Welcome to Fractal Forums

So, for a while, I've been wanting to see a zoom movie of stardust4ever's Magnum Opus Ex.

http://www.deviantart.com/art/Magnum-Opus-Ex-3132-6-Zooms-274474754 (http://www.deviantart.com/art/Magnum-Opus-Ex-3132-6-Zooms-274474754)

I got the feeling from his Deviant Art profile that, although at some time he wanted to, he never got into the mood of making one and there's no way on Earth that I could ever blame him for that. Well, recently, I got a new laptop that actually has a decent amount of muscle which, in combination with Kalles Fraktaler, actually made making a zoom movie from its center value very practical. I haven't uploaded it³ to a public URL, though, for a couple reasons. For one, the linked mux of it is just something I threw together in Blender real quick because I was so anxious to watch the result and didn't really want to have to sit through a 1/2-hour-long silent video¹. Also, I wanted to make the zoom not just for myself, but as a service for stardust4ever, so I intentionally tried to imitate his palette, and I haven't quite yet been able to correspond with him on it. So, I'm not going to just toss the video out into the open air. I'd much rather use common-courtesy discretion even though I'm about 95% sure, he doesn't care as I could tell when I was writing the palette and thought out the HTML hex values² of its primary colors only to find his apathy staring me in the face. ;-) Also, I've yet to get into the mood of expediting the movie's audio track to one that would be appropriate for the sharing sites.

Anyway, this video does zoom all the way down to the corresponding minibrot for the image's coordinates. Hope, you like it and, of course, all suggestions and constructive criticisms are welcome.

¹Nonetheless, the video ends with a minute-and-a-half-long audio gap.
EDIT: The video being referred to by this footnote has since been replaced.
²000000,FF0000,FFFF00,FFFFFF,00FFFF,0000FF
³https://mega.co.nz/#!ads02Iha!5W8bU-lgjupbAI8kdpkxQjK4h1ygQkEy93TYyYqjX7w (https://mega.co.nz/#!ads02Iha!5W8bU-lgjupbAI8kdpkxQjK4h1ygQkEy93TYyYqjX7w)
https://mega.co.nz/#!OQFDCB4K!oAcWtZle7pzWLtVJrhQ1KwmsqHY5bWdocwkuEHsiRY8 (https://mega.co.nz/#!OQFDCB4K!oAcWtZle7pzWLtVJrhQ1KwmsqHY5bWdocwkuEHsiRY8)
EDIT: A public URL has since been added:
https://www.youtube.com/watch?v=gvLCT_JxCV8 (https://www.youtube.com/watch?v=gvLCT_JxCV8)

Frame map tarball:
https://mega.co.nz/#!yF8FRYRL!eD6wQqbcO9iiXwmgln00BLTv-IOVIx48yH5QwiAezy0 (https://mega.co.nz/#!yF8FRYRL!eD6wQqbcO9iiXwmgln00BLTv-IOVIx48yH5QwiAezy0)



Specs:
 Encoding resolutions:
  Spatial:
   Visual data: 1080x607
   Compiled video: 1920x1080x24
   Linked mux: 1280x720
    File length: 3.27GB
                    3.48GB
  Temporal:
   Linked mux:
    Framerate: 30fps
    Bitrate: 15.6mbps
               16.7mbps
    Runtime: 29:56
Location:
 Center value:
-1.479796270901760024425279934770411525645551054432599517909807632824286254403907594526888466099962805022975196472549771681831234491695559852583955204986197762293872769474806903995564259040667568599770094112637937857082198756105097445619350756966793793114467432032937799346574014461957202648890738272675810429928192557139660619600296009564885280394001101360257346740494235037815372540293867629743448774089362874648952714837834757261669647915507055305674126984656711318786930953070950608370461895219919337664617433737293352968231749583061758258422599662604722098000000314666116027436011175894993211309068159162773679207450814155503083614827290221363244089017194859887036603611917809816800212447293119967557597851563066664929812719639801868841093522201098388080690835201401107775907424916066058853196986343451890576206748447874886892854877837677243840549640075772999621830901631124562599208718790588491784049700616627605181650170566542555529051199024120015083824033582065546753123188812619141291351068107748120740542228347936404419072491652758150552076092069598233522865072370987119006991017852479929569970704+0.001199443952813447746281973233374468444560314114132538362037569205657422216739564521471119107626453330996365067987088146663639996715939831819152248618042255824652268918299630897525386638029428706473920898358506195494798092175170947056508005125898188455476513341931256963209945768517999965001667076299333644684423677803421525767262584878869446028206903966564433218075051021122916794800140935842945782240831852572867457930764602730848767321435032630795018827988435399501930146465253643324698819626072050948659258653511129505375425645487770269836685433774043137622036169207385516752176231241602743976595972075603320408006389794749279655916354541539782745462023161731064232067674198549286361618303087153045942207147851253894840762493307442181404211167174928186254424413376758234084677871757029117172134794379472434314696191681618264140322839436244764371856772842373936969861577709408726036547469130011417964191068834221715324766522573823191722264068340809877388074085776799380824167633762846222500487351004632006267104521393235977416011673478415017751029891847931325006537698668478197435116049586294582890624ⅈ

