Personal tools
You are here: Home Science by Ear Workshop Problem sets Ising Model Data 2D Session notes

Session notes

Notes taken during the session.


  • reviewing the dataset
  • locality using Hilbert curves
  • avoid visual thinking -> what would be better auditory thinking
  • whole image into sound object (spins->parameters: data to parameters of sound object
  • do not fix on one objects, be aware of many many objects
  • serialization between many objects
  • serialization is done via time
  • matrix as stationary sound
  • statistical stuff not that interesting
  • matrix as orchestra from above
  • neighborhood should be preserved
  • avoid choices
  • taking some scanning lines
  • straight line through plot
  • we are interested in emergent properties
  • envelope generation via several lines, counting how far apart we go until we have 3, 6, 9, 12 spin flips.
  • what is a cluster: a region of similar spin alignment
  • relation to cellular automata...
  • random sampling and local shape into sound
  • what do we gain if we loose locality: we gain that the system cares less to locality than our eyes do. we loose structure of area around grain. (refocus...)
  • local shape is more or less correlation function
  • lets start with simple approach


We sample 5 positions per frame, creating one grain event for each .

At each position we compute average magnetism over a qxq patch

We perform nonlinear mapping of these values to grain properties, namely

  • value -> pitch
  • amplitude -> deviation from 0.5
  • noise -> if mean around 0.5

In the sonification a sequence of frames is played, from above MC-Temp to below MC-Temp.

Sound Example: around the phase transition the number of noise bursts decreases and the purity and loudness of clean tones increases.


Powered by Plone CMS, the Open Source Content Management System

This site conforms to the following standards: