David Forsyth wrote:
I know this is a wacky idea but...

Would it be at least theoretically possible to transmit digital (hi-fi?)
audio signals using a spark-gap transmitter?

Yes, if you use dash for "1", dot for "0" (or vice versa), _never_
mis-send, and have some unambiguous way to signal EOF, or use
some other coding scheme that unambiguously represents all the
possibilities as stuff that can come out of a spark transmitter. The
EE and mathematicians who know more coding theory than I do will be
better able to answer the question, but _yes_ is the basic answer.

If you want to send it synchronously, then it's even fairly easy,
provided you're prepared to accept a low data rate and do signaling
out-of-band for EOF, End-of-track, and suchlike.

After all, we manage to send images from probes passing Neptune. This
is, by comparison, trivial.

(I'll skip the lecture on Gold codes, sending a 65535-bit PN sequence
straight-up for a "1", inverted for a "0", and all the other stuff
that actually got the bits received with a minimal BER)

Let me know if you want to start sending low-rate digital audio using
your spark-gap transmitter, and I'll move a few thousand miles away.

Related note:

Did anyone else here in the US catch the piece on NPR about Bill
Marconi's first transmission from Poldhu to Newfoundland, about a
week ago? It was sort of spacey and low on information content, but
it _did_ have some stuff in it I hadn't heard before, including audio
of Marconi talking about the problems involved -- including the winds
at Poldhu being so bad that they repeatedly blew the antenna masts
over.

--
Mike Andrews

Tired old sysadmin