So where do SR80's "peels" blinding artefacts come from? (I was suprized how huge and bright they really are)
Probably they are exclusively fabricated by
one LED shining onto the other's "designated" reflector part.
How to test it: Method 1: The black wall could be used. Method 2: A laser pointer (low power for safety) can be used to shine from where the artefact appear - to figure out from which LED it comes from, "this", or "the other".
(Someone wrote to me these two ideas few weeks ago. I partially copy-pasted the wording here. Shame on me not getting it right back then)
Method 3: One can inspect the reflection of an LED when lamp is off (yellow), and its surrounding PCB area (white) in (each) reflector stripe. The PCB area stops with a reflector on one side of the LED, and continues white on the other. This assymetry can be seen in the mirror stripe (each), and thus one can deduce, which LED ("this", or "the other") you see at a particular angle. The reflection of the LED and PCB area is distorted and unclear, and it takes some time, but still it's possible to say with certainty which LED it is.
Briefly, I can confirm the above assumption of the cause of the artefacts.
The number of artefact peels equals to the number of reflector stripes (not surprisingly):
(here the right LED, which on first photo is on left, shines on left reflector half, which on photo is on right, and creates artefacts on the left side of the beamshot)
I'd say the black wall should not be full-height, it should be more like a fence. Its upper shape should be special, to only prevent "the other" LED from reaching "this" half's stripes. The artifacts will disappear, but I think there will be some change in the beam pattern (like less bright at the sides).
Bottomline:
1 (shining) LED per reflector please.
----------
Why the SR80 has a dark stripe "artefact" in the near field? The reflector only creates its light carpet about 2m away. The near field (before the carpet) is only created by direct (not reflected) light from LEDs. So there are
2 (obvious) reasons for that dark stripe: (1) the bigger the angle from the LED die normal - the less luminous intensity, and (2) the farther the distance to a road area - the smaller its brightness (luminance).
Why the housing is horizontally stretched? (previously I wished it to be vertically big - for a sharper cutoff). There are (above) the lumen reason, and styling, but also -
to make near field wider. Almost right at the wheel - *and* pretty wide sideways.
(these are some of my mis-understandings I decided to share for everyone)
----------
Why people call the SR80's windshield a lens? Looking thru mine at LEDs and reflector, I'd say it's perfectly flat.
----------
(blabbering below, but hopefully interesting..)
I searched for "reflector design", and found there are many books on this. It's amazing how such a primitive law as geometrical-optics mirror reflection can lead to really involved math and algorithms. To me it looks like there's no way a hobbyist can design a good-cutoff LED reflector, even if using some available (open source) ray-tracing engines (because it's an *inverse* problem), let alone from scratch.
As of now, a search engine's top reflector-CAD company still uses only (!) point-like light sources (with a configurable angular distribution of course) in their commercial system. This brings some doubts in accuracy of reflector designs. Looking at flat (along one direction) stripes of SR80's reflector, may be it was really a rather crude design process (I doubt they solved inverse problem)..
Before I left work, I did tease some friends by saying that I would ride to work, wave to people going into the building and laugh, and then ride home. Haven't tried it yet, though. 
