depth of calculation and every function (DOF, subpixelshift) independent pseudo numbers, but even only the first depth of ambient calculation brings a big profit.

The second images shows the difference with about 62 calculated rays in average, this makes it really useable!!!

Wow, MC is cool!
Thx Jesse!
Can i save animation with MC (in future, with i8 cpu :D)?

Lol, not in the near future i guess!

Currently i have to fix several intern issues because of a little feature i want to do...
to keep everything running again, it takes another week(s) for the release  :sad1:

In the future I will compare the old and new readme file with a nice little portable freeware app called ExamDiff - The freeware visual file compare tool (http://www.prestosoft.com/edp_examdiff.asp)


Thanks, that is exactly the info I needed.

Also, sending the parameters back and forth between main and MC to adjust will help me with any remaining issues.

Thanks for taking the time to answer my questions.
The new renders with with the modified random number algorithm look very good.

I do not want to divert the discussion but here is a file compare app I've used for long time, also free:
http://sourceforge.net/p/winmerge/news/2013/02/winmerge-2140-stable-released/

More to the topic: Waaaaahhh... I wish so much I had more time available to be using M3D.
It's got to be one of THE top fractal programs currently being run.
I had some little time earlier last year and used v 1.76 and a few earlier releases.
Beyond that, I feel fortunate to have been hanging around since late 2009 to witness the development of Jesse's M3D program and the growth of the ability of those who use it to create unimaginably beautiful fractal images.
(I'm fascinated as well by the interplay of so much magnificent creative energy flowing through the forums here of infinite fractal possibilities!)

Fred

Probably hitting many times the same number and clustering  reduces random number result hence less than linear increase. Artificial noise in the last attachment looks smoother than pure RNG, what looks somewhat like bad .jpg.  

In some articles  quasi- random numbers is claimed to lose superiority ower random numbers in aplications with large number of dimensions, but there were no explanation what number is large, but here it's not the case.

I guess it is not easy to avoid any correlations between all pseudo random numbers, for example if you use the halton numbers on certain bases, there are strong correlations between many base combinations, so you can use only a limited number of bases...
i am currently using only 2 twodimensional halton series, for first ambient depth and for DoF or hardshadows (if no DoF used).

Curses, you deleted the two "new" m3f of your previous beta in your new beta. I hope somebody :whistle: saved them ;D

i hope it that your want mandelbulb v 186 beta 3

PNG save with params works OK now with latest beta.  :beer:

I downloaded and installed version 1.8.7 and found a problem. When I click the "Check colors" button in the MC renderer nothing happens. I neither get a confirmation that the colors are OK nor do I get an info when Autoclip spec+diff has to be used.

EDIT: I just noticed that it's not always the case but in most cases that buttopn does nothing. Launch MB3D and import the default Mandelbulb into MC. Now click the Check colors button and you'll see that there is no message at all.

Did them here:
http://www.fractalforums.com/index.php?topic=6061.msg28944#msg28944


Ok, it only answers if at least reflections are calulated, else it is always ok!
Because the specular spots are not affecting the overall light that much, maybe if you choose a low power and high spec amount...
but this is not tested here...
(i may disable the button then if no relfections selected)

Thanks for clarifying, Jesse! :) I just wondered why I got the message to select Autoclip spec+diff in version 1.8.6 while I don't get any such info in 1.8.7 when I render the same parameter file. So I thought it is a bug. Disabling that button when colors are OK is a good idea to avoid any confusion.

Hi:  I am 68 years old and until today I have never written a fan letter.  I recently retired.  My career was in the field of analytical chemistry but my primary interest in fractals has always been aesthetic.
 
I have been using MB3D for about 10 months.    I lurk these forums regularly in search of understanding.  Immense thanks to Jesse for giving us this extraordinary artistic tool.  And thanks also to the various collaborators who have contributed significantly to the steady expansion of this amazing program. Thanks also to those who have taken the time to create tutorials and to all the users who have shared their parameter files.  The video tutorials of Don Whitaker were especially helpful to me as I began this adventure.   I am in awe of the scientific genius and artistic talent displayed in these forums.  I agree fully with a post by Apophyster where he said:
 
“I feel fortunate to have been hanging around …. To witness the development of Jesse's M3D program and the growth of the ability of those who use it to create unimaginably beautiful fractal images. (I'm fascinated as well by the interplay of so much magnificent creative energy flowing through the forums here of infinite fractal possibilities!)”

