Le Wed, 29 Oct 2008 15:25:51 -0700, Eric Jacobsen a écrit :

> On 29 Oct 2008 22:06:37 GMT, Laurent <xx@gmail.com> wrote:
> 
>>Hi all,
>>
>>Im working on an (synthetic) acoustic experiment: I have a receiver that
>>records waves emitted by two sources S1 and S2. So the receiver records
>>R=S1+S2. I would like to "filter" R to get only S1 or S2.
>>
>>Is there any way of doing that ??
>>
>>S1 and S2 are defined in the same frequency band. So i cannot use a
>>simple filter. Im thinking about modulating/demodulation the signals S1,
>>and S2. However the modulated signals have to be in the same frequency
>>band than S1 and S2. So i cannot use FM or AM modulation.
>>
>>
>>I tried to modulate S1 and S2 by two different white noise. It works
>>more or less, but it is not very accurate.
>>
>>See you,
>>Laurent
> 
> Yes, you can use the same channel to send both S1 and S2.    Perhaps the
> most straightforward thing to do would be to spread each a bit with
> orthogonal spreading codes and correlate on the separate codes in the
> receiver.
> 
> Elsewise you can use some example of Space-Time codes or spatial
> multixplexing assuming S1 and S2 are transmitted from different
> locations and the reflections at the receiver are different enough to
> separate them.    Since that depends on the channels cooperating
> favorably it may not always work, so I'd think just using different
> spreading codes would provide a suitable method.
> 

Thank you very much for your answer. 
In my case S1 and S2 are emitted from different locations, but have quite 
similar waveforms. Ultimately i would like to use about 150 sources, and 
be able to retrieve their waveforms accurately. 

I didnt know about orthogonal spreading code ! I have just read a few 
things about it, and it may work in my case, i have still to try it. 

See you,
Laurent

On 29 Oct 2008 22:06:37 GMT, Laurent <xx@gmail.com> wrote:

>Hi all, 
>
>Im working on an (synthetic) acoustic experiment: I have a receiver that 
>records waves emitted by two sources S1 and S2. So the receiver records 
>R=S1+S2. I would like to "filter" R to get only S1 or S2. 
>
>Is there any way of doing that ?? 
>
>S1 and S2 are defined in the same frequency band. So i cannot use a 
>simple filter. Im thinking about modulating/demodulation the signals S1, 
>and S2. However the modulated signals have to be in the same frequency 
>band than S1 and S2. So i cannot use FM or AM modulation. 
>
> 
>I tried to modulate S1 and S2 by two different white noise. It works more 
>or less, but it is not very accurate. 
>
>See you,
>Laurent

Yes, you can use the same channel to send both S1 and S2.    Perhaps
the most straightforward thing to do would be to spread each a bit
with orthogonal spreading codes and correlate on the separate codes in
the receiver.

Elsewise you can use some example of Space-Time codes or spatial
multixplexing assuming S1 and S2 are transmitted from different
locations and the reflections at the receiver are different enough to
separate them.    Since that depends on the channels cooperating
favorably it may not always work, so I'd think just using different
spreading codes would provide a suitable method.

Eric Jacobsen
Minister of Algorithms
Abineau Communications
http://www.ericjacobsen.org

Blog: http://www.dsprelated.com/blogs-1/hf/Eric_Jacobsen.php

Hi all, 

Im working on an (synthetic) acoustic experiment: I have a receiver that 
records waves emitted by two sources S1 and S2. So the receiver records 
R=S1+S2. I would like to "filter" R to get only S1 or S2. 

Is there any way of doing that ?? 

S1 and S2 are defined in the same frequency band. So i cannot use a 
simple filter. Im thinking about modulating/demodulation the signals S1, 
and S2. However the modulated signals have to be in the same frequency 
band than S1 and S2. So i cannot use FM or AM modulation. 

 
I tried to modulate S1 and S2 by two different white noise. It works more 
or less, but it is not very accurate. 

See you,
Laurent