Block 3. Subjective Testing

Subjective testing is the ground truth in video quality assessment — what a real test is, why metrics bow to it, and why "I showed the team" is not one.

MOS and DMOS explained: the ACR 5-point scale, absolute vs degradation rating, the DMOS formula, and turning raw ratings into a defensible score with a CI.

ACR, DCR, CCR and pair comparison explained: how each ITU-T P.910 / BT.500 subjective test method works, what it detects, and how to pick the right one.

How to design a subjective video quality test that survives scrutiny: source set, impairment matrix, viewing conditions, randomization, and ITU-T P.910.

How to run a subjective video quality test: recruit valid observers, screen acuity and colour vision, set the environment, and reject unreliable viewers.

Turn subjective video votes into a defensible result: MOS confidence intervals, subject rejection, significance tests, and how many subjects you need.

Crowdsourced video quality testing trades lab control for speed and scale. The ITU-T P.808 framework, gold and trapping checks, and when to trust it.

Eight common subjective video quality testing mistakes — too few subjects, no anchoring, biased content — each with the fix and the ITU standard it breaks.