As mobile devices with microphones (e.g. mobile phones, laptops) become increasingly common, it is possible to create an ad-hoc wireless network of such devices to make audio conferencing more mobile and practical. Such a network could then be used for efficient speech capture and source localization. However, the ad-hoc nature of the network introduces several errors, e.g., inaccurate clock synchronization, errors in estimating the sensor location, and constraints on the amount of bits available for wireless data transmission. In this paper, a framework is established to analyze the impact of such errors on source localization in the audio conferencing application.