That's an interesting one, surpassing Shannon more than 30 times! Surely good for a paper in Nature ;-)

I suspect that the noise measurement in EbNaut includes out-of-band energy, i.e. whatever is contained in the decimated IQ audio passband, but outside the symbol filter's bandwidth. In this the "local" signal from Rob was obviously accounted as "noise".  Wonder if there was another way to measure noise only inside the symbol bandwidth. If we take the complex symbol amplitudes, undo the PSK inversions after the decode, and apply a Fourier transform, the reconstructed carrier would ideally reside only in the central bin, allowing evaluation of average or quartile power from the other bins as noise.

PS: Joe's first transmisson arrived here straight away, despite daylight in Canada.

73, Markus

-----Ursprüngliche Mitteilung-----
Von: [email protected] <[email protected]>
An: [email protected] <[email protected]>
Verschickt: Fr, 7. Dez 2018 19:36
Betreff: Re: LF: LF and VLF EbNaut