10-24-2012 08:08 AM
Hi every body
any one can help me to find the signal to noise ratio of the attached signal, and how to convert it to frequency domain. I know it is easy but it is not working with me since the power must have a main amplitude at the fundemental frequeny which is 1Hz and I cann't find any component there.
any help is appreciated
qqa
10-24-2012 08:45 AM
What have you tried? Look at the Signal Processing->Waveform Measurements palette. There are some nice goodies in there.
10-24-2012 09:05 AM
Thanks for your reply Kudos
I've used most of them Ive used the FFT to find the signal in frequeny domain and then I want to find the power of the signal,but I don't know how I wil find the power if the nosie in the signal at same time??????????
have you any idea?
10-24-2012 10:29 AM
The SNR is defined as the ratio of signal power / noise power. So if you have an FFT you must decide which part of the spectrum is your signal. Then you can sum up the signal part of the spectrum giving you the signal power and the sum of the rest of the spectrum gives the noise power.
Cheers
Edgar
10-24-2012 11:15 AM
@ejkaiser wrote:
The SNR is defined as the ratio of signal power / noise power. So if you have an FFT you must decide which part of the spectrum is your signal. Then you can sum up the signal part of the spectrum giving you the signal power and the sum of the rest of the spectrum gives the noise power.
Cheers
Edgar
It sounds so easy - all you have to do is separate the signal from the noise, measure both, then take the ratio... Of course, if we could do that, we wouldn't have to worry about SNR in the first place! 😉
10-25-2012 02:13 AM
@billko wrote:
@ejkaiser wrote:
The SNR is defined as the ratio of signal power / noise power. So if you have an FFT you must decide which part of the spectrum is your signal. Then you can sum up the signal part of the spectrum giving you the signal power and the sum of the rest of the spectrum gives the noise power.
Cheers
Edgar
It sounds so easy - all you have to do is separate the signal from the noise, measure both, then take the ratio... Of course, if we could do that, we wouldn't have to worry about SNR in the first place! 😉
Well Billko,
what is your message then?
This is exactly the procedure and LV provides the tools to do it. But as always: first one needs to understand the task, then the tools. We have to worry about SNR because it inevitably spoils the signal, not because it couldn't be measured.
Cheers
Edgar
10-25-2012 05:19 AM
Thanks you edgar
10-25-2012 10:16 AM
@ejkaiser wrote:
@billko wrote:
@ejkaiser wrote:
The SNR is defined as the ratio of signal power / noise power. So if you have an FFT you must decide which part of the spectrum is your signal. Then you can sum up the signal part of the spectrum giving you the signal power and the sum of the rest of the spectrum gives the noise power.
Cheers
Edgar
It sounds so easy - all you have to do is separate the signal from the noise, measure both, then take the ratio... Of course, if we could do that, we wouldn't have to worry about SNR in the first place! 😉
Well Billko,
what is your message then?
This is exactly the procedure and LV provides the tools to do it. But as always: first one needs to understand the task, then the tools. We have to worry about SNR because it inevitably spoils the signal, not because it couldn't be measured.
Cheers
Edgar
The point was to show that it is very easy to go down the wrong path if you have no guidance. For instance, the OP painted themselves into a corner by believing that they needed to separate the signal from the noise to measure each, but you explained the real world, correct way to do it. I was just pointing out the subtle flaw in the logic that if we could separate the signal from the noise, we wouldn't need to measure the noise because it would be irrelevant.