Objective quality is separately measured for each media source (audio, video, screen-sharing, etc.). It takes two annoyances into account, i.e. variation and lack of timeliness.
Typically, real-time communication has strict timeline for playback and rendering for the communication to be interactive. ITU-T recommends the maximum threshold for end-to-end (capture to render) latency as 150ms. Similarly, 300ms end-to-end latency is imposed on video. Furthermore, most video apps consider changing frame size at short intervals (less than every few seconds) affects the quality of experience significantly.
Ergo, objective quality takes all the above into account and penalises streams that have the following features (non-exhaustive list):
- A/V out of sync and
- Variation in A/V sync at short intervals
- Variation in frame size at short intervals
- Variation in delay
- Throughput, Jitter, Audio output, etc.
The objective quality returns a value between 0 and 4dB, where below 1 is considered poor quality and above 2.0 is considered excellent quality.