Interesting fractal with really stubborn grain

[MindsEye]

I found this interesting fractal but I cant seem to get rid of the grain no matter what I do...

I got the raystep and stepwidth at 0.0009 and 0.0005 respectively and it still didnt clear all the grain.

I am still fairly new to this software, but not new at all to fractal art and software. I could just use some pointers to tell me where i am going wrong here..

Mandelbulb3Dv17{
T....../....1...M7...2Q....d5SnO9aXrzqPQz.zRO5wjPvi.SfhEjzvPdVS2FpRbzaRAqaAvrsxj
................................jH4wSG8lX12........Y./......../E........I/2.....
.....0.........U/Ek5m....6UOb...r3....E/.....2g8bGTLYiaD/..W3OtD...m/dkpXm1....E
.FkPGAcC12../2kAnAnAnAnyz0.................................u1....y1...sD...../..
.wnPGAUCORYfMkGgHtXbeGJ1aTnNyk9851ZfVOZjO32DjHYBUt1zy2petQTLyaNAsbBuW0aj09AsLAs4
Yt9EHa/Vkwf3y4fs33EIo9aj.....yHvw4............kD.2....sD/.kz0...................
.............ckpXm1...sD....zUJRV3.yp3K.MOLM/UnRV3.qr3K.sVLM/UFSV3..............
..................kz.wzzz1kN.g0.p....61...EJ....m....c3...k5....b....6X7...UPJ4I
2V3mm6YL8yzlcisz0tZWzTAe9mD.oc..zzzz.EmYEkti4xwj3Bk5eP./cz1...........U.8.UQgk0.
todIWOor/zPF1wVu4E.uz.........../613zHW7./UpfjrwdvguzoklrlkCx9xjShM3ZaYq1zP./c..
nc1D..CwQOr9PExDlX0dRIRSkz1..........2U.8.kzzzD.................................
/6U0.wzzz1....................................8cU0.E./2.zz/cU08../2E.szDU08c..2E
./EzT/8cU0.E./2.Qs5cU08../2E..bTU08c..2E./.ly/8cU0.E./2.Mw5cU08../2E.kqTU08c..2E
./.kz/8cU0.E./2....cU08.zz/cU08.yz1cU08.xz3cU08..w4PX3aPjpGMopmQptqQZFb9e/rN..rN
E.E..E...../....I....U....kLZxpA2FLNnB5......................................6U.
4M.................../.............................../........zD........kzXaNaNa
NaNez.........0E................................................................
.....................2....U/....1....MoPgFKOiR467t4RU.pPr/......................
8..0...........................E8QxckpX0jzHQxckpX0Lyz.........zD................
................................................................................
................................/....E/.........TB3QcJaQZZYPqB2BY/..............
..........................................A.....................................
................................................................................
..........................................U/....1....6.....JmZqMj7bP.waP........
.....................................................1........zD................
................................................................................
........................................................}


Thanks in advance for any tips!

[Sockratease]

Without checking your parameters, the best way I found to remove stubborn grain is to render several times larger than you intend for the final image, then downscale it.  Little dots will vanish nearly completely.

Hope that helps.Or try increasing the DE Stop, but that can change the shape of your fractal sometimes.

Good Luck   

[taurus]

couldn't get rid of all grain, but most of it i could eliminate. main problem seems to be the min iter at 50. decreased it to 1, and readjusted raystep and stepwidth to some acceptale values.


