The measurement space envelope for combined BTDF and BRDF measurements, shown on Figure 4(a), consists of a carbon fiber cap strengthened by a structural metallic frame; this frame also supports a static stainless-steel perforated sheet on which a moving synthetic strip can glide. The role of the synthetic strip is to select the elliptic hole through which the incident light’s path will be adequately controlled (according to altitude 61); at the same time, it prevents light from entering the measurement space through any other opening. Its unique aperture is therefore circular, slightly larger than the largest ellipse (i. e. the one associated to normal incidence); the chosen 10° step in altitude ensures that a 15 cm diameter hole never overlaps two consecutive entrances.

The determination of the actual position and dimensions of the ellipses cut out from the metal sheet required a multiple stages process for an optimal incident light control:

• First, the theoretical geometric properties of the ellipses were determined based on trigonometric considerations, assuming a perfectly parallel beam reaching an elliptic surface of apparent horizontal axis 15 cm and vertical axis 15- cos61.

• Then, the ellipses dimensions were adjusted to the real incident beam, of imperfect collimation and thus producing blurred regions around the uniformly illuminated area, responsible for parasitic reflections. Once the optimal source distance was determined, different elliptic shapes were tried out to compare the achieved sample surface illumi­nation. The most efficient compromise was established between optimal uniformity over the whole sample area and lower parasitic light flux; this was done for each ellipse individually, as more relative blurredness appeared for smaller ellipses. The determined shapes, cut out of cardboard sheets, were tested successfully; they led to only few percent of non-uniformly illuminated sample area while guaranteeing an
average relative blurredness area lower than 10%. It can be noted that these remain­ing parasitic reflections were reduced to a negligible level by adding a ring of highly absorbing material (“velvetine”) around the sample.

• Finally, the positions of the ellipses on the metal sheet had to account for the frame manufacturing imperfections (see above). The metal sheet was thus mounted tem­porarily on the frame, allowing to centre the ellipses thanks to a plumbline course driven by a progressive platform inclination. Their positioning was thereafter verified by pointing a fixed laser on the central axis and tilting the device to get each ellipse’s centre coincident with the laser spot; this test showed that an appropriate accuracy was achieved (± 0.05 cm deviation). Before sending the metal sheet for cutting out, these positions were adjusted to a flat configuration of the sheet (i. e. to its neutral fiber), to avoid slight shifts due to the sheet’s thickness.

The resulting perforated metal sheet is shown on Figure 4(b); its inside surface is covered with “velvetine” (reflection factor lower than 1%).