Re: sample rate

[Tom Combs <tncombs@gmail.com>]

I thought the Nyquist theorem is all about reconstructing a signal once sampled.  A signal is not reconstructed by simply "connecting the dots", it is reconstructed by passing it through a low pass filter.  If you pass a sawtooth, triangle, square, or any other variety of periodic signal with frequency f through a low pass filter whose passband is perfectly flat up until frequency f, and then perfectly 0 beyond f, then what comes out is a pure sine wave of frequency f - that's because squares, triangles and whatnot all contain harmonics in addition to the pure sine wave that dictates its fundamental frequency.

 

A signal sampled at twice the Nyquist frequency can be perfectly reconstructed in a world of perfect filters and perfect sampling devices.