Any simple method to detect noise level of a wave

[alex.]

As the attached wave, is there any simple method to detect noise level of a wave.

The sampling rate is 100M, the time duration is 20ms.

Any suggestion, thanks.

[GerdW]

Hi alex,

"any simple method"

I would say no. It all depends on what you define as noise - others may define those spikes in your pic as noise...

[johnsold]

As GerdW said, it depends on how you define your signals and noise.  What do you know about your signals which is different from the noise? Are the signals always larger than some threshold and the noise never reaches that threshold? Do the signals always occur at known times or frequencies?

Lynn

[alex.]

Actually I just want to detect the base line voltage level, the signal should be some random pulse.

Can wavelet package do this?

[ABCPrograms]

Noise is most of the time a signal with a high frequency.  A low pass filter might do the trick

But before you solve a problem like this, you should take a look at your system and question yourself where this noise is comming from.

Are you reading this signal from a DAQ card?  In which way?

[alex.]

Low pass filter may not work, need to reserve the low pass part of the pulse.

The noise come together with the signal on site. 

[ABCPrograms]

And can you tell me how you're measuring this signal?  If you're doing it with a DaqCard, you might want to think of measuring your signal differentially.  This is a lot more noise insensitive

[Henrik_Volkers]

Take your data and make a histogram of your values to get a feeling for the distributuion. Mostly a gaussians bell shape 🙂

Calc the mean and the standard deviation.  3 times the standard dev. (say an educated guess 😉 ) should be a good starting point.

or

Apply a light (5 point) median filter (filter the spikes) and look for the min/max ....

or

Get to know more about your noise and do more math ...

[alex.]

The noise come along with the source of the signal together, so can only play with software to remove it.

At different site, the noise level is different, the pulse desity is also different. 😞

[ABCPrograms]

Try the median filter like suggested.  That might give you a better signal.  Just don't make it to big, cause you will loose you pulses you need as well