Title: Lambda set symmetry Post by: megafiddle on November 04, 2011, 04:28:05 AM Greetings. New here.
I have generated the lambda set as described in Mandelbrot's "The Fractal Geometry of Nature". It's based on the function z = zc(1-z) for an initial value for z of (0.5, 0). My set is not symmetrical horizontally. It is symmetrical vertically. The main "mass" of the set is symmetrical. The asymmetry is in the branches that extend out from the main part, and also in the non-set surrounding regions. These parts become progressively more and more truncated as you move from the right part of the set (x = positive) to the left (x = negative). I assume it is supposed to have horizontal symmetry? I can't find much on the set itself, just the lambda julia sets. I am getting some interesting images, though. If there is an error in my code, I'm going to have to keep it as new variation. Anyway , I just wanted to confirm that the set should be symmetrical horizontally. Also, can the gallery be used for linking to an image? Title: Re: Lambda set symmetry Post by: bib on November 04, 2011, 02:36:20 PM Hi and welcome to fractalforums!
Looking at the standard formula in Ultrafractal, yes, with (0.5,0) as a starting point, the set should be symmetrical about the y-axis. Here the UF code : Code: LambdaMandelbrot {You cannot "link" images in the gallery, you can just upload them. You can also directly attach an image in a post by clicking on the Additional Options below the text box. Title: Re: Lambda set symmetry Post by: megafiddle on November 04, 2011, 10:42:26 PM Ok, thanks.
I suspected it should be symmetrical, but it seemed to work too well as it was. Amazing how well these things survive in spite of errors. What does this refer to, though? hint = "This parameter defines how soon an orbit bails out while \ iterating. Larger values give smoother outlines; values around 4 \ give more interesting shapes around the set. Values less than 4 \ will distort the fractal." Is it used for testing the magnitude of z, to determine set membership? It doesn't immediately follow a parameter like the other "hints". And the default for "bailout" is nowheres near 4. Title: Re: Lambda set symmetry Post by: cKleinhuis on November 04, 2011, 10:58:44 PM What does this refer to, though? hint = "This parameter defines how soon an orbit bails out while \ iterating. Larger values give smoother outlines; values around 4 \ give more interesting shapes around the set. Values less than 4 \ will distort the fractal." Is it used for testing the magnitude of z, to determine set membership? It doesn't immediately follow a parameter like the other "hints". And the default for "bailout" is nowheres near 4. yes, it is used as "bail-out" test, because it is a quadratic operation (z^2) the value can not get smaller when it is getting above 1 ... depending on the seed value, squaring of a value greater than 2 ( hence the comparison with 4, z^2=4 if z is 2 ;), to spare calculation cycles the z distance isnt squarerooted ;) ) will most certainly not come back to values below 2 the distortion is created because lower bail out values can lead to certainly wrong guessings, but it can be compared to the iteration depth... iteration is broke up as soon as some bailout is reached so, and last paragraph about why to set the bailout value higher than the maximum needed for calculation improvements ... the reason is, the value will grow really fast, and even faster ;) and the "smoothing" or smoothed color methods - those that create gradients and not stripes - work with the nice logarithm property, which works very well with big numbers, and they can then use logarithmic arithmetic ( dunno what i am talkin 'bout ) to create the finer gradients when using big bailout values, the nice thing is that the values do not grow over floating limitations e.g. the bailout is the maximum value the length of the z value may have, somehow this is then used to map the result to a real gradient as far as i know can david m. code this thing in sleep mode, or on a chalk board, and can explain you every step ;) ah, yes, and: welcome to the forums Title: Re: Lambda set symmetry Post by: megafiddle on November 04, 2011, 11:43:35 PM Thanks.
That's what I am currently doing for the mu (Mandelbrot) set, comparing the square of the magnitude of z to 4.0. Maybe that's the problem with my lambda set; I am comparing the square of the magnitude of z to 1.0. Ok, found an older working version of my program and changed the test value to 4.0. Everything is now symmetrical about both axes. (My program is undergoing some major changes and nonoperational. Had to find a an older version) My code for the lambda set goes back to the '80s when all I had available was 640 x 480 VGA. Never noticed the asymmetry until I dusted it off and tried it on a modern PC. Strange that test values of 1.0 and 4.0 both produce nearly identical sets. Thanks again, that fixed it. But I am going to have to add my "error" as an option. It did produce some interesting images. Sort of looks like frosted glass bubbles: Title: Re: Lambda set symmetry Post by: Vega on November 05, 2011, 12:06:35 AM My variant :dink:
Title: Re: Lambda set symmetry Post by: megafiddle on November 05, 2011, 10:59:56 PM Nice image! I haven't done anything with the interior of the sets, just the surronding aura or corona (as I call it). My program is limited to coloring based on escape time. All the coloring is done by operating on a 1000 entry color lookup table. That's where all the work is right now, adding editing tools for the table. Also I'm not familiar with Ultra Fractal and the other software that's out there. In the UF code above, the bailoutout parameter is set to 1.0e20. Is that related to a "normalized iteration count" algorithm, for smooth coloring? |