filtering and frequency separation

Sirs,
I have a question on the filtering.
Basically, assume a signal contains two signals that are of different
frequencies.
In this case, what are the requirements from filtering perspective in order
to remove one of the frequency.  Are there any min frequency separation
requirements in order for the filter to operate?

Thanks, manish

On Sun, 02 Jun 2013 06:04:01 -0500, "manishp" <58525@dsprelated>
wrote:

>Sirs,
>I have a question on the filtering.
>Basically, assume a signal contains two signals that are of different
>frequencies.
>In this case, what are the requirements from filtering perspective in order
>to remove one of the frequency.  Are there any min frequency separation
>requirements in order for the filter to operate?
>
>Thanks, manish

Maybe.    Much depends on how much latency and processing complexity
you're willing to spend to solve the problem.


Eric Jacobsen
Anchor Hill Communications
http://www.anchorhill.com

On Sunday, June 2, 2013 11:04:01 PM UTC+12, manishp wrote:
> Sirs,
> 
> I have a question on the filtering.
> 
> Basically, assume a signal contains two signals that are of different
> 
> frequencies.
> 
> In this case, what are the requirements from filtering perspective in order
> 
> to remove one of the frequency.  Are there any min frequency separation
> 
> requirements in order for the filter to operate?
> 
> 
> 
> Thanks, manish

also depends on the number of sensors you have. It is possible to separate signals with overlapping spectra if you know the statistical properties. This is how blind source separation works (independent component analysis)

On Jun 2, 7:04&#4294967295;am, "manishp" <58525@dsprelated> wrote:
> I have a question on the filtering.
> Basically, assume a signal contains two signals that are of different
> frequencies.
> In this case, what are the requirements from filtering perspective in order
> to remove one of the frequency. &#4294967295;Are there any min frequency separation
> requirements in order for the filter to operate?

what everyone else is saying.

but from what you're describing, if the frequency components in the
signal do not overlap, it's a matter of MIPS and of throughput delay
(assuming your filter is real time) and of how many dB the removed
frequency is attenuated.  at those three parameters is where the
tradeoff is

r b-j
.

>On Jun 2, 7:04=A0am, "manishp" <58525@dsprelated> wrote:
>> I have a question on the filtering.
>> Basically, assume a signal contains two signals that are of different
>> frequencies.
>> In this case, what are the requirements from filtering perspective in
ord=
>er
>> to remove one of the frequency. =A0Are there any min frequency
separation
>> requirements in order for the filter to operate?
>
>what everyone else is saying.
>
>but from what you're describing, if the frequency components in the
>signal do not overlap, it's a matter of MIPS and of throughput delay
>(assuming your filter is real time) and of how many dB the removed
>frequency is attenuated.  at those three parameters is where the
>tradeoff is
>
>r b-j
>.

Yes. When the frequencies are separted. The questiin is, how much do should
the fre be off?
F1 10 khz f2 11 khz or f1 10 khz f2 12 khz

Thanks, manish
>

On Jun 2, 10:39&#4294967295;pm, "manishp" <58525@dsprelated> wrote:
> >On Jun 2, 7:04=A0am, "manishp" <58525@dsprelated> wrote:
> >> I have a question on the filtering.
> >> Basically, assume a signal contains two signals that are of different
> >> frequencies.
> >> In this case, what are the requirements from filtering perspective in
> ord=
> >er
> >> to remove one of the frequency. =A0Are there any min frequency
> separation
> >> requirements in order for the filter to operate?
>
> >what everyone else is saying.
>
> >but from what you're describing, if the frequency components in the
> >signal do not overlap, it's a matter of MIPS and of throughput delay
> >(assuming your filter is real time) and of how many dB the removed
> >frequency is attenuated. &#4294967295;at those three parameters is where the
> >tradeoff is
>
> >r b-j
> >.
>
> Yes. When the frequencies are separted. The questiin is, how much do should
> the fre be off?
> F1 10 khz f2 11 khz or f1 10 khz f2 12 khz
>

let's say you're trying to keep the 10 and get rid of the 11 or the
12.  for a given low-pass filter, it's likely that 12 kHz will be
attenuated to a lower level than will the 11 kHz.

what you do is specify what the "transition bandwidth" (the space
between the passband and the stopband) is, and how much attenuation
you need in the stopband and how much attenuation you can tolerate in
the passband.  from that, you can get a handle on the order of the IIR
filter or the number of taps for an FIR.

r b-j

On Sun, 02 Jun 2013 06:04:01 -0500, manishp wrote:

> Sirs,
> I have a question on the filtering.
> Basically, assume a signal contains two signals that are of different
> frequencies.
> In this case, what are the requirements from filtering perspective in
> order to remove one of the frequency.  Are there any min frequency
> separation requirements in order for the filter to operate?
> 
> Thanks, manish

What's been said so far is true, but they don't seem to be answering the 
question you ask.

In theory-land, there is no minimum frequency separation -- as long as 
you have distinct spectra, you can filter them out from each other.

The theoretical limitation is that the more tightly you filter in 
frequency, the longer the filter will take to settle in time.

The more practical limitations are that the more tightly you filter in 
frequency, the more complicated the filter gets (more taps for an FIR, 
more stages for an IIR), and the more processing power, memory, etc., you 
need to implement it.

-- 
Tim Wescott
Control system and signal processing consulting
www.wescottdesign.com

Thanks a lot. Theoretical limit is what I was looking at.
I should have stated the background behind my question. 
In case of case of communication with 2 frequencies, I want to know how
much the frequencies be separated

Qthanks, manish

If you are using two frequencies to communicate, then you must be using some modulation technique, in which case your spectrum is no longer just two frequencies. So the minimum distance between your two frequencies is determined by how much spreading occurs in the spectrum as a result of the modulation technique.

Bob

On 6/3/2013 11:12 PM, radams2000@gmail.com wrote:
> If you are using two frequencies to communicate, then you must be
> using some modulation technique, in which case your spectrum is no
> longer just two frequencies. So the minimum distance between your two
> frequencies is determined by how much spreading occurs in the
> spectrum as a result of the modulation technique.

No.

Common mistake: taking the limit of particular method for
theoretical limit.

The minimum frq distance between comm channels is determined only by the
amount of indeterminism of the problem.

Vladimir Vassilevsky
DSP and Mixed Signal Designs
www.abvolt.com