The Internet has evolved to a stage where the number of Internet enabled devices exceeds the global population. With much of the required connectivity over wireless, contention for bandwidth is an important issue that has to be addressed. In this context, applications that have certain quality of service (QoS) requirements must be protected. WiFi-enabled mobile devices such as smartphones and tablets are commonly used by both personal and business users for voice over IP (VoIP) communications, while also supporting conventional data applications such as email, file transfer and web access. Previous research has shown that time synchronized endpoints can provide better QoS by calculating accurate mouth-to-ear delays, and using this information to better inform buffer playout strategies. The IEEE 802.11e protocol extends 802.11 by providing different traffic priorities based on traffic type. However, it cannot distinguish between traffic streams within the same category. In this paper, we present an access point-centred mechanism for further distinguishing between streams within 802.11e categories. The mechanism predicts the VoIP quality for multiple sessions using the ITU-T E-Model. Delay values are determined from RTCP packet timing information. We provide implementation details on a real world proof-of-concept and present results that correlate well with previous NS-3 based simulations. In addition, we address scalability issues for large scale deployments. Copyright (C) 2015 John Wiley & Sons, Ltd.