# Buddhabrot

Robert P. Munafo, 2008 Feb 12.

Name given by Lori Gardi to the plotting technique invented by Melinda Green, which renders the Mandelbrot Set in a way similar to the Inverse-Iteration Method but using forward iteration and plotting the iterates in parameter space. If viewed sideways (with West at the top and North to the right) the Continent appears a bit like certain images of Buddha.

*Buddhabrot image from Wikipedia Commons*

Related colloquial terms for features in this image are: topknot, head, shoulders, body.

To produce a Buddhabrot image, pick a dwell limit as usual, and select a view that includes the entire Mandelbrot Set. Then, iterate as you would for a Mandelbrot image (see iteration algorithm), and find out if the iteration escapes. If it escapes, then look at every value of Z that occurred to see which pixel that Z value falls in (some Z values will be outside the view). Keep a count for each pixel, recording how many times a Z value fell within that pixel. After all the points have been iterated (or occasionally while iterating) plot all the pixels, choosing a color for each pixel based on the recorded count.

If you want to view a low-resolution image quickly, you have to choose points (values of C) to iterate in some way other than a normal grid scan. Melinda Green, Paul Bourke and Marc Jaouen choose points at random, however this produces a noisy "snow" effect that makes it necessary to do much more iteration before the subtle details of the Buddhabrot image become visible. A grid scan, or an interleaved stagged grid scan like that used in ordered dithering will reduce (but not eliminate) the noise.

The basic Buddhabrot image is similar to a view of the entire
4-dimensional Julia-Mandelbrot Space collapsed to a single plane
(the Z plane), and showing all the Iterates of Z_{0}=0, but no
iterates for any other initial Z. This shows the orbits, which
contain different numbers of attracting points depending on the
periods of the individual orbits. However because the bounded
orbits do not get plotted, the Buddhabrot image only contains iterates
of 0 for the Fatou dusts. The points that have high dwell values
dominate, and their orbits include many points that fall near the
value of C used for that iteration. The C values of points with high
dwells are the points near the boundary of the Mandelbrot set, and
that is why the Buddhabrot image contains an image of the Mandelbrot
set.

There are also many rotated and enlarged copies in different places. All of the rotated and enlarged copies are preimages under the iteration of the escaping C values. The number of copies depends on the various periods of the orbital dynamics near the points being iterated. For example, there is one extra image of the area around R2.1/2a, two extra images each of the areas around R2.1/3a and R2.2/3a, and so on. The images of the smaller mu-atoms are fainter because fewer iterates contribute to them. Because of the Squaring that occurs to the iterate once per period in its iteration, many of the preimages have the characteristic shape of a Mu-molecule, a cardioid with mu-unit decorations, and the cardioid's cusp is near the origin.

See also: Web pages by Melinda Green, Lori Gardi, Marc Jaouen and/or Paul Bourke. (To find any of these, use a search engine to search for these names or for "Buddhabrot")

See also The Amygdalan Sects, Mandelbrot Monk.

Here is some crude pseudo-code using an array that is 1000x1000 (a million elements total):

large_array = malloc(1000*1000*sizeof(int)) for (x=0; x<1000; x++) { for (y=0; y<1000; y++) { cr = ((float) x) / 250.0 - 2.0; ci = ((float) y) / 250.0 - 2.0; zr = zi = 0.0; for (i=0; i<1000; i++) { // Iterate: (zr,zi)^2 + (cr,ci) -> new zr,zi // remember iterates iter_history_r[i] = zr; iter_history_i[i] = zi; } if (cr*cr+ci*ci > 4.0) { // consider this point to have "escaped" for (i=0; i<1000; i++) { // retrieve one of the iterate values zr = iter_history_r[i]; zi = iter_history_i[i]; // convert zr,zi to array indices J and K if (J and K are both in range 0 to 999) { large_array[J,K]++; } } } } } for (x=0; x<1000; x++) { for (y=0; y<1000; y++) { convert large_array[x,y] to a color (typically 0 is black, large values are white, other values in between) plot pixel at (x,y) } }From the Mandelbrot Set Glossary and Encyclopedia, by Robert Munafo, (c) 1987-2018. Mu-ency index

This page was written in the "embarrassingly readable" markup language RHTF, and was last updated on 2018 Feb 04. s.11