Title: Mandelbulber v2 - 2.08 Post by: Buddhi on July 01, 2016, 06:15:16 PM Mandelbulber v2 2.08
Free download (executables for win32 and win64, OSX and multiplatform source): http://sourceforge.net/projects/mandelbulber/ https://github.com/buddhi1980/mandelbulber2 Together with zebastian, mclarekin and bermarte we have finished new version of Mandelbulber. Development of this release was focused on implementation of material management and texture mapping. From now there is possible to define many different materials and assign them to fractal and primitive objects, like it is possible in most of 3D modeling applications. Materials can use textures for color, luminosity, diffusion, normal maps and displacement maps. Texture mapping uses bicubic interpolation together with mipmapping. For fans of 3D printing there is function to export fractal as point cloud (series of images with fractal slices) There is about 20 new fractal formulas and many new small features, improvements and bug fixes. What is new? - Materials: Added editor for materials as a separate dock - Materials: Added real-time preview for edited material - Materials: Added Material Manager dock which can be used to select, add, remove, load and save materials - Materials: Added selection widgets for materials (with preview) - Materials: Added possibility to assign selected materials to fractals and primitives. In boolean mode each fractal can use different material - Materials: Added loading and saving materials to separate .fract file - Shaders: SSAO and raytraced reflections are enabled by default - Shaders: Converted existing shaders to use material data - Shaders: Added bicubic interpolation for textures - Shaders: Added mipmapping for textures - Shaders: Added texture mapping algorithms: planar, spherical, cylindrical, cubic - Shaders: Added support for color, luminosity, diffuse, normal map, bump map, displacement map textures - Shaders: Corrected calculation of opacity for transparency effect - Shaders: Added "flat" mapping for background picture - Shaders: Corrected orientation of textures for background and environment mapping - Shaders: Fixed bug in FastAmbientOcclusion. Sometimes it generated white points. - Shaders: Calculating of normal vector converted from 4 point version to 6 point version which is more symmetrical and accurate. Rendering speed change is mostly not visible. - Volume export: added tool (Export Volume Layers) to export fractal shape as images with slices (e.g. to use for 3D printing) - SSAO: Much improved quality of Random SSAO effect - DOF: Corrected equations for blur opacity - DOF: Added multi-pass rendering of second phase - DOF: Corrected randomizing algorithm (eliminated stepping) - DOF: Added adjustment for blur opacity - Added new fractal formulas: * Msltoe Toroidal Bulb * Riemann Bulb Msltoe Mod2 * Transform - Surf Fold Multi * AboxModKaliEiffie * Imaginary scator Power 2 * Collatz fractal * Menger Pwr2 Polynomial * Transform - Power2 Polynomial * Mandelbox Menger * Mandelbulb_Bermarte * Mandelbulb_Kali * Mandelbulb_Kali_Multi * Transform - Rotation Folding Plane * Transform - Quaternion Fold * Transform - Menger Fold * Menger Middle Mod * Msltoe Toroidal Multi * Amazing_Surf_Multi * Mandelbox Menger - Fractals: Added checkbox on each formula tab to disable selected formulas in hybrid or boolean mode - Fractals: Corrected incompatibility for IFS formula with older settings files. - Fractals: Added detection of Not-A-Number results to avoid computation hanging - Fractals: Changed 'get Julia constant' behavior. When Julia constant is set, the Julia mode is automatically enabled and image renders. - Fractals: After selecting Julia constant, mouse click mode is changed back to "move the camera" mode. - Image: Added possibility to save an image as surface normals - NetRender: client-server are compatible until major version is the same (e.g. 2.08 will be compatible with 2.08-1) - NetRender: now there is possible to connect client during rendering of image - NetRender: Updated NetRender to send all textures through the network. - NetRender: Fixed bug: - crash - when client version was wrong it cleared message buffer after client was deleted (with message buffer) - UI: Added auto-refresh function. There is new checkbox just below RENDER button. This function should be good for experimenting with different fractal parameters, because image will re-rendered automatically when any parameter will be changed. - UI: Added image quality presets - UI: in preferences dialog now there are two tabs: general, image - UI: preferences dialog and animation player now opens only once (bring to front, when already opened) - UI: added option to load thumbnail cache from server (mandelbulber.org) which speeds up displaying previews for examples - UI: added generate thumbnails option to render thumbnails for all example files - UI: improved UI elements for selecting textures - UI: in edit field's context menu there are added options: "remove from keyframe animation" and "remove from flight animation". Added icons. - CLI: added voxel export as CLI option - CLI: added test option - CLI: fixed bug: Progress bar was not showed in CLI mode. - Animation: the player gets new images from current animation render, as they get saved to the animation folder - Animation: notification for keyframe/flight folder does not exist - Animation: Fixed problem with not updated interpolation of keyframes after change of parameters. There was missing clearing of interpolator cache. - Settings: Added error handling for missing value for any parameter in settings file. - Settings: Updated conversion of old settings (v1.21) to be compatible with new materials - Settings: Added conversion of settings older than 2.08 to be compatible with new materials. - OSX: Changed shared Dir for OSX - Performance: optimized access to light sources data - Performance: extendedAux now is not an array. It makes rendering of hybrid fractals faster - Performance: Added logging verbosity level parameter for WriteLog. This will reduce memory usage caused by log. - Performance: Changed refreshing if progress bar to be only when image is updated. This speeds up rendering when main loop is processed by every 10ms. - Performance: Added protection against using of partially allocated image (now is more thread safe) - Examples: Added animation: hybrid 02 - rectangle hieroglyphs animation.fract - Examples: Added example materials - Examples: Added example textures - Compilation: added CMakeLists.txt to build the project with cmake - Compilation: source code cleanup to eliminate cmake, clang and gcc warnings - Compilation: cleaned up structure of includes - Compilation: program prepared for compilation with MSVC - Compilation: switched to Qt 5.6.1 As always every feedback is appreciated! Source code repository (GIT, SVN) https://github.com/buddhi1980/mandelbulber2 Title: Re: Mandelbulber v2 - 2.08 Post by: Patryk Kizny on July 01, 2016, 07:15:32 PM Is this version already running on OpenCL or we're not there yet?
Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 01, 2016, 07:17:00 PM Is this version already running on OpenCL or we're not there yet? Not yetTitle: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 02, 2016, 08:56:57 AM What do you think about new possibilities with texture mapping and materials in Mandelbulber?
Now there are possible such a images: (http://nocache-nocookies.digitalgott.com/gallery/19/640_03_06_16_11_11_04.jpeg) Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 02, 2016, 11:29:39 AM There has been so much development with MandelbulberV2 in the last 12 months that it is hard to find time to really test out the features. There could be problems or possible enhancements that we have yet to notice. Therefore it would be very helpful if some of you could find time to test V2.08, and come back with any feedback. This is an example of a Menger Sponge with some options and polynomial controls. Code: # Mandelbulber settings file Title: Re: Mandelbulber v2 - 2.08 Post by: zebastian on July 02, 2016, 03:14:56 PM Here is rather unusual showcase of the formulas present in mandelbulber v2.08. A lot of fractals and all transforms are missing, since the space did not suffice :)
Here is a better version of the image http://z3bastian.deviantart.com/art/The-many-types-of-fractals-svg-619071520?ga_submit_new=10%253A1467463430&ga_type=edit&ga_changes=1&ga_recent=1 And here is the original file: cdn.mandelbulber.org/misc/The many types of fractals.svg Have a lot of fun with the new version everyone! Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 03, 2016, 04:32:07 AM An infinity of geometric patterns Amazing surf Mod1, just randomly choose one or more of the five fold types,
And if this in not enough, then change to Amazing Surf Multi, for an even greater range of infinity. Code: # Mandelbulber settings file Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 03, 2016, 03:07:51 PM I have just released Mandelbulber in version 2.08-1. There was one important bug fix, because program crashed when rendering of animation was started from command line.
I'm still waiting for some feedback from you about using of materials and textures. Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 04, 2016, 09:11:41 AM example of new formula Mandelbulb Bermarte. There are eight fabs() controls that increase the possibilities
Code: # Mandelbulber settings file Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 04, 2016, 05:07:59 PM It's starting to be waste of time to post about new releases here. I'm getting more feedback about Mandelbulber from another websites than from FractalForums! You only want to have the software, but nobody is interested to support developers just by giving a feedback. You are wondering why MB3D will be no longer developed, but there was the same problem - lack of feedback.
I'm writing this software for fun, because I really like programming and fractals. But you have to remember that it's also for YOU. If you want to have the software designed as you want, you have to provide a feedback. Despite everything I'm going to continue development of Mandelbulber but I'm not sure if FractalFormuls will be main platform to post announces about new releases. Title: Re: Mandelbulber v2 - 2.08 Post by: DarkBeam on July 04, 2016, 05:13:40 PM Question.
