The code above is using Away3D 4. I’m using the ‘LinearExtrude’ class for the height component of the object, and standard Plane primitives for the caps. The caps have a bitmap texture with transparency on them, which I have generated at the same time as the height component. A bitmap cap has the advantage of having a significantly lower polygon count than a tessellated extrusion cap, particularly for more complex objects, but has the disadvantage that the engine can’t do the same object culling with the cap triangles – because some of the faces are transparent, it has to draw everything that might be behind one of these planes in order to make sure that an object isn’t being incorrectly missed out.  Another downside is the resolution of the cap texture – you can see when an object is too close or when the angle is too oblique that the polygon cap is slightly blurry and doesn’t quite match up with the height component. It’s not perfect, but it’s working well enough to keep going with.

I’m really pleased with how Housifier’s behaviour is working out. It’s tremendously satisfying to see the parts that will obviously fit together to make a working object. Something I read recently made me think that I won’t want to include such a prominent view of the pieces at all. After reading James Bridle’s post on the fact that book publishers are promoting the wrong things, I realized I was promoting the wrong things, too. It’s nice for end users to know that the engineering logic exists, and they may even feel clever for being able to make sense of the pattern. Ultimately, though, they’re more invested in what it’ll be like to have a shelf in their house, what they could put in it and what it’ll look like on their desk. I made a quick model of a bedroom scene to place a shelf in. It definitely proves the point.

Even without filling up the scene with much detail, the mock-up makes the point of the system clear. You can imagine what it’d be like to have it in your hands as a final product. Imagine if you could resize the shelf while it’s on a desk, and put the junk on the table into the cubbyholes. You could even place the simulated shelf in a picture of your real room to see how it fit in with your décor. The prospect of being able to make an app like that is thrilling, but also means starting again with the interface!

Usage: Move the mouse around in the flash window to rotate around the model as it assembles. 

Glitch is pretty cool, both as a musical and a visual aesthetic. I think it formed as a response to so much of our experience being mediated by digital technology. Without noticing, we’ve become extremely familiar with the idiosyncrasies of digital delivery, and developed a sensitivity to what happens when it goes wrong. Anyone who has Windows should try listening to “Welcome to Windows” to see if it makes them feel funny when it gets to the alert sounds.

People like James Bridle and the people at BERG write a lot about post-digital media. Internet artists like Pixelfucks and Quickhoney (both moderately NSFW) cover the discussion visually, too. They like to go into the fact that now that we can do anything with bits and bytes, we can focus on what we want to do, rather than being restricted to what computers and pixels are good at. One of the things that I really like exploring is the impact of some of those old limitations of the digital form, particularly when those previously hard-and-fast limitations become more of a rough guideline, as the transition from 2D pixels to 3D volume pixels (voxels) suggests in a game like Voxatron. Something about pattern diffusion and the textured-but-clean surfaces they produce are extremely appealing. I’ve spent a lot of time in front of this latest experiment and find it very entertaining.

Are there any features of the old digital limitations that you have a fondness for? Has there been any adequate re-examination of what made those features interesting? I loved all the complex dither patterns on the black-and-white mac classic in the 80s and the lurid vibrance of palette-cycle animations in the early 90s.

Not a full-blown facebook app, but it is certainly entertaining – I’d love to see what you make with this one too!

Note: To use the smudge brush, click and drag in the direction you want to smudge in. For the pinch/bloat brushes, you can just click. They also work better at a larger size. Have fun!

I’m still working on Housifier, and one of the things that Stage3D is going to make available is detailed ‘normal maps’ . Normal maps are used by modern graphics cards to fake a higher level of detail on 3D models by changing the apparent angle, or ‘normal’, of an object’s surface.  I want to use this system to generate better textures for the materials that Housifier can use. Because Ponoko has so many materials on offer – and are updating their catalog continually- I need to be able to generate normal maps automatically in order to keep up. The math required to generate these maps is simple, but involved. In working it out, though,  I remembered that a 2D Displacement Map uses a very similar set of rules to generate distortions. On top of that, a number of discussions about Kaleidoscoper have ended up on the subject of a Displacement map-based solution, particularly for the funhouse mirror-like modes. I thought I’d give it a try and the result is what you see before you.

My question to you, the still-interested reader, is this: would you want this integrated into Kaleidoscoper? Is the app authorization process for Kaleidoscoper too demanding? Let me know in the comments below!

Knowing that it’ll work is still qualitatively different from seeing it work, though. Having the object in hand, and knowing that it was produced from an almost totally automated process is fantastic. I say almost because the part of the process where the shapes are packed into the minimum dimensions possible is not yet complete, but little by little the pieces are falling into place – not unlike the animation above!

A note on using Unity3D to view this: My apologies! I actually had this animation ready almost a week ago, and have been struggling with various conversion process from Blender and other applications through to either Away3D (broomstick) in Flash or some kind of webGL-based display. I could push through static geometry, but neither framework seems to have support for an animation format that I could find. If anyone has succeeded in converting Blender animations into something for Away3d Broomstick or any kind of webGL framework, I would love to hear from you!

UPDATE: I have replaced the Unity file with a Unity-produced SWF file! the down-side is that the files are relatively large, because the engine binaries are effectively packaged inside the swf. I look forward to playing around more with what can be accomplished with it!

Housifier basically consists of two parts: the ‘engineering’ part, which decides what shapes to make, and the ‘visualization’ part, which shows you what it’ll look like. They’re both equally important – while the engineering actually generates the object, it isn’t much good without being able to give people a realistic idea of what they’re getting!

Toward that end,  Adobe recently added hardware-accelerated 3d graphics to Flash. It means you can fit a lot more graphical fanciness into a flash app – I’ve been experimenting with it for a few weeks and I’m really pleased with what’s possible. There is a slight down-side, though – because the hardware acceleration is integrated into the system weirdly (it has to be underneath everything else), it’ll mean starting parts of Housifier from scratch. As you can tell from the proof-of-concept above, though, it’s definitely going to be worth it!

What it looks like

Icon for the page

I’m a big fan of M.C. Escher’s work  - in particular the tessellations – and have always wanted to come up with a way to handle the difficult-but-mechanical task of making sure that the tiles actually match up properly.  It’s a reasonably simple problem, so while the internet was disconnected this weekend I tried to build it, and so far it seems to be working!  I’d be very interested in either animating a blend from one set of points to another, or blending over distance (like something in Escher’s Metamorphosis series.) In any case, I 1. Have had a  a great time with it and 2. thought other people would like it too.  I hope that that is the case!

Simple house tessellation

A lot of things end up looking like either fish or birds.

(and this is the icon for the page)

A friend of mine bought a Fujifilm W1 a while ago, and has been taking very entertaining stereoscopic images with it.  The only problem with taking stereo pictures, though, is that it’s next to impossible for anyone without very special hardware to appreciate them.

Enter the “Jitter Gif”.  If you take the two images that are captured from the left lens and the right lens and put them into a fast-looping animation, you can get some sense of the space that they capture.  It’s demonstrated really well at 3erd.  While it’s possible to make such Gifs in Photoshop or a specialized tool, it’s often too much of a hassle for people to bother with – Photoshop has a serious pricetag and is total overkill if you just want to see what it will look like.

Toward that end, I have made this generator. Feed in your jpg frames and get out your gif movie.

This is an example of a jitter-Gif of the junk on my desk, although not taken with a proper stereo camera. Even so, you get a much more interesting sense of the objects in the scene! Note that you will want to shrink your pictures down to a manageable size before constructing the animation.