Freezing of gait (FOG) is difficult to measure due to its unpredictable occurrence. This study investigated: (1) whether the new freezing of gait questionnaire (NFOG-Q) is a reliable measure of freezing by comparing patients' ratings with those of carers' and (2) whether adding a video improved its reliability. Non-demented people with Parkinson's disease (PD) (N=102) and their carers of similar age and cognitive status were recruited from movement disorders clinics in three countries. The NFOG-Q was administered to carers and patients independently before and after watching a video showing several examples of FOG. Patients had very high agreement between their pre- and post-video detection of FOG (Kappa=0.91). However, this was less than in carers (Kappa=0.79). The video had a significant influence (p=0.01) on the rating of FOG severity (duration) but not on the estimation of its functional impact. Post-video freezing severity scores in the 69 freezers showed high agreement with carers' scores (ICC=0.78 [0.65;0.87]). We conclude that the NFOG-Q is a reliable tool to detect and evaluate the impact and severity of FOG. Adding a video does not add to the sensitivity and specificity of FOG detection but influences the estimation of FOG severity.