Diversity stacking is another technique used to achieve improvement by excluding noise. Records in high-noise areas, such as in cities, often show bursts of large-amplitude noise while other portions of the records are relatively little distorted by noise. Under such circumstances amplitude can be used as a discriminant to determine which portions are to be excluded. This can take the form of merely excluding all data where the amplitude exceeds some threshold or perhaps some form of inverse weighting might be used. Such noise bursts often are randomly located on repeated recordings so that sufficient vertical stacking after the weighting tends to produce records free from the high-amplitude noise.