# :BOXING MOTHLIGHT

*Boxing Mothlight* applies the procedure of box counting, often used in fractal analysis, to the avant-garde film Mothlight by Stan Brakhage.
Both the image track and the audio track are generated using this mathematical idea. The result is a hypnotic meditation on the relationship between
natural and digital texture.

The box counting method can be employed to analyze graphical patterns. It is particularly interesting when applied to highly textured images, such as those in Brakhage's work, a film made using insect wings, leaves, and other organic matter. The mathematical concept is closely aligned with the visual aesthetic of the source film.

The basic idea is very simple. The first step applies an edge and corner detection algorithm to every frame in the movie. Every pixel that represents an edge or corner is converted to black. All other pixels are converted to white. The result is a binary image, where every pixel is either black or white. The next step covers this binary image with a fixed grid of non-overlapping rectangular cells. If within one cell, there is at least one black pixel, the entire cell is rendered black. Otherwise, the cell is rendered white. The process is repeated with grids consisting of a greater number of smaller cells. This video consists of nine images generated in this manner.

The image on the top left represents a covering of the frame with a very coarse grid, consisting of only four cells. As we move from left to right and top to bottom, the number of cells in the grid is increased (4, 16, 256...) and the size of each cell is correspondingly reduced. The nine images represent progressively finer representations of the same image.

The soundtrack is produced using a similar procedure. The source is an audio recording of bees. This clip is divided into various non-overlapping fixed-length segments. A Discrete Fourier Transform is performed on each of the segments and all frequencies are filtered out except the zero frequency. The signal is then reconstructed by appending the filtered versions of the various segments. The procedure is then repeated several times, always on the original clip, using a progressively larger number of (correspondingly smaller) segments. In this way, we obtain a sequence of clips that represent progressively finer representations of the original sound.

The soundtrack is made by concatenating these clips.

No attempt is made to synchronize the sound with the image on a frame-to-frame basis.

Artist and image/audio programmer: Hector Rodriguez

Sound mixing: Sam Chan

Mathematical adviser: Felipe Cucker