Welcome to Fractal Forums

I have just released version 0.30 of Mandelbulber program for Linux

Download binary and source:
http://sourceforge.net/projects/mandelbulber/

new features:

    * very easy close-up by just clicking on image
    * added Tglad's "folding" algorithm which modifies all formulas to look like mandelbox
    * added sliders for scrolling image in render window
    * added edit field for changing colour frequency
   
fixed bugs:

    * corrected some spelling mistakes in GUI

other changes:

    * changing image scale in render window is possible during rendering
    * fake ambient occlusion is blocked in formulas other than trigonometric
    * source code was little cleaned up
    * all settings files was rebuilded

Makemeunse has prepared binaries for Widows 32bit version of Mandelbulber 0.30. All files are available for download on SourceForge.net
http://sourceforge.net/projects/mandelbulber/

Hi

First of all, thanks a lot for making this program open-source !!

Now, while playing arount with version 0.30, i think i've found some kind of a bug.
If you use one of the three "tglad ..." parameter files from the "settings" directory, the program crashes after rendering an image:

*** glibc detected *** ./mandelbulber: double free or corruption (out): 0xb6de03d8 ***
======= Backtrace: =========
/lib/tls/i686/cmov/libc.so.6[0xb6d08ff1]
/lib/tls/i686/cmov/libc.so.6[0xb6d0a6f2]
/lib/tls/i686/cmov/libc.so.6(cfree+0x6d)[0xb6d0d7cd]
/usr/lib/libstdc++.so.6(_ZdlPv+0x21)[0xb6ef96f1]
./mandelbulber[0x804bb9c]
....................

The main reason for the crash is the missing "background3.jpg" file in the "texture" directory. If you edit the settings file and change "background3.jpg" to "background.jpg", everything is ok, and the program does not crash.

But what's behind the crash. After playing around with the GDB, if found that it crashes in the following routine in "texture.cpp":

//destructor
cTexture::~cTexture(void)
{
        delete bitmap;
}

So, i think the real reason for the crash is because the "bitmap" struct is never used because of the missing JPEG File in the constructor "cTexture::cTexture(char *filename)"

Sorry, when i'am wrong, but my C++ Knowledge is not so good, i live more in the C - World.

Solution:
- Exit the program when there is wrong path to a texture
- Only "delete bitmap" when bitmap is realy used (don't know how to do that)

EDIT:

Sorry, i forgot to mention my environment:
- Ubuntu 9.10 on a Intel I7-860 with NVIDIA GTX 280

Regards,

Daniel

if (bitmap != NULL) delete bitmap; bitmap = NULL;

hehe, was also my first idea, but does not work. This is my actual version:

//destructor
cTexture::~cTexture(void)
{
        printf( "File == %s Line = %ld\n", __FILE__,   ( long )__LINE__ );
        if (bitmap != NULL ) delete bitmap;
        bitmap = NULL;
}

Don't know if it's because "bitmap" is a struct:

struct sRGB8
{
        unsigned char R;
        unsigned char G;
        unsigned char B;
};

Probably because bitmap is not set to NULL in constructor.

Ups, that's it, if there is no texture file, the "bitmap" struct is not even created...

Fix:
--- /data/download/mandelbulber0.30/src/texture.cpp   2010-02-26 20:20:14.000000000 +0100
+++ ./texture.cpp   2010-04-04 01:20:00.000000000 +0200
@@ -40,6 +40,7 @@
    else
    {
       printf("Cant load texture: %s\n",filename);
+      bitmap = NULL;
    }
 }

Thanks makc

Thank you for information about bug. I have known about this problem. Program hasn't got error handling for missing textures (or other files). I will use your suggestions and add additional conditions for situations when some file is not loaded. I will also add some error pop-up windows.