Mandelbulb3Dv17{
V....../....1...K0...2Q....d5SnO9aXrzqPQz.zRO5wjPvi.SfhEjzvPdVS2FpRbzaRAqaAvrsxj
................................jH4wSG8lX12........Y./......../E........I/2.....
.....0.........U/Ek5/....6EC2...W2....E/.....2g8bGTLYiaD/..W3OtD...m/dkpXm1....E
.FEnAnQD12../2kAnAnAnAnyz0.................................u1....y1...sD...../..
.w1...kDORYfMkGgHtXbeGJ1aTnNyk9851ZfVOZjO32DjHYBUt1zy2petQTLyaNAsbBuW0aj09AsLAs4
Yt9EHa/Vkwf3y4fs33EIo9aj.....y1E7.............kD.2....sD/.kz....................
.............ckpXm1...sD....z.MRQ9.6qlh..P5r0.qRQ9..slh.UW5r0.ISQ9.W3OtDp....UZ.
..................kz.wzzz1kN.g0.p....61...EJ....m....c3...k5....b....6X7...UPJ4I
2V3mm6YL8yzlcisz0tZWzTAe9mD.oc..zzzz.EmYEkti4xwj3Bk5eP./cz1...........U.8.UQgk0.
todIWOor/zPF1wVu4E.uz.........../613zHW7./UpfjrwdvguzoklrlkCx9xjShM3ZaYq1zP./c..
nc1D..CwQOr9PExDlX0dRIRSkz1..........2U.8.kzzzD.................................
/6U0.wzzz1....................................8cU0.E./2.zz/cU08../2E.szDU08c..2E
./EzT/8cU0.E./2.Qs5cU08../2E..bTU08c..2E./.ly/8cU0.E./2.Mw5cU08../2E.kqTU08c..2E
./.kz/8cU0.E./2....cU08.zz/cU08.yz1cU08.xz3cU08..w4PX3aPjpGMopmQptqQZFb9e/rN..rN
E.E..E...../....I....U....kLZxpA2FLNnB5......................................6U.
...................../.............................../........zD........kzXaNaNa
NaNez.........0E................................................................
.....................2....U/....1....MoPgFKOiR467t4RU.pPr/......................
8..0...........................E8QxckpX0jzHQxckpX0Lyz.........zD................
................................................................................
................................/....E/.........TB3QcJaQZZYPqB2BY/..............
............5.............................A.....................................
................................................................................
..........................................U/....1....6.....JmZqMj7bP.waP........
.....................................................1........zD................
................................................................................
........................................................}


guess that's a good starting point  

[blob]

Try also increasing the stepcount for binary search. 

[MindsEye]

Wow thanks for all the usefull tips.. i learned a few things here...

Here is another question, which smooth normals value is highest 7 or 1 ? (which smooths more i mean) and what does 0 do?

Also does anyone ever change the R bailout value? If so in what circumstance.?

Other than those 2 things I am pretty comfortable with the rest so far... really excellent program..

http://mindseye69.deviantart.com/ is my gallery if anyone wants to see what I have done with the software since I got it like a week or so back.. I would love to hear any comments or critiques.

Thanks again for the help, as a new guy here I feel very good.

o/

[MindsEye]

couldn't get rid of all grain, but most of it i could eliminate. main problem seems to be the min iter at 50. decreased it to 1, and readjusted raystep and stepwidth to some acceptale values.

Firstly thank you very much for taking the time to load up the params and check it out, I appreciate your effort. I think I misunderstood the iterations count. I was under the assumption that it was like upping anti-aliasing values in a 3d program... that it fired more virtual rays at the object per square inch or mm or whatever... but I guess that was wrong. Fiddling with it now I see that changing it has some pretty heavy effect on the actual geometry. I wasnt expecting that. Guess I dont know what it does still lol. So much still to learn.

[taurus]

...
Also does anyone ever change the R bailout value? If so in what circumstance.?
...

a general hint:
change everything in sight and see what happens - especially those things others would disadvice!
you can't really destroy something and only when you tweak it in some unusual ways, you have the chance to find something really individual (while you'll find a lot of rubbish inbetween)
example: set the bailout of your parameter (or my mod) to 6.
you'll loose some more grain without changing the personality of this fractal too much. and you gain expierience with every foolery you do!

have fun! 

[cKleinhuis]

wow, you have a grite selection of contrast colors, fractal art is much more than finding a nice formula or view, as mentioned somewhere else, it is like photography, and photography needs decent aspects to grab ones mind, and by just quick browsing through your gallery  i see not soo interesting formulas, but stunning viewpoints and excellent creative uncommon colors, keep on!

[MindsEye]

...
Also does anyone ever change the R bailout value? If so in what circumstance.?
...

a general hint:
change everything in sight and see what happens - especially those things others would disadvice!
you can't really destroy something and only when you tweak it in some unusual ways, you have the chance to find something really individual (while you'll find a lot of rubbish inbetween)
example: set the bailout of your parameter (or my mod) to 6.
you'll loose some more grain without changing the personality of this fractal too much. and you gain expierience with every foolery you do!

have fun! 

Yea thats kind of how I got this far.. pretty much tweaking everything, sometimes the potential for so many different images from even one simple formula modded can feel almost overwhelming, but I love the hunt... and I tend to be pretty brutal about tossing anything that isnt just perfect. Sometimes I find an angle thats nice, but not quite right, and I know that once i go away from it i will probably never ever find that particular angle or effect ever again. Like a needle in an infinite haystack. hehe.

I love it, even though there's just not enough time in a human life to explore even a tiny fragment of whats possible. Thank you for the advice. when i saw 1024 i thought it might have something to do with memory, thats why i was afraid to touch that specific one.

[MindsEye]

wow, you have a grite selection of contrast colors, fractal art is much more than finding a nice formula or view, as mentioned somewhere else, it is like photography, and photography needs decent aspects to grab ones mind, and by just quick browsing through your gallery  i see not soo interesting formulas, but stunning viewpoints and excellent creative uncommon colors, keep on!

