Modeling Perceived Audio for Cellular Phones Based on Field Data

Even though, the current cellular network provides the user with a wide array of
services, for a typical user voice communication is still the primary usage. It has become
increasingly important for a cellular network provider to provide the customers with the
clearest possible end-to-end speech during a call. However, this perceptually motivated
QoS is hard to measure. While the main goal of this research has been on the modeling of
the perceptual audio quality, this thesis focuses on the discovery of procedures for
collecting audio and diagnostic data, the evaluation of the captured audio, and the
mapping and visualization of the diagnostic and audio related data. The correct
application of these modified procedures should increase the productivity of the drive test
team as well as provides a platform for the accurate assessment of the data collected.