I have just released version 0.40 of Mandelbulber program for Linux

Download binary and source:
http://sourceforge.net/projects/mandelbulber/

New features:

- flight speed is adjustable during record mode by mouse clicking (left/right button)
- flight path record can be continued from last position stored in path file ("Continue recording" button)
- each shading effect is rendered into separate layer (higher memory consumption but possible post-processing)
- all shaders can be adjusted after finishing rendering
- added "Save image as..." button for saving image after final adjustments
- added post rendering effect based on z-buffer: fog
- added post rendering effect based on z-buffer: screen space ambient occlusion
- added "Post effects" tab in GUI
- added analytic distance estimation for Tglad's Mandelbox formula (faster rendering)
- added analytic distance estimation for Menger Sponge formula (faster rendering)

Fixed bugs:

- fixed bug in memory allocation for colour palette
- added error handling for missing textures (error messages)

I guess I am kind of late to join this wonderful party. I just downloaded the win32 0.30 version (and source code). One thing that puzzles me is that there is no License.txt file attached to this project at all. What does this mean for third party modifications?

The reason is that I would like to try accelerating this with CUDA, however I do not know under which conditions such a "fork" could be republished, if at all (In case I start modifying it, I would be using the Linux 0.40 version)

Anyway, thanks a lot for publishing this open source. This seems to be one of the most advanced dedicated 3D fractal renders out there.

Christian

Program is published with GNU GPL and uses components with the same licence (GTK+, JPEGlib, etc...). Of course you can modify the program but if you want to release your program you also have to use GNU GPL licence and publish source code.
Now this program is completely GPU unfriendly but it will be very nice to accelerate it.

GTK+ is actually under GNU LGPL 2.1 and libjpeg has a BSD style license. Nevertheless I respect your decision to release mandelbulber as GPL'ed software. Your software may not yet be GPU friendly yet, but is definitely GNU friendly ;)

It would be better to actually bundle a copy of the GPL license text with the distribution, just to avoid any confusion.

One of the biggest issues I am facing in the conversion is the limited floating point precision of most consumer GPUs. Your source code is very nice (apart from some non-english comments that are difficult to parse for me) and it has few dependencies other than GTK+ and libjpeg. Great! I was able to wrap this into a newly created kdevelop IDE project in just 20 minutes (I am very familiar kdevelop because we use it a lot at my workplace).

By the way I have some left over nVidia 9600 GSO graphics cards with 396MB RAM, PCI Express. Want one (for free)? They have about 400 GFlops peak performance each and they do support CUDA (and OpenCL). Consider it a donation for this wonderful software. All I need is your mailing address.

Christian

I have just released version 0.41 of Mandelbulber program for Linux

Download binary and source:
http://sourceforge.net/projects/mandelbulber/

New features:

- added high quality Depth Of Field effect
- improved coloring method for Mandelbulbs
- slider for fog visibility changed to logarithmic scale
- added information on progress bar about rendering post effects
- stop button is able to terminate rendering post effects
- added "About..." tab in GUI

Fixed bugs:

- fixed problem with g_thread initialization on some systems

On my OpenSuSE 11.1, this version hangs and becomes unresponsive after opening its GUI.
I tried your binary, as well as a recompiled binary.

Before that I had installed the 0.40 version.

I made sure the entire $HOME/.mandelbulber folder was updated with the version from
your new .tar.gz file

Is there anything else I should try?

Unfortunately there was fatal bug in post-render fog effect. Program tried to refresh this effect during initialisation of GUI when default settings file has another fog parameters than default for GUI. I have fixed this bug. Now program should run properly - I tested it on clean system. Unfortunately I didn't found this bug before because I tested it on my computer and I had little another default setting file.

