How can I calculate the SNR of a signal from its frequency spectrum?

I am pulling data from a spectrum analyzer and am trying to use the snr function from the signal processing toolbox to calculate the signal-to-noise ratio of the signal. The problem is that the value I am getting for the SNR makes no sense. I am calling the function like this:

snr(power, freqValues, rwb, 'power')

And the value of the snr I get is -28.0343 +13.6438i with rwb = 1.8e4;

The graph created by snr looks like this:

And for reference this is what the frequency spectrum I am capturing looks like:'

(This is created from plot(freqValues, power))

The frequency ranges from 0-2MHz and the major peak is at 902kHz @10.4dBm.

Any help is appreciated thanks.

[Mathieu NOE]

hello

you have to come back to power first , compute the "good signal" power from your fft amplitude between 0.8 and 1 Mhz

and do the same on the remaining portion of the spectrum => you get the power of the background noise (+ other peaks)

SNR = power signal / power background noise

SNR in dB = 10*log10(SNR)

[Mathieu NOE]

Question

in your code :

snr(power, freqValues, rwb, 'power')

is power expressed in dB as in your graph ? it should be in linear (units²) not in dB