"Frank Dresser" wrote in message
news

"Tom Holden" wrote in message
...

[snip]


Moreover, the DRM spectrum is rectangular - even energy distribution. If
we
take our 10 kHz channel and use SSBc to fill it (carrier at channel
edge),
we will get nearly 10 kHz audio bandwidth. If we also use pre-emphasis
(as
is done in FM), or one of the (ancient) Dolby or dBx type noise reduction
systems to lift the treble energy in the sideband, we will get a much
improved S/N. Still very listenable on a conventional DSBAM radio with
typical IF and AF passbands.

They do use premphasis with AM radio transmissions, at least in the US.
The
NRSC has come up with a suggested premphasis scheme:

http://www.nrscstandards.org/Standards/nrsc-1.pdf

This is NRSC 1. As I understand it, NRSC 2 is similiar, but with a a cut
off to the treble boost above some frequency to reduce adjacent channel
interference.

Frank Dresser

Seems like a good starting point - they obviously were seeking a
pre-/de-emphasis curve that would be a workable compromise between the ideal
for best S/N that could be achieved with new radios and listenability on
typical radios. Of course, with a new standard implemented in new radios,
one could use double-ended multi-band companding for greater benefit.

Tom