Please download the newest version from:
http://sourceforge.net/projects/mandelbulber/

Great program Buddhi, I've made part of this pic (its a big image) http://fc08.deviantart.net/fs71/f/2010/101/4/e/Mandelbox_DOF_by_KrzysztofMarczak.jpg (http://fc08.deviantart.net/fs71/f/2010/101/4/e/Mandelbox_DOF_by_KrzysztofMarczak.jpg) my desktop background. My girlfriend thought it was an underwater photo at first.
Anyway, I can't find an option for depth-of-field blur in 3.0, and could you perhaps provide an example parameter file for a similar image to the one above, as I'm not sure what sort of values to put in some of the fields. I tried opening the sample parameter files, but it crashed the program. Thanks  :hungry:

Depth of field effect was added to Mandelbulber v. 0.41. Problem with loading settings files with Mandelbox parametes was fixed in version 0.40. Windows version is delayed because Makemeunse is on vacation. Please be patient (or install Linux :) - Ubuntu Linux is very easy to use)

If you want to try with Mandelbox fractals in Mandelbulber 0.3 please use following settings:
power = 2.0 (when Mandelbox formula is being used, scale = power)
folding limit = 1.0
folding value = 2.0
Close-up (scale) = 15 (Mandelbox is much bigger than Mandelbulb)

To zoom some interesting region you just only have to indicate it on image and press left mouse button.

Could you give us any wort of time line as to when the windows version of .4x will be ready? You have done some really great work, my friend :dink:

And is there any way that Makemeunse might be able to also port the 64-bit version to Windows? I have a quad-core Phenom II machine at home with Vista64, 8 gigs RAM. I specifically built the system for the purpose of deep fractal rendering and ray tracing apps. I find Linux difficult to work with because I don't understand how to utilize the command prompt terminal window thingy (I know that's not what it's called, so please don't flame me about it). In order to do stuff, I have to type commands into the box. For example, if I plug in a hard drive, I have to "mount" it first in order to access it. With Windows and MAC, I just plug something in and the computer instantly recognizes it, and simply shows up in My Computer or the Mac desktop.  Especially with the Macs, you just plug it in and it works. Or I just download a program, and either runs an install wizard that does everything for me, or I dump/extract it in a directory and run it. And then there's the problem with no drivers for 80% of the available hardware on the market. Don't get me wrong, I am very comfortable tinkering with computers; I've just never learned a damned single thing about coding or typing commands (the only programming language I know how to use is TI-BASIC), which seems pretty much a requirement to do anything more with Linux than simply booting it up. It doesn't help matters much that there's like 50,000-something versions of Linux available either.

So any help porting the Mandelbulber software to windows is greatly appreciated ;D
My thanks to Makemeunse for his and your efforts as well ;D

That makes sense Buddhi, I'm tempted to install linux just to try out the latest version! but probably better to practice being patient instead. Good luck with the porting job Makemeunse, I don't know much about PC rendering techniques so wouldn't be much help.  

Buy a little external USB-Disk (10-20GB for less than 50$) and install Ubuntu 9.10 on this disk. DO NOT install the Bootloader on your internal Disk, instead install it on the external one. Change your BIOS-Bootorder in the way that it boots from the external disk if this is available and otherwise it boots from the normal Windows Disk (bäh....)
This is always my first step to bring poeple to Linux  ;D

I have successfully compiled it for windows with a little modification in the code. I had to add the exp10() function, I gess it's equivalent to pow(10,x). Am i right? (I gess the answer is yes because I haven't noticed odd things while testing.)
What should I do to make it available to people who are waiting for it?  ;D

I'm really surprised. Thank you. Please send it to me. I will prepare package and put in on SourceForge.

Thanks to @knighty we have Mandelbulber 0.42 for Windows. I have just upload it to SourceForge.
http://sourceforge.net/projects/mandelbulber/

Full list of changes since version 0.30 you can find in attached txt file.

Excellent, works like a charm.

Have you noticed this propsed acceleration for ray marching? Do you think it is applicable to Mandelbulber? In this proposed technique it allows you to advance multiple adjacent rays by doing a single Distance Estimate.  A speed-up of factor 10 or so might be possible.

http://www.fractalforums.com/mandelbulb-implementation/major-raymarching-optimization/

 ;D
(http://www.fractalforums.com/gallery/2/932_18_04_10_5_50_40_0.jpeg)
That's what motivated me!

