Shan Huang

16 Jan 2014

Project I profoundly admire:

The V Motion Project

I remember seeing many digital scratcher projects done in a full array of interaction techniques (touch screens, Leap Motion, Kinect, etc…), but I find the V Motion Project to be the most amazing of them all. For one thing, the project allows dancers to control audio with their full body motion, unlike traditional scratchers that limit motion to hands. This brings much more expressive performance into the project and makes the show so engaging. I realize that they primarily use the hands as controls even though they track the full body. For example, nearly all buttons are pushed by hands. The distance between two hands and the distance between hands and the floor both determine some aspects of the audio. Other body parts (head, foot, etc) are rarely used. I think this is a wise simplification to make because if all body parts control something, the body becomes over constraint. Dancers may not be able to freely move their bodies without causing unwanted changes.

And ohhhh the graphics! It blows my mind. I love the tessellated shape of the body and all the tiny mesh-like fragments that emerge when buttons are punched.

Project that surprises me:

Karl Sims – Evolved virtual creatures, evolution simulation, 1994

This video shows a virtual creature evolution simulation from twenty years ago. The video shows results from a research project that studies Darwinian evolutions of virtual block creatures. Hundreds of virtual block creatures are created within a super computer (twenty years ago) and tested on several tasks. Only the outperforming creatures survive, after which their virtual genes and mutations of their virtual genes are preserved for further evolution. The video strikes me simply because the result seems to make sense┬ádespite the creatures look so absurd! I think it’s very impressive that their virtual creatures move and evolve in believable ways. I’m very interested in seeing what algorithms Karl Sims adopted to make this work though I wasn’t able to find more details.

Imagine what we can do with this idea with nowadays technology. With some distributed computing facilities like MapReduce, we can produce massive amount of such creatures (say millions) and test them against a gigantic skill set… So curious to know what the winning creatures would be and if they share similarities with real world species.

Project that could be cool but (slightly) disappointed me

A hundred years of rock in less a minute

This project has a beautiful concept and the results are rendered beautifully by an animation that walks people through the music genres between 1900 and 2000. After playing around with the beautiful graph for a while I was disappointed to find that no band information is shown at all. It’d be so interesting to see the bands behind each genre and how they changed their styles and influenced their peers overtime. After all it was these musicians that propelled the evolution in music.