Also, if you're doing more detailed daylighting simulations using Radiance, and the precision of the angular dependence is important, you may need to consider more complexity of the glass itself. Refer to this post: https://unmethours.com/question/1037/converting-transmittance-to-transmissivity-for-igus/