A module for the After Dark screensaver from Berkeley Systems. Written in 1994 by Ping-Wen Ong.

SAD Graph stands for Sound-aware After Dark. When activated, the module listens to whatever sound input your computer has set: either the CD audio, or the mic inputs, or standard RCA input(s), or a built-in microphone if one exists. There are controls for sensitivity and sample rate as well as color and float options for the display.

Output from the module is in the form of a waterfall graph, familiar to acoustic engineers. Others may remember the cover of Joy Division's Unknown Pleasures. That was actually a nuclear magnetic resonance display, but the offset lines look similar.

A line is drawn at the bottom of the screen for each sample, with the x-axis representing frequency from 20-20khz, and the y-axis showing amplitude at each frequency. For each successive sample the previous line is moved upscreen by a small offset and another line is drawn representing the current sample. Each line is thus an instantaneous picture of the waveform, and the full display shows a pseudo-3D history of your computer's acoustic environment over a period of about ten minutes.

What is this good for? Well, like most After Dark modules, not much. It's eye candy. However, I put it to use as a workspace security monitor. When I'd go for a smoke I'd activate After Dark (with mouse movement ignored and a password lock in place). Sensitivity of the Apple PlainTalk microphone is easily good enough to pick up doors or cabinets opening or closing, keyboard clacking etc. Not only does any such noise stand out dramatically from the noise floor on the SAD Graph display, but I learned to recognize some sounds' particular signature from the waterfall. When I came back to my desk I could see whether anyone had been at my other computers, rooting through drawers, or whether the phone had rung.

SAD Graph was updated once by the author to give it more display options, but unfortunately its principal drawback was not addressed. Because the sounds that people are interested in occur mainly below 12khz, the right third of the graph rarely showed anything but noise, unless you were playing with a dog whistle. Restricting the range to 50-12khz or so would map the display better to the inputs and allow better discrimination.

Still, a bit of fun with synaesthesia.