QoE and MOS
Traditional QoS measures are important, but only a piece of the puzzle when assessing real-time service quality. These measures do not enable effective management of voice and video quality to contracted service levels or identification of issues leading to poor quality as perceived by the user.
Quality issues can occur as a result of a single technical fault or a number of impairments which combine to degrade the overall customer experience. Common to all communication technologies are impairments such as background noise, echo, delay, clipping, coding errors and mismatches in volume levels. In VoIP, IP Telephony and Unified Communications there are also unique sources of degradation including codec compression, packet loss, jitter, and latency while cell phone calls may additionally be degraded by bit errors, frame erasures and various compression schemes. So QoE requires monitoring and assurance of more than network parameters for QoS. The Schematic below shows how we need to build up a complete picture of voice and video QoE:
Consider an international long distance call from a cell phone through the PSTN to an enterprise VoIP network and terminal. All the network impairments described in this schematic are possible. A true voice quality measurement would consider the extent of all these factors, whether they occur in the network or outside it and determine the overall impact, in the opinion of the customer, of quality. This measurement is known as a Mean Opinion Score (MOS).
Many vendors measure MOS, but not all MOS measurements are the same. Most solutions provide an e-model MOS. The e-model standard is designed as a network planning tool. The e-model standard clearly states in its definition that it is unsuitable for measurements on per user, per session basis – in other words real-time measurement. That is why Psytechnics designed its MOS score specifically for real-time measurement. It is the only MOS designed for real-time measurement. If you don’t have a Psytechnics MOS measurement you don’t have a real-time QoE assessment of user quality.