Construction and building materials are generally considered Lambertian, namely they reflect as perfectly diffusive the radiation incident onto their surface. Thus, reflectance and absorptance are assumed to be constant and independent on the incidence and view angles. This assumption is not valid for specular materials, like glass or polished surfaces; where an angular dependence of their optical-radiative properties is observed. However, many opaque construction materials often show a mixed behaviour, which includes specular and diffuse (or scattering) reflectance components. Moreover, the apparent roughness of the materials surface changes according to the angle of incidence of the solar radiation. This issue is relevant for some cool materials, which are polished or treated with other methods to offer a very smooth surface, to increase the solar reflectance. Herein, the angular dependent solar reflectance of the opaque materials and the impact on the solar gains and energy balance of the building envelope is investigated. The off-normal solar spectral reflectance of four typical construction material for roofing and facade systems is measured with a spectrophotometer, equipped with a large integrating sphere, and with a goniophotometer to characterise the directional reflectance. The comparison between reflectance values in the visible band obtained with both devices shows relative variations lower than 2% for materials with prevalent diffusing behaviour. Discrepancies up to 6% are observed for the samples with a perceivable specular component. Solar reflectance angular dependent curves are calculated, starting from measurements, to compare different models to compute the solar gains. A building energy simulation tools is then used to compute the differences between heat fluxes and energy needs obtained with reflectance angular dependent models and with constant solar reflectance. Discrepancies of the thermal fluxes up 0.7 kW h/m(2) in summer and 0.5 kW h/m(2) in winter are calculated, with relative variations exceeding 7%. Discrepancies of 1.7 kW h/m(2) in winter and 1.2 kW h/m(2) in summer are calculated for the heating and cooling energy uses of a reference building. (C) 2015 Elsevier Ltd. All rights reserved.