Is it possible to add a 3d perlin noise generator that can perturb the fractal formulas? For example the C addition constant... You should calculate it just once then it stays the same. There are fast premade codes that can generate 3d perlin for Fragmentarium. It should be very easy to port it. Congrats for the software Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 04, 2016, 05:41:47 PM Question. Is it possible to add a 3d perlin noise generator that can perturb the fractal formulas? For example the C addition constant... You should calculate it just once then it stays the same. There are fast premade codes that can generate 3d perlin for Fragmentarium. It should be very easy to port it. Congrats for the software Thanks for your input. It's interesting idea. It will be easily possible. I think it could be added as a first formula in hybrid sequence. Do you have any images which show expected results? I will look into Fragmentarium code how to do perlin noise. I had also a plan to use perlin nose as a surface texture or clouds (some kind of haze). Title: Re: Mandelbulber v2 - 2.08 Post by: quaz0r on July 04, 2016, 09:57:43 PM just curious, what are the websites which you have found more active than fractalforums?
Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 04, 2016, 10:28:25 PM just curious, what are the websites which you have found more active than fractalforums? Just DeviantArt, Facebook, G+ and direct email/phone feedback. Sites not fractal related, but people are also fascinated in fractals, and many of them are professional artists. For instant some of them use Mandelbulber for professional purposes like creating fulldome animations for planetariums. So that's why I'm wondering why fractal related service like FractalForums almost doesn't give me any feedback (the same was for MB3D). It's difficult to develop the software, especially cross-platform, when you event don't know if it works properly on different systems. Fortunately we are getting from FF many ideas about fractal formulas or algorithms which can be implemented into Mandelbulber. And fortunately I have zebastian and mclareking in the team, because they give me strong support in application development. Zebastan is helping me in adding different features like in last release it was Voxel Export, and mclarekin is adding a lot of new formulas and do a lot of testing. Title: Re: Mandelbulber v2 - 2.08 Post by: quaz0r on July 04, 2016, 11:20:18 PM ah yeah. it seems like maybe fractalforums has more people that are more interested in their own projects than participating in other people's projects. since i started my own programming project, i tend to spend my time working on that than trying to participate in other things. i havent looked much at mandelbulber2 code yet, but i remember one thought i had about mandelbulber was what the best way of vectorizing some of the code might be. i seem to recall in mandelbulber1 seeing you wrote your own matrix class, and i thought there are projects like eigen (http://eigen.tuxfamily.org/) which provide matrix classes, quaternion classes, and such things vectorized for various architectures. also for writing your own high-level vector code there are C++ libraries like Vc (https://github.com/VcDevel/Vc) which provides a very nice way to write abstracted, portable vector code. i use Vc in my program, and i quite like it.
Title: Re: Mandelbulber v2 - 2.08 Post by: DarkBeam on July 05, 2016, 12:10:50 AM as for perlin you can start to look here!
http://www.fractalforums.com/fragmentarium/high-performance-glsl-noise-functions/ :D Title: Re: Mandelbulber v2 - 2.08 Post by: taurus on July 05, 2016, 02:21:20 AM So that's why I'm wondering why fractal related service like FractalForums almost doesn't give me any feedback (the same was for MB3D). It's difficult to develop the software, especially cross-platform, when you event don't know if it works properly on different systems. Sorry to hear that. At the moment my time for experimenting is pretty limited. It would take months to convert my workflow to a new paradigm, especially in animation. Time I currently do not have. I am pretty familiar with the old GUI and mostly happy with the features, so it is hard for me to find a convincing reason for the change. I hope I can spare some time in near future, but I can't promise...Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 05, 2016, 05:03:56 AM Thanks guys for feedback, (albeit you three are not feedbacking newbies, having already helped us a lot in the past), and I fully understand the time constraints.
pause,.... then I ramble on ..... The latest version has only been released a few days, so in some ways it is a bit early for much feedback, other than "Material Manager looks great". I still need to test this feature out more myself, BUT, I do know that it is a powerful enhancement for Mandelbulber, and means to me that Mandelbulber is now finally functionally complete. :) :) :) @Buddhi, feedback from V2.07 .... I have only recently tried the Fresnel's equations for reflectance, and liked what I found. :beer: :beer: :beer: Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 05, 2016, 10:27:17 AM This is an example of MandelboxMenger hybrid, with some basics for newbies to the program.
1) In the example the MengerSponge part is run only on iteration 4. A single iteration of another fractal to make a hybrid is often the best practice. 2) In the Statistics (enable in View menu) you can see Percentage of Wrong Distance Estimations ("Bad DE") is 0, which is good!!. As a general rule less than 0.01 is good, but it is case specific and 3.0 sometimes is OK and .0001 sometimes is not. The Raymarching step multiplier or "fudge factor" (Rendering Engine tab) is set at 0.6, which is good for a hybrid. If I change it to 0.7 the Percentage of Bad DE leaps up to 0.25 and you can see the areas of quality loss on your image. Now if we disable the addCpixel Axis swap Constant Multiplier, we find we can now increase the Raymarching Step Multiplier to 0.9, and get a faster render and visually the same quality. So monitoring Percentage of Wrong Distance Estimations is a guide to managing quality. ( Note when doing animations you may want to drop the Raymarching step down a bit to allow for what might happen between keyframes.) 3) MandelboxMenger Hybrids can behave a bit differently to a lot of hybrids, in the fact that the Percentage Bad DE often improves when you zoom in. Code: # Mandelbulber settings file Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 06, 2016, 09:33:34 AM This example is using Msltoe's toroidal formula at the beginning of the loop with Benesi T1 PT as the main frcatal formula.
Here the Percentage Bad DE is 0.28, which normally is real bad, but here just a bit bad, I try and hide it with fog. It looks OK-ish but if I zoomed in it would look worse. Another issue is that the Percentage Bad DE is possibly more like 0.8 because only two thirds of the image is fractal. The other third distorts the percentage by being being recorded as good DE. (if i have this right?!) I think that probably the best use of Toroidal, is to hybrid with other mandelbulbs or julia bulbs. Often Force Delta DE is required (render engine tab) Code: # Mandelbulber settings file Title: Re: Mandelbulber v2 - 2.08 Post by: Caleidoscope on July 06, 2016, 02:19:52 PM I received a warning from SourgeForge that a new version was available of Mandelbulber v2 - 2.08,
http://www.graphic-unlimited.com/photos/1467809524_808898130.jpg Maybe it would be an idea to do the same here. That whenever there is an update or a new program all members that are interested receive an email with this news. And it would be nice when a thread about updates or new programs are discussed in a way that it stays in the spotlight so to speak. Perhaps a special section that is an eye catcher on the side to draw attention. It also would be nice (if possible) that members for example like me that don't understand all the ins and outs of a program could be more interested for these programs. It can be quite intimidating for members/newbies to listen ( 88) to the flood of technical details. I understand that you can't make a version for 'fractal dummies' ;) but maybe there is still a way to make it more attractive. It would be good for more attention and stimulate people to dare to use a program they don't understand or just for a bit. Whatever, I'm glad you fractal-wizards take the time and the effort to make these great programs for us to use. :beer: :beer: Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 06, 2016, 05:39:59 PM I received a warning from SourgeForge that a new version was available of Mandelbulber v2 - 2.08, http://www.graphic-unlimited.com/photos/1467809524_808898130.jpg Maybe it would be an idea to do the same here. That whenever there is an update or a new program all members that are interested receive an email with this news. And it would be nice when a thread about updates or new programs are discussed in a way that it stays in the spotlight so to speak. Perhaps a special section that is an eye catcher on the side to draw attention. It's good idea about mailing list. I have to find some solution how it could be managed, because actually I don't have any list of users. Maybe cKleinhuis can help us in this matter. I have also to think if I can implement this directly in the software to inform users if new version is released. The program could check if newer version is already on the SourceForge server. Title: Re: Mandelbulber v2 - 2.08 Post by: cKleinhuis on July 06, 2016, 06:23:16 PM theoretically it would be possible through some kind of webhook that is triggered through a master github branch release or something, BUT it would be some handwork needed, we do not really have a close connection to development servers, and i am not aware of a board plugin that could do such, informing users wont happen throughout the forums because sending of emails is always a pain, a notification system on a per user base is not known to me at the moment as well
we could implement some auto webhook callback that someone could connect tio github, but i am not aware of what others things might come out of it, creating a thread for a new official release is the only thing we can offer, and it is not too different from current method :( Title: Re: Mandelbulber v2 - 2.08 Post by: zebastian on July 06, 2016, 09:08:36 PM What about a classical mailing list at gnu:
