I believe the main reason Peter Wright wanted unconventional
snooping devices was to avoid detection by sweepers who
regularly checked for the usual, known devices.
Intercepting signals from radiated objects was one of those
methods for it avoided having to plant a device within the
targeted space. However, such methods eventually became
known to sweepers and countermeasures were taken
though Wright does not describe them.
Wright wrote of one instance when Soviet sweepers of
their Egyptian ally's code room discovered a British mike
concealed in a phone set. The Brits could hear the
sweepers unscrewing the cover, pausing, then screwing
the cover back on without removing the disc mike. Wright
presumes the Soviets wanted the Brits to hear about
the strong opposition the Soviets were planning to
British Middle East ventures.
Which raises the question of what devices/methods
Wright and others used to supplement the ones that
have been written about. Seldom was only one
method applied.
On Steve Bellovin's point to "look for the plaintext,"
when examing encrypted signal: The NONSTOP
document lists that threat as a required testing procedure
of encrypted signals, as well as for emissions from
nearby cabling and equipment that may not themselves
be processing encrypted or RED signals.
RED/BLACK requirements for separating classified
intelligence bearing data from those carrying unclassified
data repeatedly emphasize the threat of seemingly
innocuous, unintentional antennas snarfing both
plain text and cipher text. Those requirements are
described in "NSTISSAM TEMPEST/2-95 RED/BLACK
Installation Guide":
http://cryptome.org/tempest-2-95.htm
It is this doc which states that NONSTOP is the principle
TEMPEST vulnerability of "transportables, aircraft and ships."
It may be that the many surfaces of these objects is what
makes such good emitters of unintentional signals, especially
when radiated. The changing environment of each probably
also complicates countermeasures.
What I've not seen are reports on the emissions of individual
warfighters who may be heavily clothed with electronic and
antenna-like devices. Unless the term "transportables" covers
such mobile, multi-faceted objects.
Then, there are the plain voice signals emitted by a vehicle in
which an encrypted cellphone user is riding. Say, Potus or
Russian bear. And so on.
The full electromagnetic spectrum is presumably capable of
transmitting more covertly than overtly -- that is, there are far
more signals being sent than there are devices to receive
and process them. So we are disinformed.
