Audio analysis turns pressure waves into answers: how loud, where in the spectrum, did something just hit, what tempo. Every audio-reactive system is an analysis layer feeding a rendering layer, and the analysis quality sets the ceiling on how musical the result can feel.
The standard toolkit stacks up from cheap to costly: level and envelope following, band splitting, onset detection, beat tracking, then pitch and harmony features. Most working systems stop in the middle; events and bands carry surprisingly complete musical information.
Photism’s analysis runs eight bands with transient detection inside the device, so the renderer receives music already parsed into the vocabulary producers think in.