https://lists.gnu.org/mailman/listinfo You can Register/unregister in the webpage of GNU (really easy and intuitive) then you can send a mail to mandelbulber@gnu.org and all subscribers will geht the message delivered. I am in a couple of mailing list and this system works Quite well. Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 07, 2016, 10:38:01 AM Example of using Transform_Menger Fold On this transform UI, the standard menger sponge formula is split into a start and end function. The simplest way to use this transform is in Hybrid Mode, having the menger fold transform in slots 1 & 3. In slot 2 place any linear type formula or transform. (ie more mengers, kifs, mboxes, amazing surf, folds, rotation , Benesi T1 etc). In slot 1 disable the stop function and in slot 3 disable the start function, resulting in a standard menger sponge with something in the middle. BTW in fact you can mix around with the start and stop functions have all enabled if you wish. Generally linear functions all work well together in making hybrids. In Statistics, maximum is 79 iterations. Generally hybrids take longer to render than standard formulas. As well as adjusting formula parameters, you can use the iteration controls to tweak hybrids. In this example the first slot is set to repeat for 5 iterations before moving to slot 2. Slot 2 is set to stop at iteration 12, whereas slots 1 & 3 can continue to default end iterations = 250 (but actually have met termination conditions (bailout) before iteration 79). Code: # Mandelbulber settings file Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 07, 2016, 11:15:34 AM In the example above, slot2 of the hybrid sequenced ended at iteration 12. 12 was chosen because how it fitted into the iteration sequence, as follows:
Slot1 x 5 0,1,2,3,4 iteration numbers (note first iteration is iteration number 0.) slot2 5 slot3 6 slot1 x 5 7,8,9,10.11 slot2 12 // last use of slot2 sequence continues slot1 x 5, slot3, .....to bailout. So slot2 is used only twice in the iteration process. If I had entered 11 instead of 12 for slot2's stop_iterations, then the slot would have been used only once, if I entered 19 then it would run three times. Title: Re: Mandelbulber v2 - 2.08 Post by: paigan0 on July 09, 2016, 02:15:21 AM Buddhi,
I've very much enjoyed Mandelbulber and I've been using it daily for over a year now, through several different versions. I've been struggling to figure out the OpenCL version, and I think I've got most of it figured out. I wish I could do more OpenCL fractals than just the Mandelboxes, Bulbs, and a couple of other minor algorithms. I set up a render farm with a couple of hundred cores and I've been producing about 1 CPU video and 3-4 GPU videos a month. I'm a multi-instrumentalist composer, writer, and producer and musician, and I wanted to make cool videos to go along with cool "soundtrack" music. I've put the two together and have done a bunch of "3d Fractal Animation and Music Videos" and post them on Youtube and better quality Vimeo at my channel called "Fractals." https://vimeo.com/channels/fractals CPU render farm: 215 cores GPU: 2 sets of GeForce 980s overclocked. I've never posted here to the forum because I didn't want to get in the way of the pros. But reading about how you didn't get much feedback, I can certainly do something about that! I've had a whole host of issues that I've resisted asking questions about and just tried to figure things out for myself. I'd love to give you more feedback. But just wanted to say Hi and thanks for this wonderful program. I've done almost 50 different 30 FPS videos averaging around 7000 frames per video or so. My early vids are primitive compared to finally learning how to use the focus effects, and textures. I can't wait to dig into more textures here in the new version. Thanks again for the awesome Mandelbulber software, both the CPU and GPU versions. And please I'd love it of the GPU version could do more fractals than just Mandelboxes and bulbs. It's just SOOOO much faster than brute force CPUs. Thanks again! Cheers! Stephen Sink paigon0 Paigan Productions Proud Chapman Stick player https://www.youtube.com/user/paigan0 https://soundcloud.com/stephen-sink-1 https://vimeo.com/channels/fractals Title: Re: Mandelbulber v2 - 2.08 Post by: taurus on July 09, 2016, 11:43:35 AM So, finally I had a little bit of testing. Here some observations:
positive I see the translation. The native german version is very convincingly translated. Maybe some things - ambient occlusion, just to give an example - should stay english, as they are more common as the translation. A serious issue, I had in the material editor. Placing the texture center is hard, if not next to impossible.
Annother issue I had with the slider for transparency of volume. it starts to jump around between two visible scales, when leaving 1 (the slider's length on screen jumps between two states). I attached a screenshot, to show what I mean. At this point I think it's time for a general note about the slider concept. For me, many of them (or should I say most) are useless, as they are not fine enough, have an unappropriate scale or simply do not act as expected. If this is an attempt, to create a m3d-like GUI, it is more than half-cooked. M3d users (and I am one) won't feel familiar with it - sorry to say that. All in all, the new material manager is a pretty good start. It acts like I am used to from professional 3d bundles like c4d or blender. Further features could be image maps for transparency or an alpha channel to stack more than one material. Besides that you're on the right way, I think. System: WIN 10 home Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 09, 2016, 12:56:18 PM Thank you @taurus. You gave me a lot of useful information. It's my first attempt to implement texture mapping, so it needs many improvements.
About positioning of textures, what you think about following solution: - Indicate point where top left corner should be placed - Then indicate top right corner to set horizontal scaling - Finally indicate bottom left corner to set vertical scaling Title: Re: Mandelbulber v2 - 2.08 Post by: taurus on July 09, 2016, 01:11:09 PM Thank you @taurus. You gave me a lot of useful information. It's my first attempt to implement texture mapping, so it needs many improvements. About positioning of textures, what you think about following solution: - Indicate point where top left corner should be placed - Then indicate top right corner to set horizontal scaling - Finally indicate bottom left corner to set vertical scaling Sounds great, if you leave the manual adjustment untouched. They are maybe useful for fine tuning. Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 09, 2016, 04:40:23 PM If anybody wants to report bugs or proposals of enhancements directly on our development site, you can add issues here: https://github.com/buddhi1980/mandelbulber2/issues
Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 12, 2016, 06:23:06 AM Example of MandelbulbKali. This makes good Julias.
Note: If you initially enable fabs() nothing will happen unless a parameter is used to map some point back into negative, e.g. use rotate about z_axis Code: # Mandelbulber settings file Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 13, 2016, 12:30:37 PM here is an example of AboxModKaliEiffie. I have yet to really explore this formula.
Code: # Mandelbulber settings file Title: Re: Mandelbulber v2 - 2.08 Post by: joe on July 14, 2016, 01:47:17 PM New version is amazing, a banquet of new features! I was stunned by how much was added to the previous version, but is this the most featured update yet?
I have yet to gain full understanding of how the hybrid system works, but I'm catching on faster than with the 1.x versions and liking it. The material editor is very good, I have been experimenting with textures with some success. A lot of ideas building. I think your 3 point placement would work, then it's down to the photoshopper's finesse to make it work. As potent a method as hybrids. It would be cool to see glow in the material editor, so primitives can be without the volumetric glow. Is that possible? I keep getting grey patches when using transparency, even with spheres etc. Maybe I don't yet understand what I'm looking at. There are two functions lacking, the only things keeping me from migrating completely to v2: one is the lack of PNG+alpha saved from keyframe animation. The other is minimum iteration. Both of them are required to make flame fractal animations - by the way these look epic in 3D. Here is an example. https://scontent.fakl1-1.fna.fbcdn.net/t31.0-8/12829115_10153585879029926_8868523938067441143_o.jpg (https://scontent.fakl1-1.fna.fbcdn.net/t31.0-8/12829115_10153585879029926_8868523938067441143_o.jpg) And here's a pic I made just now, using luminosity mapping for the fractal and diffusion and displacement mapping for the sphere. Nom nom nom. https://scontent.fakl1-1.fna.fbcdn.net/t31.0-8/13680237_10153885115579926_2930260561985397146_o.jpg (https://scontent.fakl1-1.fna.fbcdn.net/t31.0-8/13680237_10153885115579926_2930260561985397146_o.jpg) EDIT - how about distance threshold detail level per material? Or per formula. I've just discovered boolean... oh my. Yet another dream realised, thank you. I would like to see accurate light ray behaviour with transparent and reflective materials, so materials act well as lenses and mirrors for moving lights. How cool would it be to use prisms in a scene. Also a "virtual projector" would add huge value. A lot of ideas I have can be solved using luminosity maps, but currently it would be very difficult to recreate a scene where a projector shines an image on a group of objects. Should I keep adding to my wishlist in the feature requests? Title: Re: Mandelbulber v2 - 2.08 Post by: zebastian on July 15, 2016, 08:18:32 AM hi joe,
cool images, i like the texture on the sphere in your second one :beer: PNG+alpha is also available for animation, in fact the PNG+alpha is more of a deprecated thing and you should only work with the new image system: under Menu > File > Program Settings > Image Tab you will find all settings regarding image channels. Activate the "alpha" checkbox and the checkbox "append alpha to image". Now "save PNG / TIFF / JPG / EXR" will all honor these settings and try to save all image channels (in seperate files if necessary and with most fitting image quality). This will also make yur images saved by keyframe / flight animation be in this format. I know this is kind of confusing right now and we will work on this to make this more intuitive for the next version. Cheers Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 16, 2016, 11:16:22 PM Do you know when the OSX version of 2.08 will be available?
Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 17, 2016, 02:26:14 AM @ joe
Quote Should I keep adding to my wishlist in the feature requests? yep, thanks for feedback, all worth investigating O0 Thanks for remoning about mimimum iterations, a while back I needed that function for a new formula, then forgot about it. Quote I've just discovered boolean.. The potential of animation is incredible with the boolean feature O0 @ starmute I do not know, it is not compiled by Buddhi, Zebastian and me. And it is European vacation time and some team members are on vacation (buddhi!), probably lying about in the sun being served cold beers, meanwhile down here in the southern pacific, the rain is pouring, the wing is howling, thunder and lightning fill the skies, and it is deathly cold. The wild moas are roaming though the snow, breaking down the outer fences looking for food (they only eat humans, ). Have fun in the sun you lucky North Hemispherians. ;D Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 17, 2016, 10:34:13 AM With the MandelBox Menger UI we can disable the menger component and just work with the top half as a standard Mandelbox, BUT this mandelbox is set up with iteration controls for the Box Fold, Spherical Fold, Scale & Rotation transforms. So we can "late start" and "early finish" these functions.
ALSO the Box Fold transform has a z scale that morphs from an amazing surf to a full mandelbox shown in jpg below The code below is for a negative mandelbox with a near to zero z scale Code: slowpoke mcslowsauce ooops, better reload clipboard Code: # Mandelbulber settings file Title: Re: Mandelbulber v2 - 2.08 Post by: visual.bermarte on July 17, 2016, 04:52:32 PM Do you know when the OSX version of 2.08 will be available? Hi, I think next week, sorry but I was dealing with some complications. :embarrass:Title: Re: Mandelbulber v2 - 2.08 Post by: Snicker02 on July 18, 2016, 11:50:57 PM What do you think about new possibilities with texture mapping and materials in Mandelbulber? Now there are possible such a images: (http://nocache-nocookies.digitalgott.com/gallery/19/640_03_06_16_11_11_04.jpeg) How do you get different materials on different shapes? I just downloaded new version today and one issue I have seen if you choose certain materials the fractal just shows up black. Maybe I am doing something wrong?? I have shared the new release on FB in a couple groups I am in... I think a mailing list would be helpful Title: Re: Mandelbulber v2 - 2.08 Post by: Snicker02 on July 19, 2016, 02:08:27 AM A bug with cloud voxel output it puts the files in the folder above the chosen one
Title: Re: Mandelbulber v2 - 2.08 Post by: Snicker02 on July 19, 2016, 02:46:48 AM Also JWildfire can import the image sequences from the created voxel stacks please see this tutorial https://youtu.be/HP_zCyuMEao youtu.be/HP_zCyuMEao
Title: Re: Mandelbulber v2 - 2.08 Post by: Snicker02 on July 19, 2016, 02:49:25 AM https://www.youtube.com/watch?v=HP_zCyuMEao (https://www.youtube.com/watch?v=HP_zCyuMEao)
Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 19, 2016, 01:32:31 PM Another example of manipulating a Mbulb with one iteration of Msltoe Toroidal. Often need to find settings/locations where the DE calc quality is acceptable
Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 20, 2016, 10:10:43 AM This example is manipulating the surface of a mandelbox by inserting iterations of a menger sponge late in the calculation. Disable the menger checkBox and you can see what it looked like before.
Often you need to zoom in to see these effects. The menger influence is controlled by changing the iteration numbers, (and monitor the DE calculation statistics) . Various other fractals can be used late in the calculation to manipulate the main fractal without changing it's basic shape. Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 20, 2016, 09:00:54 PM Wow, mclarekin! That menger manipulation is fascinating. It might be good for one of my flights. How much longer than a normal Mandelbox does it take to render?
Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 21, 2016, 02:30:12 AM Don,t know I haven’t looked. OK, 67 seconds with the menger, 86 seconds without the menger
But it could be faster or slower with hybrids, it depends on such things as if the extra maths gets to bailout termination conditions faster or slower or the extra maths has expensive functions, (menger is just + and * functions so fast.) I zoomed into 8.695239557752599e-06, so the resulting image is completely different image (not similar like the previous example). With menger was 105 seconds without it was 146 seconds. So in these cases the hybrid is faster. Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 21, 2016, 03:10:35 PM Quote How do you get different materials on different shapes? I just downloaded new version today and one issue I have seen if you choose certain materials the fractal just shows up black. Maybe I am doing something wrong?? I have shared the new release on FB in a couple groups I am in... I think a mailing list would be helpful This is how I have been doing it. Rectangle at the bottom marked A. This is where you start a new material or load an existing. The active material is highlighted in blue. Meaning it is active in the material editor where you create or modify the material. Rectangle at top left marked B. One way to use a material is to go to Global Parameters, click on the material preview image, and the Material Manager UI will appear with the materials you have loaded or created. Click on the one you want to use, then close that UI. Similarly with primitives, click on the material preview image.And with Boolean Mode each fractal/transform has it's own material preview image when you scroll down. Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 23, 2016, 10:10:06 PM Hi. I have just returned from vacation, so I'm starting answering your questions and discuss proposals.
It would be cool to see glow in the material editor, so primitives can be without the volumetric glow. Is that possible? It should be possible, but needs a lot of changes in volumetric shader. But it's great idea to have glow parameters in material editor.I keep getting grey patches when using transparency, even with spheres etc. Maybe I don't yet understand what I'm looking at. If you provide any settings file (attach to post) where you observe this problem, then I will tell you what is wrong.how about distance threshold detail level per material? Or per formula. I'm not sure if this will be possible with actual ray-marching algorithm. I have to think how to define new conditions for surface detection.I would like to see accurate light ray behaviour with transparent and reflective materials, so materials act well as lenses and mirrors for moving lights. How cool would it be to use prisms in a scene. This is very interesting topic. I'm going to do this when I will implement OpenCL in Mandelbulber v2. These effects need much more computation power, so GPU support will be needed. However I'm going to check if this will be also possible using CPU.Also a "virtual projector" would add huge value. A lot of ideas I have can be solved using luminosity maps, but currently it would be very difficult to recreate a scene where a projector shines an image on a group of objects. Wow! You are very creative. This would be great effect especially with volumetric light effect. It should be quite easy to implement.Thank you joe for all inputs. Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 23, 2016, 10:11:21 PM Do you know when the OSX version of 2.08 will be available? OSX version is during preparation by @bermarte. Should be released within few days. Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 23, 2016, 10:13:59 PM A bug with cloud voxel output it puts the files in the folder above the chosen one Thanks for reporting it. We will fix this error. Title: Re: Mandelbulber v2 - 2.08 Post by: visual.bermarte on July 23, 2016, 11:04:36 PM OSX-release http://bit.ly/2alSPxJ :beer:
Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 23, 2016, 11:37:47 PM I am getting lots of errors "too many errors in settings file" when trying to open the attached fract file (from 2.07). This is on mac version of 2.08.
Title: Re: Mandelbulber v2 - 2.08 Post by: visual.bermarte on July 24, 2016, 12:00:11 AM I am getting lots of errors "too many errors in settings file" when trying to open the attached fract file (from 2.07). This is on mac version of 2.08. This is what I get finally past the errors. :hurt:Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 24, 2016, 12:05:38 AM That's what it's supposed to be, all the right colors and everything!
Do you have all 158 keyframes too? On my side it refuses to load any of them. That is my flight #20. :hurt: Title: Re: Mandelbulber v2 - 2.08 Post by: visual.bermarte on July 24, 2016, 12:24:13 AM Do you have all 158 keyframes too? On my side it refuses to load any of them. No, no keyframes at all, the cause of the trouble should be investigated and analyzed next week, I think. :/edit: @Starmute: sorry, I have chosen the wrong emoticon, better this one :hurt: Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 24, 2016, 12:28:59 AM No, no keyframes at all, the cause of the trouble should be investigated and analyzed next week, I think. :/ Thank you, I suppose this gives my computer time to render the rest of the flight... :snore:Is it just me or is rendering in 2.08 far faster than in 2.07? Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 24, 2016, 12:31:51 AM I think I have solved this problem, but if you have an answer please give it anyways :D By the way, switching between hybrids is very difficult on mac, you have to move them into #1 spot using arrows then move them back. Clicking on them does not work. Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 24, 2016, 01:29:39 AM @ starmute. Please send an example fract file of heavy noise.
