A retrospective on a popular audio-reactive LED strip project originally built in 2017 by Scott Lawson. The post explores the technical challenges of sound-to-light systems, including DSP techniques borrowed from speech recognition, human perception of sound and light intensity, and why matching all musical genres remains unsolved. The author is considering using a neural network trained on accelerometer data from people listening to music to better capture the human 'foot-tapping' response.
Sort: