Dual-axis Sound Pendulum

v1.0

In version 1.0 (January, 2019) I used an early prototype from Mictic to measure the angle of the pendulum. The resulting data was outputted as a stream of Midi CC values over Bluetooth and read by Max for processing and sonification. Using the anglular displacement around the X- and Y-axes, I computed a projection of the position of the pendulum onto the X-Y plane.

Using this result as an input, I used the nodes object max to compute the distance to eight points aranged in a circle around the space. This object outputs a list of 1 - the distance to the center of each node when a pointer is within the radius, and 0 otherwise. The output was used to control the amplitude of eight harmonically related (from a scale) oscillators. A ninth center node applied a distortion effect to the output signal - the maximum amount of distortion could be manually adjusted over time (via the size of the node center). In order to account for the decreasing magnitude of deviation of the pendulum over time, I manually adjusted the position and size of the nodes in real time as the process progressed. Finally, the output of each of the individual 8 nodes was played from a corresponding speaker arranged in a circle around the audience, while the summed output was played from the swinging speaker.

v2.0

In version 2.0 (June, 2019) I replaced the sensing system with and Arduino Nano coupled with a 9 DoF IMU that send out a wealth of orientation data in floating point format via OSC over wifi. This implementation offered a substantial improvement upon the 7 bit resolution of the Mictic (particularly at the end of the process when the total deviation was relatively small).

In this version I updated the synthesis method to include to include 2 elements:

1. 1. The swinging speaker played the output of a series of several detuned oscillators (programmed in supercollider, using LFNoise to ___). Specific parameters of the oscillators (e.g. __, __, __) were controlled by data (derived) from the pendulum.

   2. The 24 speakers in the room were again mapped to independent audio channels, this time the result of a composed granular synthesis process. As in the previous version, each channel was ‘activated’ when the pendulum entered the node. The size of the nodes were inversely proportional to the maximum deviation from the Z-axis over the previous 15 seconds, so as the magnitude of the motion of the pendulum decreased over time, the size of the nodes increased and the granulation ceased to be localized to the area near the moving speaker and tended to fill the entire room.