Sort:  

The same thing happens with the ships on the water in the last video segment between 3m 57s through 4m 26s. This is the portion covered at 47.9 km away. Note that before the optical zoom of the camera lens is started, you do not see the ships at sea on the water, or the tower. They are all "below the horizon" as you say (because of perspective). Then, as the camera optical zoom is enabled, you suddenly see the ships not seen earlier and then finally the tower. The lens is not powerful enough to get to the bottom of the tower at this distance. That Canon model used claims a 65x optical zoom factor. Now, watch the zoom out process. As the lens backs out (i.e., reducing the optical zoom factor), first the tower disappears completely, and then so do all the ships. If there was ever a "curve", the camera would never be able to see "around the curve." Watch that portion of the video a few times, and then just think about it. There are countless similar demonstrations.

In my last thread, I suggested using a different (more powerful) device. For example, a Nikon Coolpix P900 has an 83x optical zoom, more than that Canon model. At the 47.9km distance, you would see differently.

Another angle. Let's assume the Earth is indeed round, and given its commonly accepted size and radius, the Earth's alleged curve can be calculated from this. There are many other sources. https://dizzib.github.io/earth/curve-calc/?d0=30&h0=10&unit=imperial
The rule - The rate of curvature is about 8 inches x miles squared distance away. So, even at the first measurement of the tower at 25 km (15.5 miles) away, if the Earth actually does curve, you would not be able to see the bottom (8 x 15.5 x15.5 / 12) 160 ft (nearly 49m) of the tower at 25 km away, but clearly, you do see it the full tower to its base.

Do some of your own real experimentation with different optical devices. That is the only way to believe, by doing it yourself. There are lots of great videos online for guidance.

There are also many people doing this with lasers as well, as another way to show there is no curvature.