What is MOS?

Detailed explanation

MOS originated as a subjective quality metric where panels of listeners rated voice samples on a 1 to 5 scale (ITU-T P.800). Today, objective algorithms like POLQA (Perceptual Objective Listening Quality Analysis, ITU-T P.863) and PESQ (Perceptual Evaluation of Speech Quality, ITU-T P.862) automate this process by comparing the received audio against the original reference signal.

In mobile networks, MOS is the definitive measure of VoLTE and VoNR call quality as experienced by the end user. It captures the combined effect of all impairments: codec compression, packet loss, jitter, delay, echo and background noise. A single MOS value summarizes the complex interplay of these factors into an intuitive quality rating.

The MOS scale ranges from 1.0 (unintelligible) to 5.0 (perfect quality, rarely achieved in practice). Wideband codecs like AMR-WB typically achieve MOS scores of 4.0 to 4.5 under good radio conditions. The EVS codec can reach even higher scores thanks to its super-wideband and fullband modes.

MOS monitoring during field testing reveals geographic areas where voice quality suffers. By correlating MOS drops with radio KPIs (SINR, RSRP, BLER), engineers can identify whether quality issues stem from coverage gaps, interference, handover failures or network congestion, enabling targeted optimization.