Magnification: 2.90735489713E1078 (2^3582.57819347876)

Spectacular! Thanks for rendering. Stardust4ever's work was amazing, especially considering the render speed at the time. His images are the reason I started exploring the Mandelbrot set.

plynch27, too bad stardust doesn't seem to be active anymore. But I think you should publish your amazing movie, I have downloaded it and watched it  :o
If he would mind, you can simply remove it.

If you have kept the data frames and are willing to redo the movie, let me suggest you speed it up but slow it down on the main view at e-943, and do it really slowly there.

I just found stardust4ever a couple of weeks ago, I knew about him before but hadn't explored it much before, and yes his work is very amazing and land-breaking.
I searched if there were any high power mandelbrots already published on youtube and found his deep 3rd power render. Not HD but still excellent.
And I found many other very interesting movies from him, really inspiring with other amazing types of 2D fractals.
And I also found some really cool movies from you Dinky with high power Mandelbrots ;)

These are very good points. First of all, don't worry; whenever I render zoom sequences, I never delete the visual data output from the render. All the intermediary data between that and the end product are, obviously, gonna have to go at some point, but the visual data is effectively precious to me. The rest can be re-rendered in less than a day. Second of all, just forewarn, I'm not a very big fan of color cycling or rotation, though color cycling is great use temporarily for emphasis and illustration, because it beautifully reveals the image's structure which is great in images that are so complex that it's very difficult to sort out exactly how all the structures are simply connected.

Anyway, </digression>, I'll find some AGPL music to mux into this video ─ I actually still have all 156GB of its AVIs simply because I've been meaning to make it publishable but haven't gotten there, yet. Then, I'll render out a 2nd version for ya guys that actually has an amount of production put into it.¹ ;-) I'll let you know about them on this thread.

Thanks, guys.

¹I've also always been a fan of constant speed zooms, for some reason, especially at a speed of 2 zooms/sec. which for me has always been the happy medium between too fast and too slow.

If you like constant speed, you should do it with constant speed :) ^-^

No, don't you go tryin' to talk me out of it, now. I will be making two versions: ... Options  :dink:

cool project!
but no matter in which player I open it, the movie seems squeezed.
I just noticed that you can force VLC-Player to use 16:9, looks fine now. But I never had to do this before..
Am I the only one who has this "problem"?

And do you have any idea why there is this vertical green line on the right?

I don't mean to criticise with this, it's just that I noticed and wondered..
I really like this video! especially that you uploaded it in real hq, not just youtube that destroys all the great details.. there's nothing like

I'm also sure that stardust would love to see this - someone who was crazy enough to zoom that deep in those "early days" without pertubation definitely can appreciate this.

cheers!

edit:
what is agpl-music? like creative commons?
if it meets your taste you can use this track I made, it would span the whole length: http://www.evermusic.de/media/temp/Hypnose32min.mp3

I also had problem with the ratio of the movie...

Yeah, I know. In the end, it's really just a bad piece of information in the movie's metadata and I still haven't been able to find the setting in Blender that's causing it. Unless I'm just misunderstanding how the "aspect ratio" setting works, I've got it set to, specifically, 16:9 with the resolution at 1280x720, but, nonetheless, the end result's "aspect ratio" field is getting set to, like, 3.25:1 or something. I think I might have metadata editor around here somewhere, now that I think about it, but it would be nice to figure out what's wrong. This isn't the first time I've put together a video in Blender, but it's the first time I've come across this problem. If I post the .blend file, does anyone think they might be able to find it?