Edit, Starmute you are correct, it is noticeably faster @ Joe, I think I know what you mean when using alpha PNG. There is still some enhancements to do for Materials Manager and this is one of the issues. Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 24, 2016, 02:03:48 AM I attached the noise sample in my first post but I will do it again. :)
Do you know about the problem with switching between hybrids? Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 24, 2016, 02:41:26 AM Noise, mixing linear type DE formulas with logarithmic is difficult
quick look at Noise 1) ray marching step multiplier of 1 is too high for these, often you are down to 0.1 2) Detail Level of 2 is too high for these hybrids 3) Look at percentage wrong DE calculations in statistic UI, generally you want 0.01 or smaller. 4) As the mbox is the main hybrid fractal in your example, you should probably use Force Delta DE Linear I am not 100% sure about the switching between hybrids in Mac OS. Could you give us a step by step desrctption of what is happening, I know that when using the arrows the status of the fractal enanble/ disable checkbox does not move with the rest of the formula. Thanks for the feedback, this helps a lot. A detailed explanation to users on how to use DE settings is needed. Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 24, 2016, 03:21:10 AM Thank you mclarekin.
About the switching between hybrids: whenever you click on a hybrid except number 1, it becomes semi darkened, but not actually selected. The fractal listed does not actually change. To actually change it you must move it back to number 1. Also - to actually see the end of each slider for fractals, you must stretch that side of the window to huge size because of the fact that there are 9 hybrids. Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 24, 2016, 03:29:12 AM One more thing, there is a lot of lag before rendering when using a textured background, like this one.
http://i.imgur.com/nVS9jVJ.jpg Probably at least one second delay after clicking in the fractal or pressing 'render'. Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 24, 2016, 03:47:32 AM @ starmute have you got Enable hybrid fractal "ticked" on the hybrid tab?
Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 24, 2016, 03:59:14 AM @ starmute have you got Enable hybrid fractal "ticked" on the hybrid tab? The problems happen regardless.Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 24, 2016, 04:12:38 AM Maybe supply a screenshot and example setting file, we may need the real programmers to look at this?
I don't have any problems in Ubuntu, next time I am in Windows V2.08 i will double check that it is working there. It may be Mac OS only. BTW, We are considering having enable checkBoxes on the main Fractal UI for enabling Hybrid and Boolean modes. Another possibility would be if more than one fractal slot is ticked , then the program would default to Hybrid Mode. What do you think? Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 24, 2016, 04:23:13 AM The nine fractals issue is annoying. I think one solution is to make the names shorter, definitely "Transform" should be shortened, and the use of Mbox and Mbulb would help, I have added this feedback to the list of issues to look at.
Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 24, 2016, 04:35:03 AM Here is a screenshot of the issue, in that example I just clicked on #2 but it stays on #1.
I have to stretch it about that far in order to see the entire slider. Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 24, 2016, 04:46:59 AM looks weird, I cannot see any checkboxes for ticking enabling the formula slot?
My screenshot is Ubuntu OS Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 24, 2016, 04:49:17 AM looks weird, I cannot see any checkboxes for ticking enabling the formula slot? Oh, that is of 2.07, sorry. I have both open right now and accidentally picked 2.07. But other than that the problem is the same. :)My screenshot is Ubuntu OS Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 24, 2016, 05:02:26 AM OK we have to wait until the Europeans wake up, we need the experts.
Title: Re: Mandelbulber v2 - 2.08 Post by: zebastian on July 24, 2016, 12:23:15 PM Hi guys,
thats an OSX UI bug, the tab-bar with the fractal names at the top should be scrollable like in linux or windows. OSX does not get it right and wants to show all tabs at once, this causes the whole window to be ridiculously wide. i dont know where the checkboxes go :) This will require some rework... Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 24, 2016, 09:10:06 PM One more thing, there is a lot of lag before rendering when using a textured background, like this one. http://i.imgur.com/nVS9jVJ.jpg Probably at least one second delay after clicking in the fractal or pressing 'render'. Do you know what is going on here, with the textured background lag? It happens even when none of the background is in view. Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 24, 2016, 09:25:16 PM Do you know what is going on here, with the textured background lag? It happens even when none of the background is in view. I have no idea where could be the problem. I have just tested it (with your texture) under Linux, Windows Vista and Windows 10. There is no lag. Maybe this problem exists under OSX. Bermarte needs to check it. By the way, background texture is loaded from disk every time when you click render button, so it can take some time on slower machines. Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 24, 2016, 09:29:35 PM I have no idea where could be the problem. I have just tested it (with your texture) under Linux, Windows Vista and Windows 10. There is no lag. Maybe this problem exists under OSX. Bermarte needs to check it. Thank you buddhi. Another bug:By the way, background texture is loaded from disk every time when you click render button, so it can take some time on slower machines. Materials do not form a grid on mac, they are all in a line, so you have to scroll all the way to the right. Attached image shows the problem. And is it possible to smoothly interpolate materials in 2.08 (keyframe animation), without all of the hassle of changing each individual property? Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 24, 2016, 09:53:30 PM Thank you buddhi. Another bug: Materials do not form a grid on mac, they are all in a line, so you have to scroll all the way to the right. Attached image shows the problem. And is it possible to smoothly interpolate materials in 2.08 (keyframe animation), without all of the hassle of changing each individual property? Actually materials are displayed in one line, not a grid. This is not a bug but something to improve. Because of some another problem under OSX they are displayed without names (I hope to fix this). About interpolation, there is not possible to interpolate between different materials. It is difficult to implement (but should be possible). But you can interpolate each parameter of material. Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 24, 2016, 10:45:20 PM Actually materials are displayed in one line, not a grid. This is not a bug but something to improve. Because of some another problem under OSX they are displayed without names (I hope to fix this). How about having multiple materials, with a weight value for each one (like is done with hybrids)? That might make it easier to implement.About interpolation, there is not possible to interpolate between different materials. It is difficult to implement (but should be possible). But you can interpolate each parameter of material. Title: Re: Mandelbulber v2 - 2.08 Post by: Snicker02 on July 25, 2016, 02:22:00 AM @Buddhi and everyone else... Speaking of Facebook did you know there is a Mandelbubler group? https://www.facebook.com/groups/mandelbulber/ (https://www.facebook.com/groups/mandelbulber/)
Title: Re: Mandelbulber v2 - 2.08 Post by: paigan0 on July 25, 2016, 03:11:46 AM @Buddhi and everyone else... Speaking of Facebook did you know there is a Mandelbubler group? https://www.facebook.com/groups/mandelbulber/ (https://www.facebook.com/groups/mandelbulber/) Facebook Mandelbulber group! Now I can bug people on Facebook with my fractals and 3d fractal animation music videos that actually care! :beer:Title: Re: Mandelbulber v2 - 2.08 Post by: joe on July 25, 2016, 03:35:16 PM zebastian, I had forgotten about the program settings. Major upgrade. zbuffer export was one of the main reasons I was going to learn M3D... was ;D
Buddhi I'm happy to hear the virtual projector is possible, music videos and art manipulation will open right up. New functions like this will make mandelbulber sing. It would be a powerful feature if video could be used for background and texture mapping media, by synchronising the video frame by frame with the keyframe animation output. It could be done now, but I would have to generate a settings file for each frame... I don't like spreadsheets that much. Here is the transparency issue: (http://nocache-nocookies.digitalgott.com/gallery/19/9484_25_07_16_3_10_58.jpeg) Code: # Mandelbulber settings file By the way, I do want to request a much higher limit for reflections depth, 100 would be a start, 1000 good. Possible? I just had a thought, it would be nice to include primitives in boolean mode. This would also solve my request for spherical limits. Simple functions are awesome because they can be well understood. Cheers Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 25, 2016, 05:41:11 PM Buddhi I'm happy to hear the virtual projector is possible, music videos and art manipulation will open right up. New functions like this will make mandelbulber sing. It would be a powerful feature if video could be used for background and texture mapping media, by synchronising the video frame by frame with the keyframe animation output. It could be done now, but I would have to generate a settings file for each frame... I don't like spreadsheets that much. Here is the transparency issue: (http://nocache-nocookies.digitalgott.com/gallery/19/9484_25_07_16_3_10_58.jpeg) Notice the grey area on the right of the globe. Almost any amount of fine detail results in this, is it to do with reflections depth? By the way, I do want to request a much higher limit for reflections depth, 100 would be a start, 1000 good. Possible? I just had a thought, it would be nice to include primitives in boolean mode. This would also solve my request for spherical limits. Simple functions are awesome because they can be well understood. Cheers About your issue with transparency it was easy to solve. You forgot to set reflectance of material. Glass has to refract the light (transparency) and also reflect the light. Without reflectance there was visible inner surface which you set to light grey (it's just surface color). To the program there is attached example glass material (use Load material option) where you can see how it can be defined. I had also changed maxiter (limit of maximum number of iterations). It was 250, now is 25. It enough to provide good quality and speeds up raymarching inside transparent object where there is always needed to calculate maximum number of iterations (like inside Mandelbrot Set) Try these settings: Code: # Mandelbulber settings file About animated backgrounds and textures I have this in my mind. I need to look for some easy solutions to implement it. I think it should be enough to be able to use sequences of JPG or PNG files (it's easy to export any mpeg/avi animation to sequence of jpg frames) About your request to increase maximum allowable reflections depth. I do not agree here, because it's not reasonable (and horribly slow for rendering) to use higher number than 10. It will only produce heavy noise and could lead to big numeric errors (even generation of Not-A-Number). If you give me any example where it could be used, I will change my mind. About booleans for primitives I also though about that. It would be possible to implement it but there will be needed to develop some nice UI for it. Mayby in v2.10. Instead of that I can easily implement spherical limits. Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 25, 2016, 05:41:54 PM @Buddhi and everyone else... Speaking of Facebook did you know there is a Mandelbubler group? https://www.facebook.com/groups/mandelbulber/ (https://www.facebook.com/groups/mandelbulber/) Wow! It's great to see that there is such a group on Facebook. Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 25, 2016, 07:18:05 PM Another bug, it is not possible to zoom to any percentage that isn't 25, 50, 100, etc. 75% does not work.