My ‘handle’ is Horbgorbler.  I took that from a definition I encountered in the parlor game Balderdash.  The definition of Horbgorbler is: “One who wanders aimlessly.”  Believe me; you all have given me countless hours of enjoyment wandering around in awe within the nooks and crannies of remarkable fractal environments created with MB3D.  Jesse, you have made one old man very happy. I’m a grateful fan.

Hi Horbgorbler,
And welcome!

I'm pleased that you found my words agreeable. :)
(But I do not like to make OT comments like that so I just leave it at that, except that I try to say what I feel and feel what I say.)

A suggestion though: Usually one of the mods comes along and suggests to a new poster that they say a few words in "Meet and Group" topic.
I'm not one of the admins here, but I thought I'd make the suggestion since maybe they're busy with other things.  Here's the handy link:
http://www.fractalforums.com/meet-and-greet/

Fred

New beta is available: version 1.8.7 beta-2

Most of the latest issues are solved hopefully, please report any issues with this file, thx!

Hello Horbgorbler,

welcome and thank you very much for the kind words, i am also astounded and pleased by all the results others made and with learning so much new maths and tricks from all the creative people here.

Thank you so much for updating! The Author option works properly now. You changed the MC renderer dialog quite dramatically. Super improvements! I did some test renders and everything seems to work fine. If I come across any problem I'll report it here.

Sorry, i have to update again, thats the problem when changing internal structs that must deal with all possible hybrid options...
m3d version1.8.7 beta 3
is out!

I have a problem rendering fractals when "Interpolate" is selected. When I click "Calc" it renders properly. Rendering it with the Monte Carlo renderer works also but when I click "Calculate 3D" I get an empty picture.

I discovered also some problems with interpolation hybrid and other combinations... could you send me the parameters please?

Thanks!

Here is the parameter file.

Thank you Teena,

in my tests with some beta's it renders ok, have you tested the latest version 1.87beta-4?

Can't keep track of all betas by myself, what might cause problems is rendering MC and main beside because both uses the same maps.
Will change something so maps are reloaded only when they really changed.

But there are still some strange issues i have not found yet because the debugger won't tell me anything about it even all checks are turned on  :-\

So all reports - best with parameters, you can send them also via pm - are welcome.

Oops! I didn't realize that there's a new beta. I'll download it right away to check it in the morning (it's almost 1:00AM in Germany).

Here is something OT but I don't know where to post it here. Some forum members may be interested in XenoDream and/or Hatchery but hesitate to buy these products because of their price. Garth Thornton informed us XenoDream beta testers that there's a March sale of XenoDream and Hatchery with 40% off (http://xenodream.com (http://xenodream.com)).

Beta 4 renders properly now when Interpolate is checked.  :thanks2:
If I find more issues I'll report it.

Hi all, and thanks so much for the new version of Mandelbulb Jesse! I wonder if anyone can help me with the newest version though... I just downloaded and ran the .exe, but when it asked me if I wanted to clear out the old version of Mandelbulb, I said no - so I could try to run both. (It may have not said "clear the old version" out - but whatever dialog box came up, I said "No" figuring that would allow me to run both versions, 7 and 8 beta.)

Now when I try to go into the formula tabs (on 8 beta) to change formulas, not many come up at all under any of the formula tabs. It looks like it "wants" to bring the drop down lists up to select formulas, but no formulas show up in the list - only a small "box" where the drop down list of formulas should be. Hope I explained that correctly. On the first formula tab I get about 5, but that's it. Do I have to load the formulas in somehow? Or can I not have two versions on my computer at the same time?

Thanks!

One way of doing it would be to put both executables in the same folder, you'll need to rename one of them of course.

Thanks Blob, I'll see if I can figure that out. For now I just went back to using 7. :)

Hmmm, well not sure if I'm the only person having this issue - but here is what I get for formulas on the beta version: (attachments formula1, formula2 and formula3 - not sure how to get the image to post without attaching it). If you look at the images you can see how the formulas pull up for me. I am running Win7, with a pretty decent machine, so I'm not sure what I'm doing wrong.

The message I get is on the attached md3d.jpg, I have tried using both yes, and no as options. I have totally moved all Mandelbulb3D files to another drive, and tried just using the beta from the C drive. I have tried re-downloading the beta file to make sure I didn't get a bad .zip file. I tried Blob's suggestion, and still I get the same issue with the formulas. (There are not many to use at all.)

Anyone else have any ideas why I can't seem to get the beta version to pull in formulas so I can play with them?  :embarrass:

Thanks!
Michelle

Re: Formula directories problem (http://www.fractalforums.com/index.php?topic=15094.msg58133#msg58133)