Mandelbrot Zoom 999 (no sound)

[antimasonic]

this is amazing

[dkmann1987]

What is this program? I've just started playing around with some of this stuff. I'm looking for something that can generate these and other fractals using your GPU to do the crunching.

[Conseguidor10]

Wow, what you to be ?

[thekurlzz]

this is the trippiest thing i've ever seen in my life

[supergsx]

Is this the same type of rendering as they did at the beginning of The Matrix Revolutions?

[Whitefire240]

Do the same thing but for the mini mandelbrot at the end.

[winetam12]

I like to watch this one on full screen with Snatum Kaur playing in the background on Pandora.com..........a very beautiful meditation. Thanks for posting the video.

[ncuxap12444]

it gets even more complex on higher zoom levels. very interesting video, im waiting for the 10^10000 :D

[Oenols]

Soundtrack suggestion: Philip Glass’ Music With Changing Parts.

[Teraflop4096]

Technically well done, and the colors are also simple yet nice. Unfortunately, the zoom itself is very repetitive.

[expandranon]

Do you have a higher resolution version?

[tharqal]

isn't your interpolation taking a bit of an artistic licence here? there are some images that I don't think actually exist in the Mandelbrot set, like at 1:31

[phaumann]

@tharqal

The two slits at 1:31 are the consequence to the "peak fly-by" at 0:22. By the way, the "slit fly-by" at 1:31 has consequences too: the bended slits at 3:37 for example. Those and further considerations are really helpful to hit a Minibrot at a certain zoom level.

I have to admit, that the video contains errors :-} At 3:16 there should have been a symmetry of 16 slits but the slits are not symmetrically arranged -- my fault.

[Rinfiyks]

@phaumann How did you write the program? I wrote a mandelbrot set program in C# but I cannot zoom deeper than 10^13 because the numbers are beyond the power of the double precision library I'm using.

[tharqal]

@Rinfiyks standard double precision will fail out pretty quickly for jobs like these, so you need so called "bignum" arithmetic which is pretty much code that takes apart really long values to make calculations in multiple steps that are all below the limits of normal CPUs. it's a lot slower of course but you can theoretically have unlimited precision

[Rinfiyks]

@tharqal Okay, I'll look into it, thanks a lot

[EclipiticalVampire]

it's like being born... this is what it's like...

[ouroboros28]

the last image is the equivalent of the light filled version of the first image which was filled with "apparent" darkness/void. Darkness/void is an illusion. All is full.

[fractalzooms]

So far ahead of the curve. This is an impossibly handsome fractal zoom.

[stardust4ever]

A+++ Hands down, greatest zoom movie I have ever seen in my life!!! I am so glad you finished it off with the minibrot at the end.

I have a quad-core Phenom @ 3.6Ghz, (fractal extreme, 64 bits) yet I well know the pains of slow rendering.

Your rendering technique is gorgeous; the iteration bands are never too dense or too sparse, yet they always follow the zoom. This movie is just beyond words for me. Kudos to you as well for not cheapening the experience with crappy music. ;-)

[denelson83]

So this is what a near-death experience looks like, eh?

[MAXT3R88]

when you reach the final level of zoom God appears and says "Catcha!"

[XPac894]

I sure do like all those crazy colors!

[JyssyE]

Brilliant.

[BlockJuice]

Doing the math, and assuming the last screen is the size it is and at 96 dpi, the first screen is 8.1x10^973 times bigger than the observable universe.

Huh.

[supergsx]

This is the best video on YouTube.

[supergsx]

3 MONTHS!??!?!??!?!?!?!

[greengobbb]

thank you for sharing <3

[Spectre426]

Incredible, thank you for posting this.

[DeepZoomNet]

Quite a feat, to be sure! What resolution was this originally rendered at? Is this done frame-by-frame or is there some digital zooming technique involved? How many video frames are in this, and what resolution?

[phaumann]

@DeepZoomNet

The frames have 1280x720 (HD) resolution.