Buddhi, have you considered making it possible to use multiple materials with different weight for each one? That would make it much easier to interpolate materials. Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 25, 2016, 08:14:39 PM Another bug, it is not possible to zoom to any percentage that isn't 25, 50, 100, etc. 75% does not work. Buddhi, have you considered making it possible to use multiple materials with different weight for each one? That would make it much easier to interpolate materials. About zooming of image, I have never implemented another zoom values than those predefined. I though that they would be enough especially when when you have option "Fit to window". But if you think that it's beneficial to have another zooms, I can add it to our TODO list. About multiple materials I have already considered it. It would be good to have possibility to assign several materials as a layers which will use alpha channel of textures to control overlapping. Then weight (opacity) option would be used for interpolation. But it's quite a lot programming work, so it won't be soon. Title: Re: Mandelbulber v2 - 2.08 Post by: zebastian on July 25, 2016, 08:59:29 PM @Snicker02 sorry for my late reply.
There was a issue in the voxel export which caused the folder name to be concated, when there was no trailing folder separator: so folder name "my/great/location" saved all files to: my/great/locationlayer_00000.png my/great/locationlayer_00001.png ... which is actually in folder "my/great" and has "locationlayer_00001.png" as file name. My bad. I changed it, so in new version the folder gets parsed correctly no matter if specified with trailing separator or not. The program will also check if folder exists and give a warning if not. For the time being you should be fine to type the missing trailing folder separator by hand into the textbox. Here is the fix: https://github.com/buddhi1980/mandelbulber2/commit/6496e838fc9edb38ac4450d48cd0df28d3eb630b Cheers! Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 26, 2016, 07:57:05 PM Buddhi, would you be able to add another digit to the box folding of kaleidoscopic IFS? It seems to me that this would be better with 3 decimal places than with 2.
Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 27, 2016, 01:00:02 AM @starmute. Have updated UI, ( commit 1865902) and thanks. There is other places that need extra decimals that you might notice as well, probably mainly pre V2.05. These days I generally lazily use a default of 6 decimals, and promise myself that one day I will come back and shorten if necessary.
Keep the feedback coming. :beer: :beer: :beer: Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 27, 2016, 07:02:02 AM @starmute. Have updated UI, ( commit 1865902) and thanks. There is other places that need extra decimals that you might notice as well, probably mainly pre V2.05. These days I generally lazily use a default of 6 decimals, and promise myself that one day I will come back and shorten if necessary. Why are the decimals limited in the first place? I recall in 1.xx that they were unlimited. Is it a limitation of the Qt interface?Keep the feedback coming. :beer: :beer: :beer: Also, have you considered adding a means of sending crash reports? It might be very useful in the long run. :) Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 27, 2016, 09:30:07 AM I can only assume Buddhi was trying to keep the UI's small, as it is not an issue in QT. A lot of what I have coded has 6 decimals because I don't really know when a parameter might become "decimal place critical" under hybrid conditions. I have sometimes had to tweak the sixth decimal place. For the same reason you will find that even though a slider may only go thru 0 to 10 (an assumed normal operating range) my newer spinboxes will allow , maybe -99 to 99 to be entered.
I remember a parameter where all the major change happened between 0 and 10, but if I zoomed really deep I would still see a noticable change between 98 and 99. All of the sliders and spinboxes should probably be checked, when time and inclination allows. But iin the meantime if you find any that needs changing, we can make sure they get updated before the next release. BTW. I have never had time to properly test most of the new formulas, I have little idea of what they are capable of doing or what the optimum normal working range is for parameters, I get easily sidetracked onto the next tempting formula in the pile, (and it is still an ever increasing pile :o.) Quote Also, have you considered adding a means of sending crash reports? It might be very useful in the long run This is for Buddhi or Zebastain to answer. Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 28, 2016, 07:56:13 PM After further testing it seems that the box folding of IFS needs even more than 3 digits. 5 decimal places seems like the right number. :)
Also, is it possible to allow negative placement distance of lights? I often want to do this but am limited by the software. Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 28, 2016, 08:55:45 PM Also, is it possible to allow negative placement distance of lights? I often want to do this but am limited by the software. As I understand you want to place the light behind the fractal? Is it correct? I have never thought about it, but it would be easy to implement and also beneficial. It will allow backlighting in easy way. I'm adding it to our TODO list: https://github.com/buddhi1980/mandelbulber2/issues/117 Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 28, 2016, 10:13:46 PM I am getting lots of errors "too many errors in settings file" when trying to open the attached fract file (from 2.07). This is on mac version of 2.08. I have found where was the error. There was lack of conversion of material parameters when old settings were loaded. I have just finished writing and testing fix for this. I hope to release Mandelbulber 2.08-2 by the end of next week. It will include all fixes for bugs reported last days. You can observe our work progress here: https://github.com/buddhi1980/mandelbulber2/commits/2.08-2-bugfixes Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 29, 2016, 02:50:34 AM .
Quote After further testing it seems that the box folding of IFS needs even more than 3 digits. 5 decimal places seems like the right number. Increased scale, offset and box fold to 6 decimal places. Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 29, 2016, 03:20:06 AM As I understand you want to place the light behind the fractal? Is it correct? Yes, that's exactly what I want. Thanks!I have never thought about it, but it would be easy to implement and also beneficial. It will allow backlighting in easy way. I'm adding it to our TODO list: https://github.com/buddhi1980/mandelbulber2/issues/117 Also, would it be possible to split volumetric lighting into 2 parameters - 'range' (how fast it decays) and 'brightness'? This would make it much easier and much less frustrating to use. . Great! That is good news.Increased scale, offset and box fold to 6 decimal places. Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on July 29, 2016, 03:49:56 AM I just wonder what other parameters need adjusting. It would be helpful that whenever we are in a deep zoom situation, to test the last decimal on the various spinBoxes and note if any are still making noticeable change. Scale and spherical minR I know for sure can be very sensitive deep down. And that reminds me to also check rotations.
Thanks again Starmute for the feedback, it sure gets issues found and sorted out quicker. Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 29, 2016, 04:40:04 AM I just wonder what other parameters need adjusting. It would be helpful that whenever we are in a deep zoom situation, to test the last decimal on the various spinBoxes and note if any are still making noticeable change. Scale and spherical minR I know for sure can be very sensitive deep down. And that reminds me to also check rotations. I will make sure to test the last decimals in my flight #21. :DThanks again Starmute for the feedback, it sure gets issues found and sorted out quicker. Title: Re: Mandelbulber v2 - 2.08 Post by: taurus on July 29, 2016, 12:11:07 PM I just wonder what other parameters need adjusting. It would be helpful that whenever we are in a deep zoom situation, to test the last decimal on the various spinBoxes and note if any are still making noticeable change. Scale and spherical minR I know for sure can be very sensitive deep down. And that reminds me to also check rotations. Thanks again Starmute for the feedback, it sure gets issues found and sorted out quicker. Good thing. I would have suggested that myself. I think most parameters should not be less sensitive than 6 digits. Some of them - the texture placement for example - need the full double float accuracy to make sense in deep zoom situations. I lately experimented with mandelbox parameters close to -1. And the limit of 500 iterations was an unmanagable barrier there, so I returned to v1 to avoid. In my special case, I needed 900 Iters (found by trial and error) so I would suggest to increase that limit to at least 1000. thanks. Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 29, 2016, 08:20:08 PM I lately experimented with mandelbox parameters close to -1. And the limit of 500 iterations was an unmanagable barrier there, so I returned to v1 to avoid. In my special case, I needed 900 Iters (found by trial and error) so I would suggest to increase that limit to at least 1000. thanks. In v2.08-2 you will have limit 99999 for maximum number of iteration. I have jus changed it. I have no idea why it was only 500. Probably it was something which was taken from old Mandelbulber. Thanks for this suggestion. Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 29, 2016, 09:52:52 PM Buddhi, did you see what I asked about volumetric lighting? Is it possible?
Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 29, 2016, 09:58:46 PM Buddhi, did you see what I asked about volumetric lighting? Is it possible? Yes, it's possible. I will do this change in 2.09. Title: Re: Mandelbulber v2 - 2.08 Post by: taurus on July 29, 2016, 11:53:18 PM I have no idea why it was only 500. Probably it was something which was taken from old Mandelbulber. At least not from 1.12 and 1.21 which I still have installed. Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 30, 2016, 01:51:39 AM Also, is it possible to add 'remaining time' for both single image and entire render, like was done in 1.xx?
Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on July 30, 2016, 09:41:25 AM Also, is it possible to add 'remaining time' for both single image and entire render, like was done in 1.xx? I'm not sure what you mean. There is already double progress bar (see attached image). Top one is for total progress of rendering of animation. Bottom one is for single image. Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on July 30, 2016, 08:01:27 PM I'm not sure what you mean. There is already double progress bar (see attached image). Top one is for total progress of rendering of animation. Bottom one is for single image. It is not there on mac. I see the progress bars but no indication of time remaining.Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on August 02, 2016, 07:09:38 AM Quote Is it just me or is rendering in 2.08 far faster than in 2.07? @ starmute. You are correct, some parts are far faster than 2.07 ;D Since you mentioned it, I have noticed some increase. Example a 3200 x 2400 that previously took 6 minutes now takes around 3 minutes. This should greatly improve doing animation. Title: Re: Mandelbulber v2 - 2.08 Post by: visual.bermarte on August 02, 2016, 03:21:23 PM It is not there on mac. I see the progress bars but no indication of time remaining. Here is a screenshot of the issue, in that example I just clicked on #2 but it stays on #1. The two matters, the one relative to the lack of indication of time on the progress bar and the one on the scroll area that is not scrollable like it should, appear only when the default mac skin is selected, otherwise there's no issue at all. Those troubles disappear when selecting another skin.I have to stretch it about that far in order to see the entire slider. Strictly speaking the first is not a bug. Macintosh widget style (http://doc.qt.io/qt-5/images/macintosh-progressbar.png) Fusion style (http://doc.qt.io/qt-5/images/fusion-progressbar.png) Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on August 02, 2016, 07:08:38 PM Edit: Sorry, I am a fool, I have found the program settings and changed the skin. :) Also, I have a few more bugs to report: - Any light with intensity 0 causes all other lights to turn off. - Any light with negative intensity causes image not to render at all. This is a serious problem when trying to interpolate light intensities using catmull-rom or akima. Also - any updates on when new version of 2.08 will be released, with the ability to load 2.07 files (like the one I uploaded before?) Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on August 02, 2016, 10:15:58 PM Another bug to report, when main_mat1_coloring_speed and main_mat1_coloring_palette_offset are added into keyframes, they do not load properly upon loading a settings file. All I get are white boxes in place of actual numbers.
Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on August 03, 2016, 11:54:35 AM This is an example of zooming in past the debris. Looks terrible when zoomed out, but you can get areas of excellent DE quality when zoomed in.
Title: Re: Mandelbulber v2 - 2.08 Post by: visual.bermarte on August 03, 2016, 04:19:18 PM This annoying issue is somehow fixed too. ^-^
Quote http://www.fractalforums.com/releases-b233/mandelbulber-v2-2-08/msg94786/#msg94786 Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on August 03, 2016, 08:01:57 PM Another bug to report, when main_mat1_coloring_speed and main_mat1_coloring_palette_offset are added into keyframes, they do not load properly upon loading a settings file. All I get are white boxes in place of actual numbers. I have already fixed this bug. I'm going to release new packages for Linux and Windows today. I'm now sure when @bermarte will be able to release it for OSX. All known bugs for OSX are also fixed, so now it's only a packaging work. Title: Re: Mandelbulber v2 - 2.08-2 - bug fixes Post by: Buddhi on August 03, 2016, 10:51:31 PM Bug fixes are released for Linux and Windows. Mandelbulber 2.08-2 is ready for you.
Thanks to your very strong feedback there was possible to eliminate many problems and increase performance of application. Mandelbulber 2.08-2 renders images even twice faster than 2.07 (speedup between 10 and 160%, depending on what is rendered). On my PC rendering time of default image is reduced from 2.6s to 1.1s. Note: to get good performance under Linux there is recommended to call following command before compilation: export CXXFLAGS="-march=native -msse2" I have added this note because some people complain why Linux version is slower than for Windows. This command forces compiler to optimize application just for your computer. New version can be downloaded as always from following locations: http://sourceforge.net/projects/mandelbulber/ (http://sourceforge.net/projects/mandelbulber/) https://github.com/buddhi1980/mandelbulber2 (https://github.com/buddhi1980/mandelbulber2) release 2.08-2 includes following changes: - Settings: fixed bug: importing of old settings didn't worked properly when rotations in IFS fractals were used. - Settings: fixed bug: loading old animations (lack of material conversion) - UI: Fixed bug: Some sliders are jumping under Win 10. There is changed size policy for edit fields. - UI: Corrected calculation of statistics - Shaders: Corrected positioning of textures. Now position is calculated based on texture coordinates (before it was based on object coordinates) - Shaders: Program crashed when texture scale was 0 - Fractals: corrected number of decimal places in spinboxes for IFS - Fractals: Increased limit for maximum number of iterations to 99999 - NetRender: Fixed bug with loading textures from QByteArray when NetRender was used on different systems. - Performance: optimized Random() function and usage of cLights::GetLight(). Speed increase about 10-40% - OSX: fixed problem with missing text on progress bar - Voxel export: fixed bug: slices was saved in parent folder of indicated folder. Thank you all for very good support. Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on August 03, 2016, 11:35:40 PM That is great, Buddhi. Now we just need the OSX version from Bermarte. :)
Buddhi, what about this bug? Has it been fixed? - Any light with intensity 0 causes all other lights to turn off. - Any light with negative intensity causes image not to render at all. This is a serious problem when trying to interpolate light intensities using catmull-rom or akima. Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on August 04, 2016, 11:05:06 AM Quote Note: to get good performance under Linux there is recommended to call following command before compilation: export CXXFLAGS="-march=native -msse2" for people like me this means go to Home directory, open file .bashrc with a text editor paste in export CXXFLAGS="-march=native -msse2" as the new last line in the file, and save I have done some testing on about 20 old V2.07 setting files ,and rendering for me was on average 3 times faster, the best was 5 times faster, on V2.08 after inserting the line of code.. Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on August 04, 2016, 11:28:58 AM Buddhi, what about this bug? Has it been fixed? - Any light with intensity 0 causes all other lights to turn off. - Any light with negative intensity causes image not to render at all. This is a serious problem when trying to interpolate light intensities using catmull-rom or akima. I have missed this error. That's why I recommended to use Github to report errors (https://github.com/buddhi1980/mandelbulber2/issues (https://github.com/buddhi1980/mandelbulber2/issues)). On this site we have bug tracking. I will fix this and release fix in few days. Title: Re: Mandelbulber v2 - 2.08 Post by: zebastian on August 04, 2016, 07:48:19 PM negative intensity should be prohibited in the first place (since it does not make sense)
that the other lights are not showing, when one light has 0 intensity is an issue to address... Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on August 04, 2016, 08:20:59 PM negative intensity should be prohibited in the first place (since it does not make sense) Correct, it should be prohibited, but it frequently happens anyways when using catmull-rom or akima interpolation (because they go over/under a bit.)that the other lights are not showing, when one light has 0 intensity is an issue to address... Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on August 04, 2016, 08:36:59 PM Correct, it should be prohibited, but it frequently happens anyways when using catmull-rom or akima interpolation (because they go over/under a bit.) I have already found the error and solution. I'm checking if 2.08-2 is working properly (and waiting for feedback). If won't be any new bugs reported then I will release 2.08-3 during the weekend. Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on August 04, 2016, 10:07:33 PM I have already found the error and solution. I'm checking if 2.08-2 is working properly (and waiting for feedback). If won't be any new bugs reported then I will release 2.08-3 during the weekend. Okay. Do you have any word yet from bermarte on when the OSX version will be available?Title: Re: Mandelbulber v2 - 2.08 Post by: visual.bermarte on August 05, 2016, 01:08:47 PM Okay. Do you have any word yet from bermarte on when the OSX version will be available? http://bit.ly/2aD15uy :beer:Title: Re: Mandelbulber v2 - 2.08 Post by: taurus on August 05, 2016, 06:13:39 PM Thanks for the new version. Nice to see, that also the feedback situation is gettig better. In principle the new version works fine.
