[amsat-bb] Re: Icom D-Star

[Gordon JC Pearce]

On Sat, 2011-04-23 at 10:42 -0500, Gregg Wonderly wrote:
> In the end, digital compression of spectrum space is going to happen more and 
> more.  AM style broadcast is hugely inefficient even though it is painfully 

Okay, but *why*?  Why are we so obsessed with squeezing bandwidth down
and down, at the expense of intelligibility?

I've got my spectrum analyser hooked up to my 2m aerial at the moment.
For the past half hour it has indicated the odd little spike at
144.800MHz indicating a little bit of (weak) APRS traffic, a big spike
at the output of GB3CS (because it's line-of-sight), a couple of
slightly smaller spikes from the other two local repeaters (PA and KE)
and a bump where FE, FF and AY are supposed to be (they're quite weak
here).

Other than about 300kHz of repeater outputs and 25kHz of packet, the
rest of the 2m band is *empty*.

If I switch it to scan 70cm, I'll see GB3KV (co-sited with CS) and
nothing else, except the odd satellite up around 435MHz and a brief
burst when my heating oil tank gauge decides to tell me I need to buy
more oil.

I could safely use channels 250kHz wide on 70cm, if I had a mind to do
so.  It wouldn't get in anyone's way, because there's no-one there to
annoy.  This is even more true of 23cm, and higher.  We've got loads of
space to play with, on all the bands except 30m and 60m which have their
own kind of charm.  

> simple to do.  I don't really believe that D-Star is the right choice for 
> "everything" because it is single source.  But, so is Microsoft windows, 
> MacOS-X, and many other software based systems.  If you are an FPGA programmer, 
> perhaps you can build an FPGA based CODEC for amateur radio that would do voice 
> compression etc.  But in the end, you also have to have an transmitter with the 
> appropriate bandwidth output to reduce the spectrum used.

This is where D-Star falls down - it's *still* just a 12.5kHz-wide
channel.  Without getting into linear PAs and the like, it's going to be
quite hard to do anything else and have a useful data rate.

As for FPGAs, why not just use a cheap general-purpose DSP or even CPU?
That's what people tend to end up implementing on the FPGA anyway.

Gordon MM0YEQ