I have some digital recordings of some received over-the-air CW broadcasts of known plain-text (W1AW code practice) at known WPM code speeds, but in moderate to heavy RF noise and fading.

Can I describe a S/N ratio for the Morse Code in these recordings? If so, how would I calculate this S/N ratio? What bandwidths should I use for the signal and for the noise? Should I use recordings with or without the radio receivers narrow-band CW filter enabled?

If possible, I would like to use these S/N numbers to try and plot the decoding error rates versus the S/N ratios for various "demodulators" (including both software decoders, as well as humans at different experience levels).


1 Answer 1


Personally, I'd use whatever filtering scheme the actual receiver uses. If the narrow-band CW filter is exactly this, or even close, I'd use that. Then try to automatically (or manually) determine the regions where the signal is present. Measure the average power in the "off" regions, and that will be N. Measure the average power in the "on" regions, and that will be S+N. S = S+N - N, and then take 10*log10( S / N ). This will give you S/N ratio for the bandwidth of the narrow-band filter.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.