The learning of movement from video is exhausting me. I remember this from when I used to haha choreograph. Somehow I’ve got quite good at doing this method of capturing movement onto bodies. Nonetheless, it is painfully slow, mentally grueling and usually we have to re-remember it the next day. And especially when there is so much more to do it feel like panic.
But, we’ve tried a few things and … had a couple of conversations and sometimes just get quite dizzy about what it is we’re attempting. I seem to be either rehearsing, reading, writing or thinking about this project continually until Saturday morning when I plunge into a stupor … or maybe a coma.
I had a chat with Sarah Maddison yesterday, about her paper Galactic Cannibalism: The Origin of the Magellanic Stream that I was particularly interested in because of this sentence: “To simulate the merger of the Milky Way–LMC–SMC system, we used the TreeSPH code of Kawata (2001) which includes a self-consistent treatment of self-gravity, gas dynamics, radiative cooling, star formation, supernova feedback, and metal enrichment.” (Requiring a search for Kawata’s paper today…). I liked the idea of having a system that included a bunch of attributes in some way that would also change over time and this, along with images from the LMC led me to Annie Hughes’ paper, A multi-resolution analysis of the radio-FIR correlation in the Large Magellanic Cloud.
Which led to Sarah explaining to me about some things like column density, how red shift and blue shift show up in spectrographic analysis and give information on velocity in the clouds, and then fun stuff with TreeSPH that was the highlight of my day.
Skipping all the stuff on F=ma and the scariest things in my life at the moment – derivatives, Sarah explained how a model could be assembled in the following way: All particles are tagged as either “Gas” or “Dust”. There are a number of Forces, F that each can have – Fgrav (gravity), Fpres (pressure), Fvisc (viscosity), Fdrag (drag) that make you move and interact with other particles. When one particle meets another there are some possibilities about their interaction based on what type of particle they are and what forces are available to them because of their type: Gas-Gas has gravity, pressure and viscosity, Gas-Dust has gravity and drag, and Dust-Dust has gravity only.
Then particles can be queried for their age, and if age=n, and they happen to be in a dense region, they form a star. Queried again later their age might be age=n+old so they go supernova and cause feedback and metal enrichment, and so on.
The instantly entrancing thing about this was the idea of tagging an object, the tag having certain attributes and certain ways of interacting with other objects. Call the objects dancers, have specific operations or methods of moving available to them depending on their type (which exists independently of whatever pre-existing choreography they are doing, though can affect and modify this), and then a subset of those methods available to them depending on the type of the dancers they get too close to. Then also proximity indicates manner of interaction, too far away and it’s just an arc past, close enough and they fall into orbit, too close and they collide.
In all this, it’s the application of really simple rules and instructions that make the complexity, both for astrophysics and for what I’m trying to do. The choreography we are doing though, is fiendishly intricate and mentally taxing, so there needs to be clearer methods I think for allowing it to open up to this kind of generative approach.
We were talking about Leibniz again yesterday, and Gille Deleuze’s essay on him, The Fold: Leibniz and the Baroque, which is probably where all this started for me. I guess the next thing amidst all this is how to take what we already have that is choreography and work out how to fold, unfold and refold it, so that from any one point an infinity of possibilities can result.
That’s something for tomorrow, I think.