If you recorded yourself sitting at the computer right now, it would probably be pretty noisy. The noise in this vid could be the air conditioning or perhaps the lab’s ventilation hood, sound from outside via an open window, or even the self-noise of the microphone/amplifier. This recording likely has one microphone sitting in the middle of the room.
You can hear that the experiment’s conductor is a fair distance away from the source, and that the glass sounds are very clear. For the speech to be audible, the mic’s preamplifier would have to be turned up quite a ways. It thus amplifies quieter background sounds in the room (A/C, wind, a fan) by just as much as the speech, making them, obviously, quite loud. It actually sounds as if the audio has been amplified digitally (i.e. after the original recording), which can be an incredibly noisy process.
Silence is rather subjective. There’s plenty of noise in a ‘quiet’ room that we just ignore. The only rooms that are truly quiet require a lot of expensive treatment, especially to the AC, because most of this treatment traps heat. Oftentimes in audio engineering you’d just use a noise gate, which measures the minimum level you want to hear (in this case, the conductor’s speech), and shuts off the rest (hence the word ‘gate’).