Okay, I found out what the problem was. I misinterpreted Blender's "Aspect Ratio" setting as the video aspect ratio when it was referring to the pixel aspect ratio. So, when I set that aspect ratio to 16:9, it got multiplied by the 16:9 aspect ratio implied by the video's render resolution setting causing the aspect ratio tag in the mp4 metadata to be set at (16:9)^2 or 3.16:1. I'm encoding the muxes now and I'll drop another post on this thread when they're uploaded to give you e-mail notification.


Attribution-Only General Public License. So, effectively, yeah.
and, yes, thanks for the track. It sounds like Orbital ─ you know, "Hackers", "Mortal Kombat" ─ and brings me back to the 90s.

Alright, here's my next mux of it:

https://mega.co.nz/#!OQFDCB4K!oAcWtZle7pzWLtVJrhQ1KwmsqHY5bWdocwkuEHsiRY8 (https://mega.co.nz/#!OQFDCB4K!oAcWtZle7pzWLtVJrhQ1KwmsqHY5bWdocwkuEHsiRY8)

This one has the correct aspect ratio in its metadata and the track Chillheimer made and linked to four posts above this one in its audio stream. I'll be trying to keep an updated list of the files' links ─ not that I plan to upload a lot of them, but there will be infrequent "version updates" and hosting site URLs ─ in this thread's originating post.


Actually, I was unable to duplicate that problem.

Alright, guys, I finally got the video a watch page on YouTube:
https://www.youtube.com/watch?v=gvLCT_JxCV8 (https://www.youtube.com/watch?v=gvLCT_JxCV8)

It's amazing, thanks alot. The x of x of x is still the deepest Julia morphing ever!

Hey, if anyone's interested, I've uploaded a tarball of the frame maps:

https://mega.co.nz/#!yF8FRYRL!eD6wQqbcO9iiXwmgln00BLTv-IOVIx48yH5QwiAezy0

It's a 9.4 GB download that will decompress to 17.3 GB.

In case you're thinking "WTF?", no, I normally wouldn't do this. It's just that because of the circumstances of this video, I'd like to make it as open as possible.

Hey guys, I'm back, baby!  O0

First off, sorry for the late reply. I guess I've taken a hiatus from fractal rendering for a while. Sometimes I go for months without logging into my Deviantart account. I'm still around, although my interests have primarily shifted back to my old hobby of retro video game collecting. After the Magnum Opus Ex fractal, deep zooming took a back seat to Mandelbrot variants with abs() commands, like Buffalo, Celtic, Burning Ship, etc. At some point a year or so ago, I guess I got bored with the fractal thing. Magnum Opus Ex was so deep that I knew I could never do a full zoom into it so I basically did a remake titled "Magnum Opus Lite." This was basically a search for a similar formation at a much lighter zoom depth. I accomplished this by shortening the arm length of the original X formation I used as the base. As expected, shorter arms equated to a more shallow zoom depth for the final formation, this time I discovered the formation at 1731 zooms instead of 3132. Using this new location as a start point for a video zoom, I successfully completed a full zoom from start to minibrot that totaled around 1800 or so zoom levels. Due to the extreme amount of zooms, each Fractal Extreme zoom frame was rendered at 640x360 pixels, interpolated to an HD output frame size of 1280x720 for cleaner scaling. Magnum Opus Lite took a little over 3-and-a-half weeks to render on my AMD 8-core FX Bulldozer (overclocked to 4.2Ghz). Zoom speed was fairly swift to keep the time from running long. I've been sitting on the Magnum Opus Lite video for a couple years, but never got around to uploading it as my interests have moved on a bit. Being 640x360 native, it was a little bit pixellated when played back on an HD screen.

Magnum Opus Ex (3132 zooms):
http://stardust4ever.deviantart.com/art/Magnum-Opus-Ex-3132-6-Zooms-274474754

Magnum Opus Lite (1731 zooms):
http://stardust4ever.deviantart.com/art/Magnum-Opus-Ex-Lite-1731-Zooms-301462979

Side by Side Comparison:
http://stardust4ever.deviantart.com/art/Magnum-Opus-Ex-Comparison-314583454

Pteradactyl Bone (2482 zooms - Periodic doubling from Magnum Opus Lite formation):
http://stardust4ever.deviantart.com/art/Pterodactyl-Bone-2482-Zooms-298663841

My Fractal Gallery:
http://stardust4ever.deviantart.com/gallery/9952329/Fractals

Anyway, were it not for my procrastination to upload the oversampled , the OP may never have taken the reigns and finished the original zoom. I'm currently downloading the files from Mega right now. For the sake of the topic, I'll get the original Magnum Opus Lite video uploaded shortly, however my upload bandwidth on DSL is painfully slow and I gotta find the video file in my archives...

Finally, to plynch27, I just wanted to say thank you. Thank you for doing this for me!!! :thanks2:

 :mandel:

My orignal MP4 video "Magnum Opus Ex Lite" zoom movie is currently uploading to Youtube. It should be live in another 6 to 7 hours if all goes well:
http://youtu.be/0WChGAK_mYg

Nice to see that you are back stardust!
I hope you will discover the amazing findings by mrflay (K.I Martin, no longer active) completed by Pauldelbrot! And I hope you want to try my implementation of it, Kalles Fraktaler 2, which is more than 100 times faster than traditional programs on deep locations. Or Mandel Machine which is up to another 100 times faster than KF on some locations, but unfortunately not yet complete.

How does this new program work? I'm skeptical that it's really 100x faster than Fractal Extreme.

Fractal Extreme was for years the fastest Fractal Rendering software out there because it utilized 64-bit longnum integer math routines written in x86-64 assembly, with various redundant calculations removed from the equation. Most competing fractal proggies used 32-bit calcs (4x penalty) written in a slow language like C (more penalty) with little low level speed optimizations (again roughly 4x penalty). I have often wondered, however, if faster implementations were possible. Julias and Perturbations have interesting properties when set to a point many many zooms deep within the set. I tried doing (0,0) Julia zooms with the seed set to the final minibrot, which allows virtually identical formations with less repitition at lower zoom depths, but it required higher precision that the software allowed at current zoom level, causing the render to fail.

Another thing I've thought about, once you're far away from any minibrots, and with the type of images I've been rendering, zooming into the centroid will double and quadruple the surrounding patterns, ultimately leading to a minibrot. Zooming into a point off of the centroid eventually leads to a repeat of all the patterns previously seen, until ultimately leading to an image resulting from the original pattern being wrapped twice around the chosen point of inflection.

Ultimately all of the intricate patterns I have created were based on blazing trails based upon te above principal. I propose it would be possible to render such detailed Mandelbrot patterns, not by zooming 1000s of zooms deep, but simply by selectively rendering key points around such a feature and distorting the shape in such a mnner as to emulate the natural contortions present within the Mandelbrot set.

To better describe this concept, think of zoom movies not as video files consisting of vast numbers of 2D frames, but as a cyclindrical tube. The tube is essentially a bitmap image of X pixels in width (distance equal to 2PI or 360 degrees) and arbitrary height. Moving down the bitmap equates to deeper zoom, moving up equates to shallow. The height is a log scale, so features get progressively larger or smaller. Start to finish from mandelbrot to minibrot you have your entire zoom sequence mapped out on a 2D cylindrical plane. Take a cross section of this plane 50% between the start and finish, and compare it to a slice 75% between start and finish. The pattern will be the same but it wraps around the zoom cylinder twice instead of once. Furthermore, if we go 87.5% to the end or 7/8, the pattern will repeat around the zoom cylinder four times. At 15/16 to the end, 8 times. 31/32 to the end, 16 times. The features become smaller and smaller until it reaches the event horizon, the minibrot which is like a black hole in that even an infinite amount of iterations can escape it's influence.

Now that we can visualize the zoom sequence as a flat but finite cylinder with repeating patterns rather than a large fractal plane with infinitely small features, the mystery of the Mandelbrot is largely unraveled. So, since the patterns repeat ad-infinitum, it should be possible to distort the ether that is the Mandelbrot Set at relatively low zoom depths, and remap these points to duplicate the itricate imagery found much deeper within the set, without doing the inane, tedious, and computationally astronomical work of zooming thousands of frames deep calculating millions of iterations per pixel.

I hope that makes sense. Just contort the set to produce whatever feature you want to create, then render just that, rather than brute force the pattern by relentless zooming. Sit and watch the entire 30 minute video and ask yourself if zooming through the exact same patterns countless times is the best strategy to create/locate these intricate formations. No matter how powerful the computer, you'll always hit a brick wall.

