FYI, I got the article from Chris, N0UK. It is a good article about automated audio signal-to-noise measurement, but not much detail about the optical photon tube I was looking for. I forwarded the article to Rob K0XL and Barry VE4MA who expressed interest. 

For those who want to know, the "photon tube" is basically a 2ft or longer PVC pipe with a LED at one end and an optical receiver at the other end. The pipe is sealed as well as possible to eliminate external light sources and has a lining of non-reflective flat-black material inside. The LED can be driven with any desired test signal and the operating current is set very low for minimum light output. The optical receiver can be swapped out to test different designs and to determine relative sensitivity and signal-to-noise with a "known & repeatable" light source. The intent is to create something that looks like a point source and eliminate off-axis reflections with the flat-black lining. One variation was to add a pinhole shield a couple inches from the LED and another shield some distance away from the optical detector with a hole diameter near the detector's size. Another variation created a folded or bent tube to reduce the size, but you also need to keep the active LED circuit away from the optical detector to prevent measurement error. A white piece of paper could be a corner reflector or an attenuator.