Philips luminaire in Hostetin, plain and adapted
– an overview of (specific) ground illuminances

Values below the scale are illuminance / 1 lx. This is for (unrealistic) “unit” case that luminaires would have lamps producing 1 klm only, would be point-like and just 1 m over the terrain.

The house-side emissions are never needed and always obtrusive in Hostetin (as they are, if not adequately faint, almost everywhere). So I took a glossy strip of a thin aluminium sheet sized about 7 cm × 20 cm and installed it (laid onto the bottom glass, with corners bent so that it could not shift) just below the lamp. The result was far less light at the house side, and quite a lot of additional light at the street (50 % to 100 % more, esp. at several pole heights from the luminaire), starting by a lane which is some 0.5 pole height far from the pole. The light output ratio went down from 80 per cent to some 66 per cent. But the useful flux increased! (Another mirror lamp positions and strip size and shape might suit situations where the luminaire is not set back far from the street. But the principle is the same: recycle unwanted light partly the desired one, with reduction in total light emissions.)

	– photometry plot from the *.ies file. White dots are spaced each one pole height, axes cross below the luminaire.
And a real imaging photometry from a Fuji S5000 camera (illuminated wall, luminaire 0.50 m from it, background luminance subtracted). To my suprise, the wall albedo of 0.2 and the 4.4klm lamp produced values (luminances / 1 cd/m2) very close to the expected specific illuminances ( / 1 lx), with no need of scaling... The axes and dots (as if each pole height) overlaid on the images should be adapted to perspective, or even better, the images rectified instead. Then the he measured luminances could be compared with the above illuminances in detail. Without rectification, the photographed sideways maxima to the street side are shortened by a perspective. For a reliable rectification, however, a grid on the wall, using real white dots, should be painted (next time...). (Luminances at far right are influenced a lot by tertiary light from another wall, going toward the observer. Without that wall they would continue to diminish away from the luminaire, as in the theoretical plot.)
Original luminaire (mirror extreme street-side, lamp at extreme house-side)	Adapted luminaire, with an added metal strip inside
		short exposures
		long exposures
		differences at short
		... and at long exposures
To make the measurement more understandable, these are real images at mid-exposure:

And this is the view just in background illumination:

(All processed data are within the ../clona directory, the summary photometric values are within the ../clona/infod.txt file. Nothing easy to grasp.)