Working Demo

For this week I put together a working demo that successfully maps gestures performed on my spandex wall to a filter of ambient noise in MaxMSP.  Getting to this point required figuring out the right interaction fabric so that a user could push into it without touching the wall, setting up a provisional lighting and camera set-up, and setting up white walls behind the fabric for contrast.  It also took me a while to navigate Jaime Oliver's patch and successfully route the data to MaxMSP via a network port.  Here is a video documenting the progress so far, with a simple right to left gesture mapped on to a bandpass filter of some cafe noise:

I had to take down the installation for this weekend's ECM concert, but I can get back to work on Sunday.  Starting on Sunday, I will install a stronger light and set up the second camera for the vertical axis.  From there I'll scale up to build my second wall and elaborate the sound design.

A quicky this week:

Been working hard at using water as an interface, finally made some positive steps in that direction this afternoon.  The first thing is a video of a mock-up of the actual patch that uses two video sources.  The real patch uses a webcam on one side, but it was hard to make the video and move the water at the same time, so I made a quick recording of the water and used that recording to show the processing.  The patch takes two sources and uses the method of lumakeying to blend them.  A lumakey essentially takes all the pixels within a certain value of one video and makes them transparent, thus revealing the second video behind.  So here's the video of the screen shot:

And the second is a video of the result projected on to a 4x8 piece of black burlap from about ten feet away.  It makes the weird masking decision on the previous video make more sense: