Median noise barriers, like parallel noise barriers, can be employed to reduce the impact of traffic on roadside communities via the direct propagation path. The performance of different shapes of median barriers was compared using reactive and passive surfaces and a 2D boundary element method (BEM). In the case of reactive surfaces, quadratic residue diffusers (QRDs) and primitive root diffusers (PRDs) were used on the top and stem surfaces of median barriers. To introduce passive barriers, two different absorbent materials including fibrous material and a grass surface with flow resistivity of 20 000 and 2500 kg/(s·m2), respectively, were similarly applied. The effect of thin absorptive barriers was similar at lower frequencies and better at mid and high frequencies to that of their equivalent rigid barriers. More improvement was achieved by covering the top surface of thick barriers with grass rather than with fibrous material. The performance of QRD and PRD barriers where the diffuser was located on the top surface was more frequency dependent than that of barriers coated with fibrous material. A comparison of the average A-weighted insertion loss in the thick barriers showed that the greatest improvement (2.59 dB (A)) was achieved using a barrier of 30-cm thickness covered with grass. Median noise barriers, like parallel noise barriers, can be employed to reduce the impact of traffic on roadside communities via the direct propagation path. The performance of different shapes of median barriers was than using reactive and passive surfaces and a 2D boundary element method (BEM In the case of reactive surfaces, quadratic residue diffusers (QRDs) and primitive root diffusers (PRDs) were used on the top and stem surfaces of median barriers. To introduce passive barriers, two different absorbent materials including a fibrous material and a grass surface with flow resistivity of 20,000 and 2500 kg / (s · m2), respectively, were similarly applied. The effect of thin absorptive barriers was similar at lower frequencies and better at mid and high frequencies to that of their equivalent rigid barriers. was achieved by covering the top surface of thick barriers with grass rather than with fibrous material. The performance of QRD and PRD barriers where the diff user was located on the top surface was more frequency dependent than that of barriers coated with fibrous material. A comparison of the average A-weighted insertion loss in the thick barriers showed that the greatest improvement (2.59 dB (A)) was achieved using a barrier of 30-cm thickness covered with grass.