Snow – A Reactive Environment Installation
Snow – Interactive Installation from chris teso on Vimeo. Commercial produced North.

Project Client

City of Portland Downtown Marketing Initiative in conjunction with North.

Project Goal

Create an outdoor reactive environment in which passersby could interact, based on their location and movements, with artists interpretations of snowfall in Portland.

Project Strategy and Equipment

Display – Custom made acrylic panels lined with photosensitive film built and erected to exact dimensions of installation window
Projection – Rear projection with flipped signal using a 5200 lumen Sanyo PDG-DXT10L Projector
Video CaptureLogitech QuickCam® Vision Pro
CPU – Mac Pro Quad
Application – Flash AS3
Industrial Design – Two large tarps sewn together to create light blocking canopy. Lining to seal off windows and acrylic. A shitload of Velcro.

Project Synopsis

As part of a larger campaign to brand Downtown Portland North was charged with creating an outdoor reactive environment in which passersby could interact, based on their location and movements, with artists interpretations of snowfall in Portland. Three traditional artists were commissioned by The Portland Institute for Contemporary Art to create their interpretation of snowfall in downtown Portland. These pieces were then transformed for use with my motion detection and tracking application built entirely in actionscript. The application was built to display 3 different scenes with randomly generated snowfall. The three scenes were set to rotate on a time interval. The application used motion detection to make the falling snowfall react and animate based on the location of an individuals movement. The application also incorporated and automated snapshot function that took a photo every 10 minutes and posted it to a private flickr account. This function was implemented solely for remote monitoring, ensuring the application was up and running. After going to the installation location and taking careful measurement of everything from window frame to projector distance, the installation was built and staged at North. We built a ‘to scale’ model of the window in the back of the office. This was necessary for accurate calibration of projector, and optimization of motion detection and tracking code. The staging was an extremely helpful and necessary step in eliminating early stage bugs such as projector calibration and camera positioning. We quickly learned the maximum distance for a USB signal, the proper use of DVI vs. DHMI to projector and the challenges of doing motion detection in every possible light range from complete darkness to bright sunshine. After testing stage was complete we moved the entire rig over to the installation spot located at 6th and Alder in Downtown Portland. We went about setting up a canopy to block out all extraneous light leak. The interior of the installation space was two stories of extremely large windows. The canopy we chose was a homemade compilation of two tarps strung up to the surrounding walls with rope. Since there was very expensive equipment residing under the canopy, the fear lingered of canopy failure collapsing down, resulting in devastating loss. We next set out to install the panels into the window frame. Since they were pre-cut to exact dimensions these flowed right in smoothly with only small light leaks needed to be sealed. We then hooked up the projector, cpu, camera, monitor, and application and set to testing. Testing in a staging environment is one thing. However, most tests are rendered moot when on-site. I quickly learned that reactive environments are just that, entirely dependent on their environment and all the challenges that come along with it. The motion detection is to run 24/7 for six weeks during all weather, lighting situations, and process location accurately with varying levels of distractions. Since the installation is located on an extremely busy downtown street corner, distractions ranged from traffic movement, pedestrians, bikers, storefront lights and weather. For testing purposes I built many functions that allows the system to be calibrated without actually going into the code, but rather using an external UI to update blurring, light detection levels, contrast, and other variables. After several days of calibration, code tweaks and testing the application has been running uninterrupted and has gained quality exposure for Downtown Portland. Many fine citizens of Portland have had a unique interactive experience, and have consequently spread kind words about the project. Furthermore, the project is a successful case study on the implementation of new technology and reactive medium into a larger branding campaign.

Project Challenges

Development – Create a motion detection application that worked in every possible light and weather condition.
Design – Making traditional raster art work work in a complex location detection application and animation.
Industrial Design – Create an environment that optimized presentation and functionality in varying environmental variables.

More information on the design and development process