**4. Conclusions**

The amount of literature on the effectiveness of noise barriers has not provided sufficient evidence on the actual attenuation achieved by these devices, and there is uncertainty over the noise reduction capabilities of existing barriers. The methods described in the ISO 10847:1997 standard have drawbacks that make it difficult to obtain reliable attenuation measurements.

The direct method ensures identical propagation characteristics since the source of noise, the barrier, and the receiver are at the same positions, but the equivalence of source and meteorological conditions may not be fully satisfied. The indirect method ensures that the same local weather and traffic conditions are maintained, but the equivalence of terrain profiles, obstacles, and ground surface conditions may not be fully achieved. In addition, the usage of provisions of the ISO standard sometimes is complicated when at the site point exists a relatively high background noise level, or adverse meteorological conditions [54].

According to the ISO standard, the recommended method is the direct method, although most studies have been based on the indirect method because the barriers were installed during road construction, and therefore it was not possible to obtain equivalent "before" measurements.

The ISO standard provides generic methods for determining Insertion Loss at receiver locations. However, there are no universally acknowledged receiver positions for measurements. It is important to note that barriers are relatively ineffective at some distance from the road. The effective distance range is limited to a few tens of meters, so it is unclear that many receivers can benefit from barrier attenuation. Many of the studies conducted have calculated Insertion Loss levels at barrier near-field distances, so the noise reduction capabilities of barriers were only partially assessed. The IL levels measured at comparable greater distances from the barrier (20–30 m) were, in most cases, very moderate. This supports the argument that the barrier attenuation levels are, above a certain distance, clearly lower than expected.

Additionally, the ISO standard specifies measurements of equivalent continuous A-weighted sound pressure levels to calculate the attenuation of the barrier. However, A-weighting tends to underestimate the effects of low-frequency noise [47]. Several studies have highlighted that A-weighting does not adequately consider the perceived annoyance produced by predominantly low-frequency noise. This is the case of road traffic noise, which is characterized by the wide variability in the relative level of lowfrequency noise [25, 55]. Noise barriers increase the relative level of low-frequency of noise on the shielded side of the barrier. Thus, the attenuation in A-weighted measured levels level may overestimate the estimated reduction in perceived annoyance due to the increase in the relative level of low-frequency sound [47].

In summary, the literature has described some critical points about the applicability and reliability of ISO methods. These points were dealing with (i) the reliability of results (i.e., direct IL measurements obtained at different moments, and indirect IL measurements obtained at equivalent locations), (ii) the equivalence of results of direct and indirect methods, (iii) the nature of the indicator used (A-weighted levels), and (vi) the relevance of operational factors such as weather conditions, traffic fluctuations, ground impedance, and background noise.

The effect of these factors on noise levels measurements needs to be better known. More research studies in this domain are required to bring improvements in measurement methods.
