Realistic Graphics Lab


RGL re­lies on a cus­tom­ized PAB pgII go­nio-pho­to­met­er for most ma­ter­i­al meas­ure­ments. This device sim­ul­tan­eously il­lu­min­ates and ob­serves a ver­tic­ally moun­ted sample from ar­bit­rary dir­ec­tions us­ing a set of four mo­tor­ized axes (double ar­rows). Il­lu­min­a­tion ar­rives from a stat­ic light source in­stalled in an ad­ja­cent room par­ti­tion (ar­row from the left); the il­lu­min­a­tion dir­ec­tion is thus con­trolled by ro­tat­ing the sample mount. The walls of the room are covered with a black fab­ric to ab­sorb light re­flec­ted from the sample.

The pgII is fast: in the de­fault meas­ure­ment mode, the sensor head re­volves around the sample at about 3 meters/second, while ac­quir­ing meas­ure­ments at a fre­quency of 1Khz with an an­gu­lar pre­ci­sion of 1 mrad (≈ 0,06°). The res­ult­ing data fully char­ac­ter­izes the op­tic­al be­ha­vi­or of the meas­ured ma­ter­i­al and can be used either as a ground truth for ex­ist­ing mod­els, or to re­pro­duce the ma­ter­i­al in a phys­ic­ally-based ren­der­er. 


The device provides the fol­low­ing ima­ging mod­al­it­ies:

The first two sensors would nor­mally be used for BRDF/BSDF meas­ure­ments, while the cam­er­as en­able meas­ure­ment of spa­tial vari­ation in ad­di­tion to dir­ec­tion­al be­ha­vi­or. Note that the cam­era and spec­tro­met­er have an in­trins­ic ex­pos­ure time and can­not meas­ure while the ma­chine is mov­ing, hence a slower pro­ced­ure is used where the device drives to spe­cif­ic con­fig­ur­a­tions and halts un­til the ex­pos­ure is done.




Light sources

Vari­ous light sources can be moun­ted on two light rails in a sep­ar­ate room par­ti­tion. They il­lu­min­ate the sample through two small cir­cu­lar holes vis­ible above. The fol­low­ing op­tions are cur­rently provided:

The lamp rails can also ac­co­mod­ate beam con­di­tion­ing op­tics and oth­er op­tic­al devices, in­clud­ing fil­ter wheels (e.g. to meas­ure fluor­es­cent ma­ter­i­als) and mo­tor­ized ro­ta­tion stages (e.g. for po­lar­ized light meas­ure­ments). The im­age be­low shows a setup where the left rail is used to meas­ure retrore­flec­tion us­ing a co­her­ent laser source that passes through a beam­s­plit­ter.

Sample mounts

Two dif­fer­ent sample mounts can be in­stalled:




Ideally, the sample sur­face to be meas­ured is per­fectly flat and cov­ers at least 10×10cm - 20×20cm. Lar­ger samples are help­ful be­cause the beam used to il­lu­min­ate the sample stretches out in­to a long el­lipse at graz­ing angles (> 85°)—if the sample is too small, parts of the sample hold­er would be meas­ured in such con­fig­ur­a­tions. The im­age be­low shows a meas­ure­ment of an iri­des­cent but­ter­fly (Morpho melenaus) whose sur­face is not per­fectly flat. In such cases, the res­ult­ing meas­ure­ments will spe­cify an “ef­fect­ive BRDF” of the il­lu­min­ated re­gion.

Ex­ample data

Click the screen­shot on the bot­tom right to see an ex­ample of the typ­ic­al out­put pro­duced by the device—the file will open an in­ter­act­ive view­ing tool. This is a high-res­ol­u­tion meas­ure­ment of a holo­graph­ic pa­per meas­ured us­ing a sil­ic­on pho­to­di­ode. Each point cor­res­ponds to a dir­ec­tion pro­jec­ted from the hemi­sphere onto the disk, whose height is pro­por­tion­al to the log­ar­ithm of the in­tens­ity. The purple ar­row de­notes the dir­ec­tion (θᵢ=30, ϕᵢ=0) from which the ma­ter­i­al was il­lu­min­ated. The bump on the op­pos­ite side is the spec­u­lar re­flec­tion, and the nu­mer­ous ridges are caused by dif­frac­tion from the pa­per's grat­ing-like mi­cro­struc­ture.