In order to obtain a driving algorithm for the Intelligent Dance Agent, dancers were filmed from above. Several exercises were tried to get a sense of what a dancer actually does when they react to their partner. The dancers were Elyssa Dole and Cristina Schenk.

Exercise 1 – Free Dance:

The dancers were asked to simply improvise a dance duet with each other. They were told to play close attention to where they were on the stage and where their partner was.

Watch the video in [streaming Quicktime] or [avi] form

Exercise 2 – Automaton Interaction:

In this exercise, one of the dancers was asked to move around the stage completely oblivious to the other dancer, acting as a sort of naive automaton. The other dancer was asked to dance and interact as much as possible with their unheeding partner. The purpose of this was to accentuate the visibility of how a dancer reacts to their partner.

Watch the video of Cristina as the automaton in [streaming Quicktime] or [avi] form

Watch the video of Elyssa as the automaton in [streaming Quicktime] or [avi] form

Exercise 3 – Automaton Switching:

This was essentially the same thing as the Automaton Interaction exercise, except at various random points within the dance the dancers were told to switch who was being the automaton. This was done several times so as to pinpoint what happens to movement when a dancer changes from acting to reacting.

Watch the video in [streaming Quicktime] or [avi] form

Exercise 4 – Opposites:

In this exercise, the dancers were asked to try and dance as differently from their partner as possible, making gradual changes in their movements so as to experiment in several types of opposites.

Watch the video in [streaming Quicktime] or [avi] form

Exercise 5 – Contact and Separation:

The dancers were asked to experiment with what happens when they got very close or very far away from each other. They tried different types of contact interactions and different ways of coming back together.

Watch the video in [streaming Quicktime] or [avi] form