Simple Software Modem

Hello,

Im at yet another DSP challenge ...

Im trying to implement a simple software modem to transmit 4kbps over a
3Khz radio channel. Any suggestions on what the absolute simplest form
of digital modulation to be used?

Thanks!
--Shafik Amin

On 18 Jul 2006 13:33:13 -0700, "Shafik" <shafik23@gmail.com> wrote:

>Hello,
>
>Im at yet another DSP challenge ...
>
>Im trying to implement a simple software modem to transmit 4kbps over a
>3Khz radio channel. Any suggestions on what the absolute simplest form
>of digital modulation to be used?
>
>Thanks!
>--Shafik Amin

What do you know about the channel?   If the SNR is high and the
channel is flat then PAM is pretty easy to get working.   The channel
assumptions are pretty important, though.

Eric Jacobsen
Minister of Algorithms, Intel Corp.
My opinions may not be Intel's opinions.
http://www.ericjacobsen.org

Shafik wrote:
> Hello,
>
> Im at yet another DSP challenge ...
>
> Im trying to implement a simple software modem to transmit 4kbps over a
> 3Khz radio channel. Any suggestions on what the absolute simplest form
> of digital modulation to be used?
>
> Thanks!
> --Shafik Amin

I don't know how to measure simplicity in your context, but I would
suggest differential PSK. This kind of modulation is bandwidth
efficient, can be detected noncoherently, and is insensitive to
amplitude variations. If you choose D8PSK, a symbol rate of 1500 baud,
and use a root-rasied cosine filter, you have a shot. Try a one baud
delay-conjugate multiply for demodulation.

Some of the old telephone modems used this modulation. V.27 is one that
comes to mind. Go download the spec from ITU web site (I think you get
three specs free) for info.

John

johns@3db-labs.com wrote:

(snip)

> I don't know how to measure simplicity in your context, but I would
> suggest differential PSK. This kind of modulation is bandwidth
> efficient, can be detected noncoherently, and is insensitive to
> amplitude variations. 

Why, then, were most early modems, such as the 103, FSK?

Though that assumes analog filters, which is probably the
wrong assumption for this newsgroup.

-- glen

glen herrmannsfeldt wrote:
> johns@3db-labs.com wrote:
>
> (snip)
>
> > I don't know how to measure simplicity in your context, but I would
> > suggest differential PSK. This kind of modulation is bandwidth
> > efficient, can be detected noncoherently, and is insensitive to
> > amplitude variations.
>
> Why, then, were most early modems, such as the 103, FSK?
>
> Though that assumes analog filters, which is probably the
> wrong assumption for this newsgroup.
>
> -- glen

FSK doesn't require analog filters, and I would say that it is simpler
than DPSK. The reason I didn't suggest it because I don't think simple
FSK can meet the BW objective.

John

glen herrmannsfeldt wrote:
> johns@3db-labs.com wrote:
> 
> (snip)
> 
>> I don't know how to measure simplicity in your context, but I would
>> suggest differential PSK. This kind of modulation is bandwidth
>> efficient, can be detected noncoherently, and is insensitive to
>> amplitude variations. 
> 
> Why, then, were most early modems, such as the 103, FSK?
> 
> Though that assumes analog filters, which is probably the
> wrong assumption for this newsgroup.

In those modems, the baud rate equaled the bit rate. The OP's 
specification doesn't allow that.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

Jerry Avins wrote:

(snip regarding the simplicity of FSK modems)

> In those modems, the baud rate equaled the bit rate. The OP's 
> specification doesn't allow that.

I don't know that it is required that only two frequencies
are used, though that seems usual.  It might be, though,
that PSK is easier with more than one bit/baud.

-- glen

glen herrmannsfeldt wrote:
> Jerry Avins wrote:
> 
> (snip regarding the simplicity of FSK modems)
> 
>> In those modems, the baud rate equaled the bit rate. The OP's 
>> specification doesn't allow that.
> 
> I don't know that it is required that only two frequencies
> are used, though that seems usual.  It might be, though,
> that PSK is easier with more than one bit/baud.

Modems usually go beyond 4 phases before amplitude becomes a variable. 
If independent quadrature carriers are amplitude modulated, simply 
turning them on(+/-) and off yields 8 phases. The 45-degree phases have 
greater amplitude, making room for half-amplitude pulses without sinking 
below the noise. Great minds took it from there and flew with the idea.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

Thanks for all the input guys.

--Shafik