Suposed vibratory cycle (one open and one closure) of a normal adult male: 100Hz Video discrediting the statement: https://youtu.be/v9Wdf-RwLcs
You cannot detect the dynamics of vocal fold vibration using standard 30 frame per second video, since they vibrate (in that individual) a dozen times per frame. This is actually a nice demonstration of the logic behind the Nyquist rate and the need to filter signals for digitization. What happens is that a camera can detect a change in glottal configuration at most once every 1/30 of a second, and in between frames, you can't tell if the vocal folds have opened and closed one entire cycle, or two cycles, three cycles, and so on. In the case of sound, you have to remove any sound component that is above 1/2 of the sampling rate; if you don't, then components will "wrap around". With a SR of 10K, a component at 9.5K is mathematically the same as a 500Hz component. With video, there's no way to filter out motion, so you have to guarantee that your frame rate is high enough to detect complete cycles of motion, which is why they use high speed cameras for observing vocal fold dynamics.
As for the question of complete closure, that is indeed a variable and potentially contrastive property of the voice. There is a distinction in some languages (Dinka, Mazatec, Hmong) between modal-voiced and breathy-voiced vowels. In breathy voiced vowels, the vocal folds are closed for a relatively short period of time, compared to modal-voiced vowels. Also, the folds separate and make contact continuously, that is, they don't instantly transition from "complete contact" to "no contact", so there are degrees of openness and closedness. Moreover, we only see 2 dimensions of vocal fold vibration in such films, but in fact the tops of the folds and the bottoms of the folds kind of vibrate independently (like two separate masses with a connecting spring).