I spent a couple weeks at the beginning of December working on the Magnetic Ink pieces. I played around with more elaborate source graphics taken from NASA images of nebulae. I originally tried using smoke as the source graphics but was dissatisfied by the lack of contrast within the images. Nebula images work well because they contain a lot of subtlety but still maintain a general look of fluid or smoke.
I played a bit more with refining the look of the ribbons. They are so incredibly hard to control sometimes. The current code allows for particles to pass back and forth between the large gravity orbs. Occasionally a particle ends up in a loop where it will bounce from one gravity to another every frame. This happens mostly because I am not using Verlet Integration or the much feared 4th order Runge_Kutta to calculate my particle movement. I stick with the much simpler Euler method but it can cause problems. The effect on the ribbon is that it forms a sawtooth pattern which takes away from the smooth organic look I am trying to achieve. However, these problems are small enough that I am willing to overlook them. Learning the Verlet or RK4 is quite the mental chore. I am putting it off for now.
A few of the newer pieces allow the ribbons to reflect the history of the audio being analyzed. Not sure if I like it more or less than the more hair-like ribbons, but I do like being able to see a bit more into the process. The pixel gap between the quads is a whole other story. Not sure how to get rid of those yet.