That was why I gave up on the deep zoom thing. I figured there's got to be a better way of doing it than the old iteration loop until bailout. Just my luck I leave the scene for a moment and somebody discovers some new rendering technique. Just some food for thought. Suns coming up. Gotta get some sleep now.

Magnum Opus Lite (the slightly shorter zoom video I've been sitting on since April 2012) is live:
http://www.youtube.com/watch?v=0WChGAK_mYg

EDIT: For some unknown reason, the forum is borking the standard Youtube url. Try this version instead:
http://youtu.be/0WChGAK_mYg

We have used this image, zoomed to 6e141, as benchmark:
http://www.fractalforums.com/images-showcase-(rate-my-fractal)/tick-tock/

On my machine Fractal extreme is rendering it in 10 minutes in 640x360.
Kalles Fraktaler is rendering it in 4.5 seconds - that is 133 times faster.
And Mandel Machine is rendering it within 1 second, i.e. more than 600 times faster!

If you want the details of the new rendering method look at this thread
http://www.fractalforums.com/index.php?topic=15559.0

It is a very long thread, so in short - only one pixel is calculated with full precision and used as a reference to calculate all other pixels in the image with only hardware double precision.
Further a truncated Taylor Series can approximate a starting value for the low precision pixels that enables a significant amount of iterations to be skipped.
But unless this reference pixel is carefully selected and the pixel's differences in iteration counts are low, this rendering method cause errors because the double precision is not enough.
That issue was solved by Pauldelbrot that discovered that the orbits of the reference and the low precision pixels are "snapped together", so by measuring the relative distance between them the incorrect pixels can be identified and recalculated with additional references.
http://www.fractalforums.com/index.php?topic=18908.0

This is an awesome development! I downloaded Kalles Fraktaler and explored well past e600 or so. I was watching the images render on the screen and there is definitely something different going on with the pixels. The iteration count nearly instantly goes up to some amount, then jumps again filling the holes as it renders. It's as if I'm watching the iteration count increase as the image renders. Just the interface seems a little clunky and it will take a bit getting used to a new software interface after using the smoothe Fractal Extreme for so long. Gotta check out Mandel Machine as well...

 :peacock:

Thanks. Yeah, I haven't used fx much so I did many things differently.
And I do not want to belittle fx since I am certainly not able to do it faster in the old way, and I know it is indeed possible to do it much faster in the new way, as Mandel machine implemented in assembler instead of C++ and utilizing SIMD for parallel operations within each thread.
But so far Kalles Fraktaler is the only available program that solves all the glitches automatically.
I might implement SIMD myself one day, but it requires restructuring a big part of my program and debugging is very time consuming :)

Well I've gotten off my Kester and installed Kalles Fraktaler. I rendered a frame of Magnum Opus Ex formation at 1920x1080. Took an hour on my quad core AMD laptop. Is there a way to change the aspect ratio as it appears stuck in 16x9. I'm gonna do some heavy rendering on my older but massive 8-core rig and see what I can churn out. It's got 4.2Ghz OC and is about 3x as fast with FX as my laptop.

((1920*1080)/(3200*3200))*(28*24+18) / 1 = 140 times faster, not considering different hardware. Cool :)

Arbitrary ratio is not currently possible, because the main target is movies, and that it requires a lot of work.
But it is indeed possible...?

Addition: KF is using long double on locations deeper than e600 which is about 3.5 times slower. Botond has shown that after the approximation ordinary doubles can be used for this location. Further much more iterations can be skipped if more terms is used for approximation. And SIMD would of course speed up with the same amount as the grouping sizes. So it is possible to render this particular location 10,000 times faster than FX!!

My 8-core beast rendered this image (3610 zooms) in 32 seconds using Mandel Machine. Days on FX literally became seconds!!!
http://www.fractalforums.com/images-showcase-%28rate-my-fractal%29/pterodactyl-vertebrae/

Sadly Mandel Machine caps at 3700 zooms so I am continuing my quest for deeper images in Kalles Fraktaler. But without aspect ratio adjustment or rotation, sometimes I get stuck, for instance if I'm trying to render an oblong shape that happens to be vertically oriented. :tongue1:

Also zoom progress is somewhat slow because I have to wait for the feeder pixel to compute with every zoom advance.

I'm probably done with the X motif for now (yeah said that one before, ha!), but to anyone interested, have a look at this little number...

"Seven Ex Magnum" - seven repetitions of the Magnum Opus Lite feature. 5210 zooms deep in Mandel Machine!