I encountered a redraw problem, I already saw some times in previous versions. When triggered, the fractal does not redraw anymore, until you resize the window or change to annother window, regardless of the status (still rendering or finished). Unfortunately I was not able to reproduce it. It seems to appear unregularly. So, don't take it as reason to delay the next build. It is not that serious and can be solved by restarting the program. My System: win 10 home. Keep up! ;D Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on August 05, 2016, 06:54:41 PM http://bit.ly/2aD15uy :beer: THANKS! :DI can finally get my flight #20 finished faster now. Title: Re: Mandelbulber v2 - 2.08 Post by: mclarekin on August 06, 2016, 01:16:24 AM @ taurus, can you remember with what version you first noticed this problem. Also similarly the progress bar can have the same problem. We have been monitoring this non-critical problem but have yet to determine what causes it.
@ starmute. I am looking forward to viewing it :) Title: Re: Mandelbulber v2 - 2.08 Post by: Starmute on August 06, 2016, 05:14:15 AM Mclarekin, I think Taurus' problem has occurred with me too. It seems to happen when clicking the 'rotate' buttons rapidly.
I have a couple non-essential suggestions: - Allow multiple Mandelbulber windows, and setting weight for each one. So if window A has weight 4 and window B has weight 1, A will use up 80% of rendering time given to Mandelbulber, and B will use 20%. This is useful for when I want to render a flight and an image at the same time. - Allow user to import a music file, and show the waveform along with keyframes. This is useful with varying framerate between keyframes (another feature that will be added). Might be useful to allow playing the music file as well. Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on August 06, 2016, 09:00:14 AM I encountered a redraw problem, I already saw some times in previous versions. When triggered, the fractal does not redraw anymore, until you resize the window or change to annother window, regardless of the status (still rendering or finished). Unfortunately I was not able to reproduce it. It seems to appear unregularly. So, don't take it as reason to delay the next build. It is not that serious and can be solved by restarting the program. My System: win 10 home. I already know this problem. It happens not so often, and unfortunately is very difficult to reproduce (no any rule when it happens). Actually I have no clue why it happens, but it looks like something which blocks processing of Qt events (signal / slot mechanism). I have a couple non-essential suggestions: - Allow multiple Mandelbulber windows, and setting weight for each one. So if window A has weight 4 and window B has weight 1, A will use up 80% of rendering time given to Mandelbulber, and B will use 20%. This is useful for when I want to render a flight and an image at the same time. There are system solutions for this. You can launch two Mandelbulber programs and set for them different system priorities. Under Windows you can change program priority in Task Manager. Under Linux for instant you can use 'nice' command to launch the program. I would not implement this solution into the program, because handling of thread or program priorities is system dependent. Until the program is multi-platform I'm avoiding system dependent solutions. - Allow user to import a music file, and show the waveform along with keyframes. This is useful with varying framerate between keyframes (another feature that will be added). Might be useful to allow playing the music file as well. I have this solution in my head. I'm going to implement this when I will do 'animation by sound'. Then I can also do this what you mentioned, just to edit animation together with waveform view with variable frames/keyframeTitle: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on August 07, 2016, 05:01:33 PM New bug fix is released.
There was fixed bug reported by Starmute: image was black when intensity of any light source was zero. It affected rendering of keyframe animation where intensity of the lights was animated. Mandelbulber 2.08-3 can be downloaded as always from following locations: http://sourceforge.net/projects/mandelbulber/ (http://sourceforge.net/projects/mandelbulber/) https://github.com/buddhi1980/mandelbulber2 (https://github.com/buddhi1980/mandelbulber2) Title: Re: Mandelbulber v2 - 2.08 Post by: paigan0 on August 07, 2016, 09:15:22 PM I already know this problem. It happens not so often, and unfortunately is very difficult to reproduce (no any rule when it happens). Actually I have no clue why it happens, but it looks like something which blocks processing of Qt events (signal / slot mechanism). I've also had the redraw problem on occasion exactly as taurus described. Also the progress bar will disappear somewhat rarely. Both are fixed by restarting the program. I also feel it happens when I'm clicking too fast in the view and the system tries to redraw things to keep up.[...] I have this solution in my head. I'm going to implement this when I will do 'animation by sound'. Then I can also do this what you mentioned, just to edit animation together with waveform view with variable frames/keyframe Also, since my main schtick is putting music (my own) to fractal animations and the reverse, I'd love to see where I could more directly control one versus the other. Currently, I can use Virtual Dub to make the frames per second of a video file match the wave file, so the two will be the same length, and then manually change the FPS back to 30, and Virtual Dub will handle the rest (I believe by inserting frames (doubling or tripling them) and repeating or dropping them automatically). But I'd love, love, love to be able to change the frame rate between keyframes, and also sync the music more closely with the fractal animation. And use music as feeder data to directly manipulate parameters in real time! ;D Great work making Mandelbulber better guys! (Although I've been using v 1.21 OpenCl lately almost exclusively because of the formulas mclarekin keeps porting over for me!) :beer: :beer: :beer: Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on August 15, 2016, 09:44:43 PM The other is minimum iteration. Both of them are required to make flame fractal animations - by the way these look epic in 3D. Here is an example. https://scontent.fakl1-1.fna.fbcdn.net/t31.0-8/12829115_10153585879029926_8868523938067441143_o.jpg (https://scontent.fakl1-1.fna.fbcdn.net/t31.0-8/12829115_10153585879029926_8868523938067441143_o.jpg) It's still possible to get this kind of flame fractals. You need to enable option "Stop at maximum iteration" on "Rendering engine" tab and set high value for "Maximum number of fractal iterations" (above 100). Then you need to set high intensity of glow effect. Try this: Code: # Mandelbulber settings file Title: Re: Mandelbulber v2 - 2.08 Post by: paigan0 on August 28, 2016, 07:00:15 PM I'd put this into the feature request section, but this thread seems to get more notice. One request for the Netrender module: we can see the names, addresses, and number of cores of each networked computer, but the PC that is the host server (which also renders lines) is not shown. Could we add the cores and lines rendered by the host server machine as well? My host machine has a 16-core 5960X chip that eats through the lines twice as fast as my nearest i7, a Dell 4790. But the PC doing most of the work isn't shown because it's also hosting the party.
Also, my standard request (prayer) for OpenCL support! :dink: Video cards get more and more powerful everyday, and the new 1080 is out. I know that might have to wait for a much later version. Cheers, and thanks, Krzysztof and all! Title: Re: Mandelbulber v2 - 2.08 Post by: joe on September 02, 2016, 07:16:19 AM Thank you, upgrade again! I knew there was something I was missing. Back to flame fractals I go.
I'm still not sure how the colouring works with glow. Would it be possible to introduce more colours? A gradient for each colour? Title: Re: Mandelbulber v2 - 2.08 Post by: paigan0 on September 07, 2016, 06:30:46 PM On the subject of future (and past versions) support for OpenCL, I made an exciting discovery this morning. I've never been able to get Mandelbulber v 1.21 to use 2 video cards at the same time on the same computer, even if they were SLI'd together.
I accidentally loaded up 2 instances of Mandelbulber v.1.21 at a time--which I've done before, and quickly closed the other windows--but I never realized that the two instances could run just fine next to each other. I used to have 2 GeForce NVDIA MSI 980s installed, one in two different computers. I used to have both 980s in the same PC, and I SLI'd them together, but Mandelbulber would only ever use one at a time. I could switch from one to the other card, but never get them both to work at the same time on the same PC. I finally started two instances of Mandelbulber Open CL and started one normally, but started the other one using the Device Index dropdown on the OpenCL tab and I changed the 2nd instance to 1 and kept the first at zero. I loaded up the exact same file and keyframes and everything on the 2nd MDBR program and hit rend animation. THE TWO VIDEO CARDS ARE RENDERING PERFECTLY NEXT TO EACH OTHER ON THE SAME PC!!!! (I'm pretty excited here, you might notice!) I have the 1080 rendering away on a pc by itself, and then the PC with the 2 980s has two separate instances of Mandelbulber running as well. That's a 980 and a 980 and a 1080 all making frames!!!! That was probably the intent behind that dropdown select menu, but I've never seen any support whatsoever for any of the features on the OpenCL tab. Trial and error is the only way I've figured anything out so far, other than mclarekin's recent help. Just wanted to share that multi-GPU knowledge, unless I'm the only person in the world that used Mandelbulber v 1.21 that didn't already know that! Cheers! Title: Re: Mandelbulber v2 - 2.08 Post by: Buddhi on October 04, 2016, 08:38:00 PM Mandelbulber 2.09 is released! http://www.fractalforums.com/releases-b233/mandelbulber-v2-2-09/ |