See also : http://www.fractalforums.com/mandelbulb-implementation/potential-optimizations-for-de-based-stepping/ (http://www.fractalforums.com/mandelbulb-implementation/potential-optimizations-for-de-based-stepping/)

is ther an example file for setting up an mandelbox rendering ?!
i choosed tglads formula, and by default it renders just a tiny sphere ...
when i enter -1.5 as power value, i get an empty screen :(

I wrote in post http://www.fractalforums.com/mandelbulb-software/mandelbulber-open-source-program-for-rendering-high-quality-mandelbulbs/msg15821/#msg15821 what you have to set up to render Mandelbox fractal

You can also find in default settings folder 4 files with example configurations ("Tglad - ****.fract").

I attached also 2 example settings files

this is what the program renders, when just loading the settings file:
"tglad scale -2 .fract"

:(

ok, the config files you posted work, awesome program :D
would love to see hybrid formulas also :D

negative values from 0..-2 produce black result ;(

Hmm, your problems might be resulting from using the settings files of the previous 0.30 release with the new 0.42 version. On Windows, don't forget to run INSTALL.BAT from the new 0.42 folder once more (acknowledge the overwrite with y for yes each time, or with a for all)

If image is completely black, please check location of the camera. Maybe it is inside fractal. You should try with high values of Close-up (zoom), for example 20. Camera will be far from fractal.

i slowly get a hand at the program , but where is the camera move navigation ? left/right would be cool to have! and up/down ;)
this program generates amazing images, but i am clearly lost navigating ;)

what happens when i click ?

on option to move directly to the surface point clicked, and zooming would be cool, how is it working now ?

Thanks Knightly! You are the best ;D chilli chilli chilli

I've been :fiery: my poor laptop's 2Ghz Core Duo processor all weekend with Mandelbulber v0.30 :evil1:
So I'll have to wait until its done rendering so I can close program before I run the *.BAT file  O0
I can't wait till I go back home to do some serious rendering on my big Phenom II x4 quad mobo - speed limit 3.2Ghz O.C. @3.6 :siren:

Hi

I have good news. I'm finishing writting user guide for my program. I will include it to the next release of Mandelbulber (it will be in few days).


Navigation in 3D unfortunatelly is much more complicated. You have several was to set-up camera:
1. Use some 2D fractal program and find interesting coordinates (very good for cose-up of areas near z=0)
2. Zoom (colse-up camera) by clicking on image. Each mouse click moves camera 5 times closer to the selected point
3. Use 3D navigator. First use "Init 3D navigator" or put some starting camera position (far from fractal), rotation and very low Close-up value (1e-8). Next move camera using arrows, "Forward" and "Backward" buttons.
Of course it is possible to mix all methods.

I'm still thinking about some improvements for 3D navigation. I'm open for some new ideas from you.

I cant wait for some interesting images from you rendered on my program.

what about industry standard wasd+mouse+arrows navigation? with low iteration count rendering while navigation is active.

I have just released version 0.43 of Mandelbulber program for Linux

Download binary and source:
http://sourceforge.net/projects/mandelbulber/

New features:

- added possibility to move camera up, down, left and right in 3D navigator
- zooming by mouse click can be disabled
- changed default camera position in "Init 3D navigator" function
- changed coloring algorithm for Mandelbox (Tglad's) formula. Now color in not depended on number of iterations and distance.
- added spherical folding mode which modify all fractal formulas
- modified shading formula
- Screen Space Ambient Occlusion (SSAO) is rendered using multiple CPU cores

Please remember about running install script before first program run

great, thank you very much, going and testing it :D

GPU (CUDA) accelerating all of Mandelbulber is a bit more than I can chew right now.

So I have begun working on a real time DOF post effect with CUDA first. I am basing this on the recursiveGaussian filter sample from the CUDA SDK, which implements a constant time (with respect to blur width) gaussian filter.  I split up the image into distinct layers (somewhere between 10 to 30 layers) each with different blur factor. In am composing the layers back together using alpha blending. This should look very similar to what Buddhi's CPU based DOF code is currently doing.

