Extension of quaternion julia sets in fragmentarium

[SamTiba]

Hello!

I've been working with Julia sets for a while now, mainly 2D.
Got nice images and thought about extending it to 3D via fragmentarium and finally tried it.
Unfortunately I am really new to OpenGsl and can't figure out every Error I get, but since I really think we can get some new interesting shapes out of it, it may be in interest to help me out.

Already got some Matlab 3D-Isosurfaces working but they .... they just suck, computation time is high since I have no raytracing there and the results are blurry.

So now lets get to buiseness:

I work with Quaternion Julia-Sets with a different Bailout-Algorithm to determine if a point is part of the set or not. Unfortunately there is no distance estimation avaiable for this method so i have to Bruteforce it, in addition to that the sets are really really thin and can sometimes fill the whole plane.
But I guess (and hope) we can work around that!

The Algorithm works for every Continous Formula so it will extend the basic known quaternion julia sets.

Attached is what I have for now, basic working code, still very clumsy.
I would really appreciate if someone could help me or give me some advices (direct advices please, I can find and read helppages by myself).

[Sabine]

Hello SamTiba,


No way I can help you with your mathematical quest, but you will get your code to do something if you change your last line return (r<Bailout); to return (arc<Bailout);. As it says in the log area, r is not defined. Though I have no clue if you really want to return arc;)

You get this error line (yellow in log area) 0(430) : error C1008: undefined variable "r" The 430 is the line number of the error. If you click on Render up in the menu, then on Output preprocessed script, you can look up line 430 (3dickulus' version has linenumbers, you might need to enable them in the application preferences first).

Hope this helps you on your way until someone steps up that really knows what he/she's doing  

[SamTiba]

Thanks for your quick response, that was actually the last compiling-error that I missed, it is working now but there is no result yet.
I have to return (arc<Bailout) so the raytracer knows if its inside or outside our fractal.

Guess I really have to take a closer look at the raytracer and how it reacts with my conditions.

[Sabine]

Hm... aren't you already checking against bailout in the conditions of the loop?
If I try that something happens (AND result reacts to changes in bailout)

Well, good luck and most of all of course have fun on your mission I'm looking forward to seeing your results!
 

[SamTiba]

Your version is not workin in my Fragmentarium, because the function inside wants to have a boolean returned (and that makes absolut sense), so I wonder what it does in your application.
We have to check in the while-loop and after the loop return, if we are inside the set or not (and that is the actual condition here).

I corrected some issues in the code (the for-loop was wrong) and I have now a version to test, I don't get any usefull images until now but feel free to try it out by yourself!
All important parameters are in the 'QuaternionJulia' group, if someone spots a known fractal that would be awesome (because then it works).
My GPU is kinda slow and I have to restart the program every now and then, is the GPU-Frame-restriction from Windows vista and 7 in Windows 10 also active? That could explain it.

[Sabine]

It does of course do utter nonsense, as you have well spotted  

But it seems the TDR is still active in win10. It says in https://docs.microsoft.com/en-us/windows-hardware/drivers/display/timeout-detection-and-recovery... from Vista upwards and the document is current (2017). I have a (win7) machine with a rather quick GPU (980Ti) but fragmentarium wants to kill it nonetheless without having changed the TDR settings;)
It also helps to reduce the size of the preview window as much as possible.
Had a look at your newest version (thank you for sharing! ) but sadly nothing much happens, not matter how I mistreat it :}

[SamTiba]

You really have to play around a lot, I guess I have to optimize some things there to work properly.
But I am getting first results, sadly just when I am inside the structure.
TDR kicks me out fast but here is my first properly result with a polynom of grade 4.

It does not Compute in full-preview-window, also turn down the iteration to 1-3 when you move around and center yourself somewhere in the middle and start playing with the factors. It's usefull to set the 'far' value from the raytracer to see more. Maybe then you'll be able to get something too.

[phtolo]

You can change the TDR settings, eg increase the number of seconds before the GPU gets reset or turn it off completely.

https://msdn.microsoft.com/en-us/library/windows/hardware/ff569918(v=vs.85).aspx

[SamTiba]

Sorry for the amount of responses but I'm currently working hard on it and I'm excited to share results with you!
The change in Tdr-Registry makes it possibly to render full-size-images.
I now got the first rational polynom working and the results of 1/(ax^2+bx+c) already stop being symmetrical and therefore start to look interesting.
Attached is a file to try it out by yourself. Make sure to set the amount of Subframes to 1.

I'm happy to hear if anyone sees potential or is interested in the enhancement of this.

[Sabine]

Thank you for making me look again, SamTiba 

Though your math is lost on me (I'm really only 'innit for da pwitty picshers'  ), your frag is fun and challenging.

Found a quaternionesque place which is nicely stable. If you are interested: attached the zip file with in it a brute-raytracer that is able to give a bit more depth. If you unzip the files into a separate folder it should work.