Performance assessment of the BTDF data compression based on wavelet transforms in daylighting simulation

The bidirectional transmittance distribution function (BTDF), which is used to characterize the light transmission of a complex fenestration system (CFS), commonly involves bulky volume of data that can be a challenge to data storage and transmission in lighting simulation tools and on compact platforms. This paper introduces a compression scheme that is based on planar wavelet transforms with three levels of decomposition to efficiently compress the BTDF data of CFSs and maintain fidelity in daylighting simulation. The root-mean-square error (RMSE) and transmittance error (TE) of compressed medium-resolution BTDF data using three different wavelet bases were evaluated for five different CFS samples at various compression ratio. Based on the results, the generic error from compression did not exceed with CDF9.7 basis for compression ratio up to 200. In the case of rendering an image of a scene in which a CFS was installed at the upper daylighting section of an unilateral façade, errors began to be noticeable at a compression ratio above 70. The uniformity factor was discovered to be relatively sensitive to compression (below 15% error), while the average horizontal illuminance and daylight glare probability (below 10% error) were immune to compression ratio below 100. Compression was also evaluated for a high-resolution BTDF data of external Venetian blinds in the simulation of work-plane illuminance (WPI). With a HDR imaging technique based sky luminance map, the WPI simulation had unnoticeable errors for compressed BTDF data within a compression ratio of 100.