Oh, zbuffer values for background pixels are a little on the extreme side, something like 1E20. Was this intended? ;)

Christian

DOF effect is one of the fastest parts of program. Rendering of DOF effect takes only from 1% to 5% total time of rendering image. I'm wondring also about quality of this effect when it will be divided to only 30 layers. Now blur radiur is in range between 1 to 100 with unlimited resolution. Smooth value of blur radius is very important neer places which are in focus.


Yes, it was intended because there is almost unlimited view distance. My program scales depth value to normalised range. When object has size equals to close-up value and is located at depth = 0 then its size on screen will equals to image width.
Size_of_object_on_screen = size_of_object / (1.0 + normalised_depth*perspective_factor)
It means that when there is very deep close-up, the nearest objects are in range of depth between around -1 and 1 but objects which are very far have very high depth value.
This solution has many advantages. It's is very easy to adjust intensity of every effect and scale value of distance threshold using this normalised depth values especially during rendering of deep close-up animations.
I know that this is strange solution, because in OpenGL every value is in range <-1.0,1.0> but for me this is the best.

Yes and no. ;-) When you render simple things like a Menger Sponge and add some extreme Depth of Fields with lots of very blurred pixels near the camera, this ratio can easily be reversed.


I intend to apply some logarithmic scale to the size of the blur bins. That should take care of that. I am more worried about whether simple (but fast) Gaussian filter creates the "correct" optical effect.

Knighty has just prepared Windows port for Mandelbulber 0.43. You can download it from:

http://sourceforge.net/projects/mandelbulber/

Please remember about use install.bat script before first launch of program.

I wrote with beeper52's help user guide for Mandelbulber. If somebody has problems with using Mandelbulber it should be helpful. It can be downloaded from SourceForge:

http://sourceforge.net/projects/mandelbulber/files/Mandelbulber%20user%20guide.pdf/download

I have just released version 0.50 of Mandelbulber program for Linux

Download binary and source:
http://sourceforge.net/projects/mandelbulber/

New features:

- added "Light sources" tab in GUI
- main light is configurable (colour, angle)
- added random auxiliary light sources
- added 4 configurable light sources
- improved environment mapping reflection. Now is calculated real reflection vector.
- improved specular highlight effect

Please remember about running install script before first program run

example images: http://www.fractalforums.com/mandelbulber/adjustable-light-sources-new-function-in-mandelbulber/

P.S. Is somebody able to make Windows port for this version?

I'll try to do something.

Thanks to Knighty now we have also Windows version of Mandelbulber 0.50.

program: http://sourceforge.net/projects/mandelbulber/

user guide: https://sourceforge.net/projects/mandelbulber/files/Mandelbulber%20user%20guide.pdf/download

Please remember about running install.bat script before first program run

Just finished 1st render using new version (0.50) of Mandelbulber.
New lighting setup took a bit of getting used to, but am loving the results and
potential of this application! :D

See: http://www.fractalforums.com/index.php?action=gallery;sa=view;id=2356

i'm in love with your software  ;D
But i have some trouble to understand how animation work  :hmh:

edit : just noticed that there is a pdf manual, i'll read it :) :)

Some general advice for GPU friendlyness.

Using structures of Arrays is much faster on GPUs arrays of structures. Or in other words, your array of sComplexImage structs is a worst case for the GPU. I converted it to an a struct of arrays like this


Of course it needs quite a bit more new/delete operators to (re)allocate, but it's worth the cost.

Another thing are the sRGB8 sRGB16 structures, which currently have three elements. Now the GPU can much better work with 4 element vectors, so I am going to use the CUDA native ushort4, uchar4 types here instead.

