On Tue, 23 Jan 2001 09:33:52 -0000, Talbot Andrew
<[email protected]> wrote:
Is it a coincidence I wonder, that the 'standard' Soundcard / CD
sampling rate is a multiple of the squares of the small prime numbers.
ie:
44100 = 2 * 2 * 3 * 3 * 5 * 5 * 7 * 7
Spooky !
Certainly simplifies decimation to a wide variey of lower rates. For
a CW tone centred on 800Hz, 11025 / 5 = 2205Hz would be ideal, letting a
CW filter of up to 400Hz bandwidth handle removal of the aliasing
products above 1102Hz. or 44100 / 5 = 8820 for a more accurate value
for voice bandwidth signals.
The original reason for the sampling rate in the 44 kHz region is that
video tape recorders were used widely in early digital audio
experiments long before the CD. There was usually three
samples/channel in every video line between the synch pulses. With 525
lines @ 60 Hz interlaced, that would give 47250 samples/channel.
However, there is a large number of lines used for vertical synch,
thus, the samples had to be averaged over a single field, giving rates
such as 44056 Hz or 44100 Hz.
In order to be able to store digital audio on both 525@60 Hz and
625@50 Hz video recorders, the sampling rate have to contain the 2, 2,
3, 5 and 5 primes. Thus, each video field would contain 735 resp. 882
samples and with 3 samples/line, 245 resp. 294 active lines/field or
490 resp. 588 lines/frame are used, the rest was taking by the
vertical blanking.
Paul OH3LWR
|
