Cube Shadow Art....

In many respectives I find this pice of plastic a masterpiece…

It contains religious symbols however which are not up for discussion, neither intended, nor allowed due to forum rules.
Having shared it I don´t bother if the thread gets closed down for religious discussion but would be nice to hear some impressions on the sculpture.


(Action at ~1’00")

Just enjoy it, start blending and send it to shapeways :smiley:

Masterpiece indeed.
should be pretty easy to make, though i’m not sure if they can print such a complex shape.

i think it is really cool :slight_smile: crap video tho.

They way I see it its just another way to shove religion down our throats.

Nice use of shadows, would be very tricky to model that.

It would also be interesting to see how such a cube was put together (as I imagine parts it could easily snap)

Really cool. The dealing of the shadow is quite skillful.

It’s a very well thought out piece, most of taken a lot of time to design!

its easier than you think
it can in fact be achieved with just three planes
having more than three images in a cube will be the real challenge

I am with Eris, I think its rather easy to produce. 2 Shadowimages is really easy.
Basically you make one plane to cast the shadow, cut it in a checkerboard and then move the normals of the tiny planes and adjust their height until the light 90° turned casts the second picture.
The 3rd one gets tricky and beyond that I doubt you can do it without a computer simulation casting light and moving bits in realtime.
It might also be possible to use reversed stencil shadows. You could use images as shadows and cast the shadows back into a cube and let the rays cut it. should be possible to write an application which you feed 2 or more images and it returns you an object casting the shadows.

It would be nice to find out who created this sculpture and to know how he or she actually did it.

First build cube out of tiny faces and edges, then in top, bottom and right view delete faces corresponding to images you want leaving edges for stiffens.

@arexma
even the third image/symbol is easy.

A theoretical approach…

  • Make three planes, divide them into,say 20*20 grids. Orient them at 90 degrees to each other (along x,y,z axes)
  • Now you have a set of coordinate planes, that define a 3-dimensional grid (a virtual grid in our case, but obviously needs to be physical for a practical implementation)
  • Hollow out squares from each of the three planes to form your three images.

Your masterpiece is ready.
For the “wow factor” and to make it look complex, give all squares random displacements (in unit multiples) along the normal into the grid.
If you think about it, squares belonging to any one plane cannot obstruct the when viewed from other angles, only their sides will be visible.

Just like when you have three perpendicular planes, and you view one normally (face-on), you only see the edges of the other 2.
Our squares retained their directions and were only displaced inwards.

That’s most definitely really cool,and really artistic. Creating more than 3 images with A CUBE would be impossible, considering you only have 3 directions (unless you’re just amazing and can do it from angled views also)

The only way I think you may be able to do it for more than 3 shapes is by taking more complex shapes such as octagons, that would possible give you the more angles needed to do it.

Just my thoughts.

I agree with you both. 3 shadowcast images is a piece of cake.

But more than that is not impossible. No one says that the shadow has to be cast by light entering 90° to a face. you could cast it 45° into one of the edges.
But then it gets really tricky :smiley:

In my post I did provide a counter statement saying that if you’re good enough it could be done at angled views. So, we’re basically on the same page.

Almost.
Your reading skills seem to be superior to mine lol. :smiley:
Sorry either overread the remaining sentence, or forgot about it being caught in my thoughts…

I tried to blend a bit, but one thing that´s making it tough is that you´r very limited by the remaining geometry.
If you got for instance a smiley in the topplane, and then try to do something very volumetric in the right/left plane you can´t really make it because adding “shadowcasters” there would change the shadow of the smiley.

Well, making these “shadow casters” I assume wouldn’t be too hard. It’d just be a bit like playing this game: http://www.freepuzzlegames.biz/game/455/AP-2-Puzzle.html

If you set up your 3 shapes, in a 3d game of this, I assume it’d be quite simple really to make them. I couldn’t find the actual game but this was the closest I could find. I may, or may not actually attempt to make one. But in theory, I believe it’d be fairly easy, especially the way they made them.

Okay, so I decided to make one, I’ve only gotten started, but if you want to see how to start (or at least how I would approach it) then here’s a picture of what you should achieve and a description of how to get to it.

Okay, so make a cube, go into edit mode, and subdivide it to your hearts content (be noted that with each subdivision it’ll get EXPONENTIALLY harder, just a warning, which you should honestly know by this point if you’re attempting this fairly easy task anyway). Select everything (a) and then hit delete, and select FACES ONLY, not faces. this should leave you with a basic wire frame of a cube.

Then make sure Auto merge editing is on (mesh > auto merge editing), along with snapping (Shift + Tab, or click the Magnet). Then go into one of the NumPad Views (1,3,7) and select the vertices of one of the sides. If you select the right side of the cube, you’re going to want to extrude it to the left be sure to snap it at every interval that there is a vertice (whatever doesn’t make sense might be explainable by the picture), being sure to lock your extrusion to the direction you’re extruding it. Continue doing this for all 3 directions: X, Y, and Z.

Kind of complicated, probably an easier way… but… here ya go. You should end up with what is pictured below.

Attachments


I think you misunderstood me…
The problem is the limitation of shapes that remain to do after… oh nevermind. I just got enlightened. For the sake of the discussion… that was my issue:

http://img535.imageshack.us/img535/2947/shadowface.th.jpg

You could randomly remove cubes from the smiley “layer” so it casts the shadow of a smiley.
If you turn to the side then (bottom right view) there is not much left for the other shadow and adding other cubes would interfere with the smiley shadow.

However I lost track of the fact, that he obviously uses the base cube grid and planes which he hides behind those grid beams for the views…
So the leftside “shades” appear as lines in the frontside view.
and leftside and frontside “shades” appear as lines in topview.
And wherever you got a section with shades from all 3 directions you can make a cube.

Basically this and my last post are brainfarts :smiley:
runs off

k, the bottom one was the first one I did, and I was half watching battle star galactica, half doing this, which is why the other 2 didn’t do so well.

  • create a new plane (plane should be facing in the z direction by default), then add 2 modifiers on it: 1 array of 15 in the x direction , 1 array of 15 in the y direction
  • click apply on both arrays - now, when you go into edit mode, you will have a grid of separated planes that you can move up and down in the plane normal direction.
  • cut out your pattern
  • select random faces ( I selected 30% random), and move these faces along the blender grid with snap turned on (I moved up 2 up every time )
  • move until square is complete
  • repeat for other patterns in the x and y direction.

you can double check your work by using the front, side, or top views (numpad 1,3,7)

Attachments


shadows.blend (279 KB)

AMDBCG, I actually like your way more.