The above modifications result in much higher speeds of memory access, where one can achieve the full 80GB/sec or so of modern GPUs (some modern ones have even higher bandwidths in excess of 100GB/sec).

a cuda version of mandelbulber ?
I'll finally have a good excuse to buy a "Fermi" powered gfx card \o/

I received reports about bug in random lights generator. Few minutes ago I have uploaded new Windows version of Mandelbulber with fixed this problem.

http://sourceforge.net/projects/mandelbulber/

Now lights are distributed properly. There was problem with difference in implementation of rand() function between Linux and Windows.

I have also fixed problem with loading settings files created in older versions of Mandelbulber.

I have just published next version (v0.60) of Mandelbulber program for Linux

Download binary and source:
http://sourceforge.net/projects/mandelbulber/

New features:
- added fully configurable Kaleidoscopic IFS formula
- added "Kaleidoscopic IFS" tab in GUI for all IFS parameters

Fixed bugs:
- fixed problem with loading settings files generated in older versions of program (all missing parameters are filled by default values)

Other:
- rewritten all algebraic functions in whole program (all use new CVector3 and CMatrix33 classes)

Please remember about running install script before first program run

Knighty or Makemeunsee, as always I would like to ask you for Windows port

For Kaleidoscopic IFS I use following formula (code)

Thank you, I've been waiting for this formula   :tease:

@Knighty or Makemeunsee, please hurry ;)

Woohoo! ;D
I'll send you the win version ASAP.

Knighty for the win!  :D

Lool. I meant win32 version.

Now there is also available Mandelbulber 0.60 win32 for download.
http://sourceforge.net/projects/mandelbulber/

Thank you Knighty.

Please remember about running install script before first program run

Many people waited for making animation using key-frames in Mandelbulber. Now I can announce that I have done it.

I have just released Mandelbulber 0.70 for Linux

http://sourceforge.net/projects/mandelbulber/

New features:
- Making animation using key-frames with Catmull-Rom spline interpolation
- saving images in PNG, PNG 16-bit, and PNG 16-bit with alpha channel
- all fractal formulas can be mixed with IFS Kaleidoscopic formula (Kaleidoscopic IFS Folding mode)
- all shaders was updated for generating alpha channel

Example animation with key-frames
http://www.youtube.com/watch?v=byCpJ82DXJQ

Impressive animation. Buddhi you are really bringing 3D fractals to new heights.

Agreed - this is fantastic. He adds features more quickly than people can even ask for them.

I think Buddhi could open a paypal account to take in a few donations to support his development, he really does deserve some gifts.

Thanks  :embarrass:, but only what I need is enough free time. Money will not help :). The best gift for me is to see your renderings generated with Mandelbulber  :D

Holy cow!

That animation almost made me cry Buddhi, so beautiful!

+ You are a great person!

I hope to get some hardware soon that is more compatible with your app!

AMD's 6 core CPUs might deliver the best bang for the buck!

Just got to say what a nice man Buddhi is - he doesn't want any money for his wonderful
program! If it was possible to give a gift of time, I wouldn't hesitate!!  :D

Knighty have made for us Windows port of Mandelbulber 0.70

http://sourceforge.net/projects/mandelbulber/

Please download and enjoy!

The program runs in Win7.

(Great job Buddhi and Knighty).

I haven't tested it on Win7. On Vista 64bit it works. If there are some problems on Win7 (or it works) please tell me.

I have found some terrible memory leak in Mandelbulber. It was around 10GB/hour during rendering of animation  :death: I have fixed it and released next version (0.71) for Linux and also for Windows

http://sourceforge.net/projects/mandelbulber/

Changes:
- reorganized buttons for key-frames - now are on bottom of window
- fixed bug: start_frame and end_frame parameters weren't refreshed in edit fields after loading parameters
- fixed memory leak: missed deallocation memory for fractalParams
- fixed memory leak: missed g_thread_join() function in PostRendering_SSAO() (8MB per each thread creation!!!)

P.S. I have learnt how to prepare and use MinGW (GCC for Windows). Now I can prepare Windows versions of Mandelbulber by myself. Thanks to Makemeunse and Knighty for preparing all Windows ports.