Fractal Forums (parameter file also available for download):
http://www.fractalforums.com/images-showcase-%28rate-my-fractal%29/triple-ex-magnum/msg76128/#msg76128

DeviantArt (high quality view/download):
http://stardust4ever.deviantart.com/art/Seven-Ex-Magnum-475692041

It's weird in a way how images that used to take days, weeks, or months can now be rendered in minutes. Finding an object now takes far longer than rendering it. Like, the first time when I spent over a month rendering the Magnum Opus Ex image, and even upgraded my PC in the interim, it really meant something to finally see it realized. In truth, I have Magnum Opus Ex (X of X of Xs) framed on my living room wall next to XX Reactor Core (the original X of Xs). Now that it seems like we can just zoom as deep as we want to, and get instant gratification for doing so, some of the magic is lost. Feels almost like cheating in a way...

I guess what one really needs to ask, is "Where do I go from here?" Abs() Mandelbrot formula variants?? 3D Mandelbulb/Mandelbox???

Oh and one final thing, thanks for getting me back into the fractal scene. :toast:

I just gotta contemplate what to do with all that extra speed boost...
:cantor_dance:

I get what you mean. It must be like collecting stamps as a child and one day receiving a briefcase from your uncle, full of stamps...  :hmh:

But it can also be seen the other way. By rendering time not being a limiting factor, the creation of interesting images depends more on one's creativity and aesthetic skills rather than patience. We have seen too many "Deepest zoom ever" animations and stills from deep but boring regions of the Mandelbrot set.

You also said that when posting your XXX Cold Fusion Cell image (http://stardust4ever.deviantart.com/art/XXX-Cold-Fusion-Cell-154728253 (http://stardust4ever.deviantart.com/art/XXX-Cold-Fusion-Cell-154728253)), just to come with Magnum Opus Ex later...   :tease:

Magnum Opus Ex started with an almost identical location to the Cold Fusion Cell render. :dink:

I can continue to zoom into the center of the Seven Magnum formation which will give me a rough "X of Xs of Xs of Xs" shape, but the central formation would have eight arms instead of four.

Or if I back up to the Triple Magnum render and use the edge of the fragment at the end instead of the center, it will eventually create something similar in concept to the Magnum Opus Ex fractal, but it will be an X of Xs of Xs of Xs. One caveat is due to the zoom methods, the center formation will have a big X in the center, then 3 smaller X of Xs, then a bend snake, then three more X of Xs, then a tiny bent fragment on the edge. The X of Xs of Xs of Xs formation would then have the three standard X of Xs of Xs (Magnum Opus Lite shapes) in each arm (12 total) followed by the bend fragments comprised of X of Xs formations ('XXX/XXX') on the edge of each leg. The stripe half way up each leg will be an artifact of the methods I used to create the Magnum Opus fractal and cannot be corrected at this point without unwinding and choosing a completely different zoom path. The image I have envisioned will need to expand the maximum zoom depth of Mandel Machine to at least 7200 zoom levels, ironically barely within the zoom limits of the original Fractal Extreme. I imagine it would also be possible to use Kalles Fraktaler, but by guesstimation the render would likely take several days to about a week to complete. Mandel Machine would take a couple hours tops.

At some point however, the images will become so intricate that much of details will be lost unless the image is rendered to gigantic resolutions.

Here is one way to go from now:
http://www.youtube.com/channel/UCupBhu5TVx3RNK3iUIdeVlg

This guy have made some amazing movies with my program. To bad YouTube has too hard compression. Because I really like patterns that appears over the spirals because the palette is synced with them.
I can hardly imagine what we can do with references with billions of iterations and perturbation, series approximation, SIMD, glitch solving and carefully customized coloring all together :)
But one thing is for sure - depth is not so fascinating anymore.

So please keep zooming :)

Here is a little attempt also from me
http://www.youtube.com/watch?v=Cik4y4s-HIM

Yeah depth only impresses so much. I have pretty much mastered the art of multi-replicating shapes, which can create mind-blowingly intricate stills, but now I need to branch out by picking different styles of inflection points. Zeep zoom movies of these multireplication points are actually boring and repetitive. Some of the better zoom movies are actually made by whimsically picking seemingly random locations for inflection points, creating a varied look throughout the whole zoom video.

And yes, I'll keep zooming! O0