There are approx. 18.000 frames (= 10 min. x 60 sec./min. x 30 fps).

With a total of approx. 12.500.000 points calculated we have a theoretical resolution of approx. 700 pixels / frame = 35 pixel x 20 pixel per frame. And that's where -- as you expected -- some digital zooming technique comes in.

[DeepZoomNet]

@phaumann So are you saying you only calculate each frame to 35x20 pixels, then digitally interpolate to make a 1280x720 video frame? It seems like you're doing something a bit more complicated than that.

[phaumann]

@DeepZoomNet

That's almost right. Keep in mind that these 700 pixels per new frame need not to be arranged in a 35x20 raster and that you can reuse already calculated pixels from other frames...

[DeepZoomNet]

@phaumann I'd like to hear more details about how you do this. My software makes a single "master" image then digitall zooms into it until the spatial resolution of the video frame matches that of the master image, then makes a new master image (well, it's more complicated than that, but that's the essence).

[phaumann]

@DeepZoomNet

In oder to gain a speedup of more than 100 compared to a "brute force raster" I choose quite a different approach.

Bottleneck of my approach, I think, is the use of Mathematica to calculate iterations. At precision of about 300 digits I get about 20.000 iterations per core and GHz. Usage of an arbitrary precision library might be useful.

Can you speedup my calculations? -- perhaps I can reduce your number of points to be calculated... :-)

[NickRoeder]

I need this as a screensaver....

Awesome.

[R1ckr011]

@NickRoeder i concur!! The ending: Spectacular. and to think our own universe is but a crude and classically wrought iteration of a fractal or two....

Speaking of which, how would one go about constructing an "anti-mandelbrot set where every "spin left" is countered with a "spin right" and the two initial patterns overlap and mesh together, since the mandelbrot Is chiral and therefore non-superimposable on it's mirrr image...

[R1ckr011]

@R1ckr011 mirror image i mean

[TAfTfilms]

This is a spectacular creation. I don't think I've ever seen one this deep.

[TAfTfilms]

@TAfTfilms If it is 10^999, than what is the 2^?

[ChazyK]

@TAfTfilms use logarithms to count it

[R1ckr011]

@TAfTfilms you tell me. 2^x=10. solve for X. then simply multiply X by 999 since (2^a)^b= 2^ab

Using my handy dandy Ti-84, the number is approximately 2^3.219281 (probly more like 3.2192809.... ). Ah here we go. Out to 13 digits--sad i know--3.321928094886 x 999 =

2^3318.6061667911

That's damn impressive. About that figure, 3.3219... dividing it by pi gives us 1.0574025537939, which is... more work than i can get into. slightly less than the square root of 1.119

[TAfTfilms]

@R1ckr011 Dude...

[R1ckr011]

@TAfTfilms how am i supposed to take that XD

BTW it was mirror :P on my other comment. I miss that calculator :'(

[TAfTfilms]

@R1ckr011 I'm just marveling at the numbers man. It is all pretty impressive. I'm sorry about the calculator man, he will be missed.

[dafaucas]

motion blur looks good, but u miss lots of details

[v1nc3nz000]

musi???

[BBCTheGreatWar1964]

2,4,8,16,32

[phaumann]

@BBCTheGreatWar1964

Good Point. Moreover we have at least 2-symmetry in the last half, at least 4-symmetry in the last quarter, 8-sym in the last eigth and so on ending with a Minibrot at the end. So, by breaking symmetry at 500 for the last time, I can ensure to see a Minibrot at 2*500, because of 2=1/1+1/2+1/4+1/8+...

[SoupIsMoney]

Awesome. Trippy. Wonderful.

[zZeiss]

amazing. thank you very much

[Kaervos1024]

This is astounding. Very well done indeed. How do you find target coordinates? Also, do you know what the record is for the deepest Mandelbrot zoom ever is? This might take that.

[johncatchp0le]

incredible, well done