With the complexity of the sky as described in the previous section, one would suppose that the only method for testing models for daylight performance would be to place them under actual sky conditions. In fact, the most accurate real world performance is measured in outdoor conditions under the dynamic and changing character of the fully integrated "sources" of the real sky. However, using this method to answer design questions can be complicated by the fact that the real sky is constantly moving and changing. As a result it is often very valuable to do testing under more controlled sky conditions. While these methods compromise absolute accuracy and reality, they reveal and clarify important issues that can inform design decisions.
For example, if testing outdoors, when comparing design alternatives A and B, how do we know if an improved performance measured in alternative B is a result of a more effective daylighting design or the fact that the sky may have changed in character and/or in illumination by hundreds of footcandles in the time that it took to switch the model to the consideration of alternative B? By testing both alternatives under artificially constant sky conditions we can attribute any measured or observed change in performance to the characteristics of the design itself. Also, the practical constraints of waiting (and hoping) for the occurrence of a specific, desired set of weather conditions, solar geometries and sky types at the testing location support the use of more controlled testing scenarios. Five types of testing are available at BSU / CERES:
Physical Testing
Heliodon: Direct Beam Sunlight/Shadow Patterning
Mirror Box Artificial Sky: Overcast Sky Vault Illumination
Outdoor Real Sky: Integrated Illumination
Digital Testing
Digital Simulation: Integrated Daylight/Electric Illumination
Digital Simulation/Calculation: LEED IEQ Credit 8.1 and 8.2