[time-nuts] getting accurate timing on RTL-SDR output

[Magnus Danielson]

Hi Jim,

> That's the sort of strategy I'm looking at.. find a set of samples where
> the transient occurs, zero pad it out (so that when I do peak picking
> later, I've effectively got interpolation), then run the correlation

You find it using cross-correlation, and once found, you have a fair
idea for the next round where it is so a shorter cross-correlation can
be used if processing is relevat.

> ifft(fft(a).*conj(fft(b))) in matlab

After that, search through for the highest amplitude, which is a linear
search.

The trick of using a reference is that the conj and fft can be done as
common processing and then only the fft, multiply and ifft needs to be
done for each other channel. This removes 1/3 of the needed FFTs.

Cheers,
Magnus