Take at the video again at around the 3:45 mark and you will see that as each truck passes the rail join, that the rail come up about 6 or 7" or maybe more and you can observe the 2 sections of rail are bolted together and are placing the joint under extreme distress and I'd guess that is why the train crosses the road so slowly.
That amount of movement suggests to me that the rail is not actually attached to the sleepers at various parts of its length. Properly attached rails will spread the weight along the length of the rail and resist the flexing as it is essentially a massively strong I beam.
The play present in the join between the two rails indicates that the fish plate bolts aren't properly tightened either. Since the train didn't derail, I think it's safe to say that the ties/sleepers are (somewhat) attached to the rails but a thorough tamper is in dire need because the ties are flapping about in the ballast.
Edit: I found an
instruction that says: "If using an impact or torque wrench then the torque setting should be set to 350 lbft
(475Nm) for black fishbolts and 650 lbft (880Nm) for 1” high tensile bolts." I'm pretty certain there are no 475Nm on those fishbolts in the video.
Regarding vertical motion it is a of course a definite difference between the small but noticeable flexing a passing train will impose on well maintained mainline track and the dance we see here. It is not debated whether track will flex; it will, but there are limits for when it's OK and not.
Having traveled at 320km/h (200mph for you older people) on a regular service, am happy that those rails (LGV Est Strasbourg-Paris, site of the world speed record on revenue track, 574,8 km/h, in 2007) are better cared for.