swing

swing - a "moving video"

Remark: The above link to the movie "Swing" is about 5.5 MB. It's in Avi Format. This is a very compressed version of the original. Here is a compressed Quicktime version of 12.6 MB. Both can be viewed with open source viewers such as e.g. Xine or mplayer.

About swing:

In analogy to the InSeries (i.e. the moving stills In2, InU2 and InKan) the videoloop swing is called a "moving video". It is an experiment which investigates the perceptory implications of mathematical information in connection with video data.

"swing" displays a "superposition" of (neverending) oszillations of different frequencies. One oszillation is given by a girl on a swing, another by mathematical transformations, the rest by music/sound. The oszillations may be thought as defining different timelines (regular ticks are providing a possibility to measure time). A human brain tries to join the different timelines, as you can see by watching the movie. The irregularities in the "oszillations" of the girl are emphasized by the regularity of the mechanically oszillating background.

swing had been featured by Doron Golan on DVblog.org (Oct 18 2005).

It had been shown at the Thailand New Media Arts Festival (May 2-4 2006).

It will be shown at the Electronic Language International Festival, Sao Paulo (Aug. 14- Sept. 3).

Mathematical description:

The involved transformations are again hyperbolic transformations, i.e. certain types of Moebius transformations (Cf. e.g. Dubrovin, Fomenko, Novikov "Modern Geometry" Part II for details). The two fix points of the hyperbolic transformations are the same for all transformations applied to the image. As for the InSeries the family is again constructed via sending one fixpoint to infinity, taking the log, using a sine-of-time function on the resulting universal covering, and then going back. Amplitude and frequency of the sine function are chosen in accordance to the given videomaterial, i.e. by an artistic choice. As one can see the sine stays in one fundamental domain. Again there is also no change of conformal structure on the torus ("twist") like in the Escher video (In particular the corresponding transformations wouldnt be any more hyperbolic in this case).

Acknowledgements: Again thanks to Tim for the help with the Java programming!