Most media professionals have the desire to express quality in a way that is commonly understood and that facilitates comparisons across different algorithms, different organizations, and time.  Mean Opinion Score (MOS) has emerged as the most popular descriptor of perceived media quality.  The 5-point MOS scale (excellent, good, fair, poor, bad) in particular is widely used.

MOS has had great success in the domain of speech quality, and consequently it has also been used for other modalities such as audio, video, and audio-visual content, as well as numerous applications, from lab testing to in-service monitoring.  MOS is used not only to express the results of subjective tests (“subjective MOS”), but also as the output of objective measurement algorithms that provide an automated alternative to subjective tests (“objective” or “predicted” MOS).

While there is a clear benefit to such a “reference quality indicator” and its widespread acceptance, MOS is often applied without sufficient consideration of its scope or limitations.  In this talk, I will critically examine MOS and the various ways it is being used today.  I will highlight common issues with both subjective and objective MOS and discuss a variety of alternative approaches that have been proposed for media quality measurement.