Title: red room, dreamscape, a deepness in the sky Post by: lycium on November 11, 2006, 05:22:57 AM finished off my first entry, not sure if there'll be more (lots of coding atm). hopefully i can put some new 3d fractals together before the end of the month :)
edit: finished off another; linking here rather than embedding since it's large. if it's somehow too large for comfort i can provide a smaller version (or you can use a high quality resizer such as irfanview or photoshop etc), and if it's too small i have a super smooth 15392 x 9632 version :P edit2: last one added, and some info on how i made them. i have a feeling that posting my entries now will make them more "ho hum" by the time the judging happens :/ edit3: hot damn, i'm some kind of moron, mistaking euler for euclid! edit4: euclid is no longer being entered sniff, replaced with "a deepness in the sky". red room: http://www.fractographer.com/wip/redroom.png dreamscape: http://www.deviantart.com/deviation/42886472/ a deepness in the sky: http://www.deviantart.com/deviation/43245472/ technical details: the first two are made with my 2d ifs app, lovingly/stupidly named "ham and mustard" (you can get some executables here: http://fractographer.com/binaries). they're made with multiple near-analytically antialiased (very smooth) layers, and everything is hardwired into the code for now :/ one day i want to add a gui and various other usability features for a proper release. the last entry, a deepness in the sky, is made with my 3d ifs app, which is currently nameless :( a great deal of mathematical/numerical analysis went into obtaining the "sketched" look while still being able to scale the resolution (more info in my later comments). most of the computational effort went into the extremely high quality postprocessing (all done in my application, no photoshop etc - i can readily produce an exe should this claim be challenged ;) which i'm currently making ~1000x faster to scale to the high supersampling rates employed by the newer, smoother variation i'm currently rendering. however, i really like the gritty look in the original so i've decided to submit it instead of the other version. Title: Re: red room, dreamscape, euler Post by: heneganj on November 14, 2006, 12:07:27 AM Euler = 404 error
Others look superb. Title: Re: red room, dreamscape, euler Post by: matera on November 14, 2006, 03:37:17 AM Yes, euler has gone AWOL :(
Hmm, ham and mustard sounds tasty :) Title: Re: red room, dreamscape, euclid Post by: lycium on November 14, 2006, 09:10:59 AM fixed; what a retarded mistake on my part!
oh and thanks all round :) i have some testing versions of h+m available here: http://www.fractographer.com/binaries/ Title: Re: red room, dreamscape, euclid Post by: heneganj on November 14, 2006, 02:24:41 PM Euclid is an inspiring image - it's very simple but I see it used as a backdrop to some magazine article about space-time-relativity-the solar system etc that sort of thing. In a way it looks like someone drew star trajectories and made a map of the sky.
The other interesting parallel with this image is that you can zoom in to reveal more detail, more trajectories, more maps - just as you would with a telescope. Very thought provoking - that's what Art is all about! Title: Re: red room, dreamscape, euclid Post by: lycium on November 14, 2006, 03:25:54 PM Euclid is an inspiring image - it's very simple but I see it used as a backdrop to some magazine article about space-time-relativity-the solar system etc that sort of thing. that's no coincidence btw, they have a hyperbolic geometry connection; see for example http://en.wikipedia.org/wiki/Hyperbolic_geometry for a basic overview and http://en.wikipedia.org/wiki/Minkowski_space for the actual connection. if you happen to have it, roger penrose's "road to reality" contains a very good presentation of several hyperbolic space representations, illuminated by m.c. escher's incredible mathematically-themed artworks. The other interesting parallel with this image *chortle* http://en.wikipedia.org/wiki/Parallel_postulate is that you can zoom in to reveal more detail, more trajectories, more maps - just as you would with a telescope. some of my newer renders are exhibiting an extreme amount of hierarchical detail: overall shape, details, lighting, etc; i'm going to look into making really really deep zooms of http://www.deviantart.com/deviation/42922304/ for example; a version featuring slightly more complex colouring and shading is rendering now, and the details are even more apparent. perhaps i'll enter that into the competition rather than one of my previous pics... it's difficult to gauge what sort of style is appreciated though! Very thought provoking - that's what Art is all about! i fully agree, and am happy you noticed the cough parallels mentioned previously ;) Title: Re: red room, dreamscape, euclid Post by: lycium on November 18, 2006, 10:31:56 AM my final submission, "a deepness in the sky", is available here: http://www.deviantart.com/deviation/43245472/ old info follows: that's made with my 3d fractal rendering app (as yet unnamed); the main feature which gives the sketched look is the alpha-mapping: since the surfaces in my app are bi-parameterised i can map any 2d function to them, and because of the way the program works it's also very easy to get the partial derivatives of any such mapping with respect to the image co-ordinates. i use this to get an estimate of the surface parameter-space ellipse, which in turn i use to sample a fractal noise function - didn't use anything too fancy here (like for example real turbulence), only statistical adding of various gaussians according to the "squashedness" (eccentricity) of the major and minor axes. when this mapping is stretched (rescaled non-uniformly), it produces those nice looking streaks. the function, in turn, dictates how opaque/transparent the local surface element is. well, that was my fun for most of last night :P i may submit another version (i've already rendered 2 others, both at 1920x1200) later once i've revisited the colouring and sped up my glow filter enough to be actually usable. edit: uploaded a testing "disco edition". right now i'm trying to raise the resolution and supersampling without losing the sketched look, it's incredibly difficult since there's some normalisation factor in the jacobian induced by the mapping (change of variables) i'm not taking into account... need to whip out the notepad... edit2: changes have been addressed, final version is rendered, submission info above changed to reflect this; the new version has been renamed "a deepness in the sky" for its similarity to a previous work of mine, "a fire upon the deep". work is ongoing to render a supremely antialiased version, but since i like the grainy effect so much that will be my submission. Title: Re: red room, dreamscape, a deepness in the sky Post by: bradorpoints on November 30, 2006, 09:53:03 PM Euclid is great. It could've been a prop if Terry Gilliam had gone ahead with a sequel to "Time Bandits" - the map of the universe.
Title: Re: red room, dreamscape, a deepness in the sky Post by: lycium on December 01, 2006, 02:51:45 AM thanks mate, glad you like it! it's a pity that the compo deadline has passed now, and i retracted it :/ btw, a bit of justification on that move: while i actually prefer euclid to redroom in some ways, i thought it doesn't exhibit enough "fractalness" - even if it was computed using the exact same (fractal) algorithm. spirals made of self-similar curves (as in redroom) are somehow canonically fractal, and i guessed that might score more highly. i'm kinda starting to regret doing that switch, and i guess if given the option to use euclid instead of redroom i'd take it... |