Years ago, I was faculty advisor of a thesis containing a viable technique
for measuring the carrier of a SSB signal modulated by a human voice when
the carrier was nulled. The student's name, as well as I can remember, was
Day and his first name was something like Louis.
Should be a link to it somewhere.
The scheme took advantage of the nature of the human voice. It required one
to have the means to inject a carrier (which was measured by conventional
means) and to look at the recovered audio on a scope.
The old FMT had one measure CW signals. They were great fun.