FAQ Frequently Asked Questions

The Mandelbrot set is a mathematical set of points that are computed from a relatively simple formula.

Its name comes from the Mathematician Benoit Mandelbrot who did not discover it but was one of the first to study it.

The Mandelbrot Set holds a special place among fractals as it is one of the more easily recognizable ones, while at the same time being very varied.

Typical fractal explorers compute the fractal on your machine, this means that exploration is slow, power-hungry or has to be compromised on quality.

The fractal here has been pre-computed to a high resolution:

• Exploration is fast and doesn't need much computing powre (or bandwidth)
• The Mandelbrot Set could be computed with a high number of iterations.

The Mandelbrot Set reveals itself more with high number of iterations, but more iterations mean more CPU time. Many areas of the fractal you can explore interactively actually took several minutes of computing time on fast CPUs.

None! The Mandelbrot Set is only semi self-similar, which means that while you may recognize some patterns, they will never repeat in the exact same way.

There is one symmetry along the vertical axis, which was taken advantage of during the computation.

The current database was computed on the smoothed Mandelbrot Set up to a depth of 65535.

It is likely that at some point in the future, relevant tiles could be recomputed to greater depth.

Each zoom level corresponds to a doubling of the resolution:

• Level 0 is made of a single 256x256 tile
• Level 1 is made of 2x2 = 4 tiles of 256x256 pixels
• Level 2 is made of 4x4 = 16 tiles of 256x256 pixels
• etc.
• Level 14, the current limit is made of 16384x16384 tiles of 256x256 pixels

At Level 14, the whole image is thus made of 4194304x4194304 pixels, or 16 TeraPixels, that is 16 millions of millions of pixels.

The database for Level 14 is currently 84 GB in size (uncompressed it would be 32 TB).
The compression ratio is achieved by:

• taking advantage of the Y symmetry
• handling tiles of uniform depth specifically
• encoding depth values as a delta of neighbours
• using ZLib highest compression

At the moment there is no plan to add it, for storage reasons :-).
Even with all the compression, Level 15 is expected to be around 300 GB, and Level 16 would likely be above 1.1 TB.