On the effect of compression on the complexity characteristics of wireless acoustic sensor network signals

A microelectronic system for monitoring areas of environmental interest through the automated creation of soundmaps, based on data from a wireless acoustic sensor network (WASN) has been recently proposed. In this context, it has been demonstrated that compression algorithms need to be employed at sensor node level due to the increasing demand in bandwidth as the number of sensors and events to be logged increases. Motivated by this finding, the effect of data compression on signal complexity is studied in this paper by employing four widely used audio compression algorithms in combination to different entropic/information measures. Several entropic/information measures are calculated for both compressed data streams and the original audio, leading to a comparison on the effect of the compression on the complexity characteristics of WASN signals. Numerical results imply that in a realistic WASN for environmental monitoring it is possible to locally compress audio data at node level prior to network transmission while maintaining the complexity characteristics of the sound signal in terms of preserving the precision of specific entropic/information metrics. However, this is not possible for all the studied “complexity metric-compression algorithm” combinations.