Thank you very much! And yes, you are correct, not every formula is the most interesting mathematically, but even a simple pattern, from the right angle, can be very dramatic. My understanding of color comes from my doctorate work on the harmonic relationship of phi in the visible color spectrum. You can download my colorset software which I used to prove my thesis work from http://www.goldennumber.net/color.htm it produces harmonic color triplets. Adobe copied my software with thier Kulor software and left out the Phi function as it was copyrighted by me. I have since given them permission to use my work in thier program, and they are adding the phi set to an as yet unreleased version.

I hope you enjoy it and find it usefull, and again thank you for the kind words.

Dr. Michael Semprevivo

[kameelian]

Hi folks,
 
Some useful tips/reminders have been offered in this matter so far.
 
I too get some issues with persistent grain on occasions, where reducing RSM and SWL so low that it takes hours to render but still doesn't seem to remove it all. I haven't done much with bailout for some time and min iterations often gets left out (as iteration count does change colours and geometry), so I will have to remember these things.
 
By the way, I'm sure I once read that setting Step Count for Binary Search past 12 or so doesn't really help. Is this correct? I can't say I've found much difference past there myself.
 
If I can add another pointer on the grain matter - surprisingly, I have found that *sometimes* the simple act of changing the colour set has a notable impact on SOME types of grain - although logically, I cannot understand why such a simple action should have such impact - unless it is just about light and angles. It may not be the answer if we want a particular colour, but it might just work in some instances where we are not too wedded to a certain colour set.
 
Hard to believe? just change some colours on these parameters (or certain other pictures you have in the main calculation window) and see. For example, see pics below. I have deliberately put RSM and SWL up to 1 for speed of render and to emphasize the grain - but just using standard colourmaps 9 and 10 - and altering the fog for the 3rd, have this much difference on the pictures using one of the custom colour presets in my list. Worth considering along with other tweaks at least.
 
Thanks MindsEye for your harmonic colour relationship software. I shall give it a try.
 
regards
Kam

[cKleinhuis]

wow, you have a grite selection of contrast colors, fractal art is much more than finding a nice formula or view, as mentioned somewhere else, it is like photography, and photography needs decent aspects to grab ones mind, and by just quick browsing through your gallery  i see not soo interesting formulas, but stunning viewpoints and excellent creative uncommon colors, keep on!

i meant "great selection ... "

[Jesse]

Here is another question, which smooth normals value is highest 7 or 1 ? (which smooths more i mean) and what does 0 do?

0 means a usual calculation of normals, higher values adds more steps of calculation plus a calculation of surface "roughness" from variation among the different offsets.  But this helps more on specular aliasing, not here.
Btw, there is a normals smooth value of 8 that calculates a volume of offsets, what is really slow.
Much faster would be to calculate the normals with postprocessing "Normals on Zbuffer" after main rendering, for such cases i would recommend this.

Some formula hybrid just gives very bad and unsteady distance estimates, so rendering can be tricky or is failing totally.
Can't do very much about this, reciprocal formulas can be often troubling for example.

[MindsEye]

Here is another question, which smooth normals value is highest 7 or 1 ? (which smooths more i mean) and what does 0 do?

0 means a usual calculation of normals, higher values adds more steps of calculation plus a calculation of surface "roughness" from variation among the different offsets.  But this helps more on specular aliasing, not here.
Btw, there is a normals smooth value of 8 that calculates a volume of offsets, what is really slow.
Much faster would be to calculate the normals with postprocessing "Normals on Zbuffer" after main rendering, for such cases i would recommend this.

Some formula hybrid just gives very bad and unsteady distance estimates, so rendering can be tricky or is failing totally.
Can't do very much about this, reciprocal formulas can be often troubling for example.

Thank you, that helped me immensely!  One more question while I have the attention of the masters, I noticed something that seems to go against everything I know about aliasing, I noticed that sometimes when I have an image that has grains along edges, that if I render the image at anything up to say around 3600 x XXX it will still have the grain, but then if I render the same image at 5760 x XXX suddenly the image is not grainy but crisp and clean. How is this so? Why is it not identical but larger? This makes my brain hurt : )

I come from a background of 3d software since the late 80's, so my brain keeps looking for an Antialiasing parameter with max/min iterations, but the iterations adjuster in mandelbulb seems to not mean the antialiasing iterations, but rather iterations of the formula, which was confusing to me at first. I am getting used to it.

I must say that this is the most fun piece of graphics software I have had to play with since zbrush! Really superb bit of coding! I am stunned and amazed... GREAT work!