[sdiy] Harmonic bandwidth

[Nicholas Gregorich]

Scott Gravenhorst wrote:
> Another thing that just occurred to me is that higher harmonics are represented by fewer samples
> and as such, I believe their approximation will be less accurate.  So as we add more and more
> sinewaves that are less and less accurate, this effect seems to stand to reason.
> 

Hmmm, are you using 100Hz as the fundamental? If so the 32nd harmonic is 
only 3.2kHz. Would inaccuracies due to sample rate become important at 
this low of a frequency? The 32nd harmonic is 40dB below the fundamental 
as well.

Although the effect stands to reason, Nyquist says that 44.1kHz should 
be more than ample to represent a 3.2kHz sinusoid. I've heard strong 
arguments on both sides of the 'oversampling/Nyquist' debate but I've 
yet to go investigating myself.

Nick.