21 Apr
2010
21 Apr
'10
6:35 p.m.
On Wednesday 21 April 2010 02:47 pm, micromoog wrote:
While I consider this to be an academic discussion since I have high
frequency hearing loss, it does seem to me that with a sample rate of
44.1KHz, a 22.04KHz sine wave is indistinguishable from a 22.04KHz square
wave despite being below the Nyquist frequency, and an 11025Hz sine wave is
indistinguishable from an 11025Hz triangle wave. I just fired up Audacity
and generated a tone to verify this.
It seems like a pretty esoteric thing to care about, but if one does care
about frequencies above 11KHz, I guess he really might need to consider
sample rates higher than 44KHz. The Nyquist frequency is the threshold
above which tones can't be represented at all, not the threshold below
which tones are represented with any kind of fidelity.
Rob
It is easy to
understand that this correlation between phase and
correct amplitude also affects frequencies below half the
sampling-rate. Might be as
low as quarter of the sampling-rate, which in case of the CD is 11kHz.
This is a
pretty bold claim, and contradicts Nyquist and other
literature. Do you have a citation for the claim that frequencies "as
low as a quarter of the sampling-rate" are damaged by sampling?