The windows version 0.71 requires libpng14-14.dll  to run but it is not included in the zip, I had to grab it here:

http://ftp.gnome.org/pub/gnome/binaries/win32/dependencies/libpng_1.4.0-1_win32.zip

Cheers ;D

Thank you for info. I fought that is not required. I used libpng12 but maybe some component of GTK uses libpng14.

No probs,  it is  LIBGDK_PIXBUF-2.0-0.DLL and  LIBCAIRO-2.DLL that both need it.

I noticed that missing file as well, but I just stole it from version 0.70...lol..

Okay, I did notice one issue today while I was screwing around with the 0.71 version at work (Windows XP).  Looking at the messages in the DOS screen of Mandlebulber, it says: Warning! g_thread NOT supported    
Okay, so what's a g_thread?  This issue is also replicated on my home machine (Win7 x64) - going back to the 0.70 version (on both machines), I get (what I guess is) the normal message, which says: g_thread supported
Like I said above, the only problem I had with the 0.71 version was with the missing libpng14-14.dll file, which I copy and pasted from the 0.70 folder.
I didn't really have time to play with the program versions and compare them to figure out what impact the g_thread has on Mandlebulber.  Buddhi, any ideas what this could be, and how to fix it, if necessary?

some other bugs in v 0.71:

- the values in "distance" (kal. IFS Tab) will not saved. If you load settings they show "0" again.

- Same "random lights" (win)-problem like some versions before: (http://www.fractalforums.com/mandelbulber/problem-with-aux-lights-%28only%29-in-windows-version/msg16963/#msg16963 (http://www.fractalforums.com/mandelbulber/problem-with-aux-lights-%28only%29-in-windows-version/msg16963/#msg16963))
The coordinates from distributed aux. lights are all the same:

3: Number of lights = 10
Light no. 0: x=0,998850, y=-0,998850, z=-0,149376, distance=0,000063
Light no. 1: x=0,998850, y=-0,998850, z=-0,149376, distance=0,000063
Light no. 2: x=0,998851, y=-0,998850, z=-0,149376, distance=0,000063
Light no. 3: x=0,998851, y=-0,998850, z=-0,149376, distance=0,000063
Light no. 4: x=0,998850, y=-0,998850, z=-0,149376, distance=0,000063
Light no. 5: x=0,998851, y=-0,998850, z=-0,149376, distance=0,000063
Light no. 6: x=0,998851, y=-0,998850, z=-0,149376, distance=0,000063
Light no. 7: x=0,998850, y=-0,998850, z=-0,149376, distance=0,000063
Light no. 8: x=0,998850, y=-0,998850, z=-0,149376, distance=0,000063
Light no. 9: x=0,998850, y=-0,998850, z=-0,149376, distance=0,000063

---------------
HEITER WEITER,

SaM

I checked this g_thread problem. It is not important. There was wrong message in program. Function g_thread_supported() returns TRUE also when g_thread is already initialised. As you saw, program worked properly with many threads.

Unfortunately you are right. This parameter wasn't saved and loaded. It will be corrected in next release.


Do you have settings file for this case? It looks like you put some wrong parameters of lights. I have to check it but it will be easier when you send me setting file (you can attach it to the post, but before you have to change filename extension to .txt)

I have just released new version 0.72 of Mandelbulber program for Linux and Windows. I have fixed some bugs and have added something...

http://sourceforge.net/projects/mandelbulber/

New features:
- added undo and redo buttons (100 levels of undo)
- undo buffer is saved in files and is available after restarting the program

Fixed bugs:
- IFS_distance parameter was not saved and loaded
- missed libpng-14-14.dll in package for Windows
- background image disappeared when close-up was lower than 1e-10

Please remember about running install.bat script before first program run

Buddhi, that is TOTALLY AWESOME that you have finally added undo and redo buttons (and the buffer save function is pretty cool, too)!!  Thanks!

I have started new forum thread in Mandelbulber section. I will do the same for every major release of program.
http://www.fractalforums.com/mandelbulber/mandelbulber-0-80/