"Richard Fry" wrote in :

In analog TV transmit systems with a typical 500+ foot length
transmission line from the tx to the antenna, a 5% reflection from a
far-end mismatch can be quite visible, showing as a "ghost" image that
is offset from the main image as related to the round-trip propagation
time of the transmission line.

Richard, read this carefully, I do NOT disagree with what you have said
in the paragraph above.

For a ghost to be visible to viewers getting signal off air, there is
more at play that a 5% reflection from a far end mismatch, the reflected
wave is heading towards the transmitter, and needed to encounter another
at least another reflection point with sufficiently high reflection
coefficient that a sufficiently large, time delayed copy of the original
modulated wave reaches the load (even if in turn part is reflected
again).

This second point of reflection could be a transmission line
discontinuity, but it is most likely that it was that the tx end of the
transmission line was not matched, that is that the tranmitter did not
terminate the line with a near perfect match.

It leaves the questions "are transmitters matched to Zo in real life, is
matching an unavoidable consequence of optimising power output?".

For avoidance of doubt, the discussion in this post is about the
transient behaviour of the tranmission line and waves, it is relevant to
long transmission lines with analogue TV modulation, but that doesn't
mean it is necessarily important to other applications.

Owen