Crafting an interactive experience can often be an exercise in taming imprecise meatspace inputs like cameras, microphones and finicky hardware. The communities surrounding toolkits like openFrameworks, Processing, Cinder, PureData, etc have contributed a wealth of useful pre-built components for managing this complexity, but it is rare that one finds a component which happens to be an exact match for the problem at hand. The process of customizing something like a camera vision or audio analysis component to your installation’s needs often results in a deep dive into the DSP wing of Wikipedia, where dense phrases like “normalized cross-power spectrum” and “complex exponential multiplied by a Gaussian window” are the name of the game.
This talk will focus on a bag of DSP tricks and techniques relevant for creative coders in the form of case studies and concrete examples. In particular, there will be an emphasis on explaining how existing DSP components can be configured and chained together to form the backbone of powerful interactive installations.