Scanned Spring 1996
Created November 27, 1996
Last updated May 1, 1999 (added Milk and Alice)
Go to the image gallery
When MIT's Senior House was renovated over the summer of 1996, the murals that graced the walls of the old building were lost with the rubble.
The architectural layout of Senior House made photographic preservation of the murals very difficult. The whole point of the murals was to decorate the spaces where people lived; these spaces, in rooms and hallways, could not be photographed with great precision because of tight physical constraints. In spite of this difficulty, many people did take pictures of the murals, and one group even combined them into a DOOM-like computer walkthrough. These photographs best preserve the murals in context of the rest of the house.
I wanted to be able to record the great detail and intricacy of some of the murals in a way that "context photographs" would not. Many murals in the house were the artist's pride and joy, a project of self expression and release from the rest of the MIT environment. This care can be seen in the color choices and individual brush strokes of their works.
To record at least some of this detail, I developed a scanning system that could record image detail from the walls of Senior House directly using a desktop page scanner, a scanning jig, and a laptop computer. Essentially, I turned the scanner "inside out", using it not as a stationary object for scanning moving paper, but as a moving scanner for imaging a large wall. The scanner was mounted to a Plexiglas jig with rails to that it could move over the wall. An entire wall surface could be scanned in long 6-8 inch strips. Resolution of the images is very high, about 100 dpi. (I will add a picture on the scanning system as some point).
The scanning process was fraught with difficulties. The walls of the old building were not very flat, so the scanner sometimes got caught on bumps or blemishes. Such a catch would result in a change in scanning speed and thus a local change in scale. Since each strip would have different bumps, alignment of adjacent strips became almost impossible. In later scans, I added a metal tape measure on the edge of the field of view to make alignment easier.
The data size of the images is also a problem for alignment. Each strip is about 850 pixels wide by up to 3000 pixels long, in 24 bit color. Even scanning one strip was difficult because the only available computer, owned by the late Professor Muriel Cooper and graciously lent by the Visible Language Workshop, had only a small amount of RAM. All the scans were done in the last two weeks before the house was closed and demolished. The Alice in Wonderland mural on Runkle 6th was an all-nighter of scanning on the crazy last night before the demolition crews moved in. Not a lot of room for error.
Piecing together the data from the murals I scanned is a time-consuming and computationally tedious process. The scanning process was good enough, though, that the raw scans can be pieced together without too many mismatches and scanning artifacts. Below links to some examples of the mural images. As I get time, I'll stitch together more mural images and work out some of the problems of the existing ones.
Atkinson 4th kitchen
Text and images Copyright 1996-1999 Michael Halle, ARR.