An evaluation of error masking techniques for digital wireless audio systems.

Abstract

This thesis investigates the tradeoffs associated with typical communication system designs for packetized wireless transmission systems. Communication system design requires judicious selection of source data rate, data compression technique, error correction method, modulation scheme, and latency. In particular, we investigate a wireless system for live music, focusing on masking methods to mitigate the affect of dropped data packets. Three different masking methods were developed and tested in combination with three packet sizes, creating nine unique test environments. The efficacy of the masking methods was then evaluated. The packet error rate threshold is defined as the lowest packet error rate for which the subject can hear noise interference. A Matlab graphical user interface was used to automate a human subject protocol, which was designed to investigate the packet error rate threshold for each condition. A comparison of the results and statistical analysis of the effectiveness of the masking methods are presented.