On 23/09/06, steven c wrote:
Prior to the late forties, and the introduction of the Decca(UK)
"ffrr" wide-range recording system (a side result of research in
SONAR imaging) there were definite limits to the accuracy of the
recorded signal. Thereafter, whatever inaccuracies existed were
probably outside the bandwidth of human hearing. However, the 44.1
standard would by definition limit the upper frequency recordable
on a CD (I'm not exactly sure of how this works).
If there are 44k samples in a second, then a wave which lasts for
1/20,000 second can only be represented by two samples - an up and a
down. A sine wave of frequency 20KHz becomes a series of alternating
high and low sample numbers. There is no indication of the exact shape
of the wave.
Anything higher than that, and there is not enough information to even
say something is going up and down.