I get the wrong signal to noise ratio using snr() from the Signal Processing toolbox.

Hi!

I need some help using snr().

When I use snr() on the column vector I want to calculate the SNR of, it returns -10.5526.

I plotted this column vector below, with the peak visible on the right. Clearly the SNR should be around 2-5.

Anyone knows what I am doing wrong?

Thanks in advance!

[AndresVar]

edit: clarifcation of snr(x) vs snr(x,y)

is that the full signal and are you saying the peak at index ~460 is the actual signal?

If you use snr(x) then the signal is assumed sinusoidal but your's is not.

So you must use snr(x,y) where y is the assumed noise.

To get noise estimate you can look at the spectrum in the window without signal, or you use the rssq as shown in the example: Signal-to-noise ratio - MATLAB snr (mathworks.com)

[Kevin Jansen]

Hey! thanks for your answer. Almost the whole range is just noise, only the peak at ~460 is actual signal.

Okay I think I just got it to work by your comment.

I have multiple of these column vectors, where the signal at ~460 stays around the same value.

I made a new vector containing all these signal peak values.

I also made a vector with the same size by taking the value at ~300 of all the data vectors I have.

I used these two signal vector and noise vector and put in in snr(xi,y), and it returned 4.17.

This seems to be correct, thank you!