|
| quote: | Originally posted by vasyachkin
CD uses a 44 khz sampling rate that places a theoretical limit of its resolution at 22 khz ( according to Nyquist sampling theorem ).
however the practical limit of CD is 20 khz and the extra 2 khz are required to implement filtering that is needed to block all frequencies above 22 khz that would otherwise cause aliasing, which is a form of distortion. ideal filters that would allow a full 22 khz range are not physically realizable only exist in theory.
thus a CD does not actually contain frequencies above 22 khz and has very little between 20 and 22. |
Um, lowpass filtering does not take up an extra 2 kHz of bandwidth. I think you are confusing this with MP3 encoding, which typically filters tracks above 20 or 21 kHz because most people cannot hear above that range and that is also where the majority of distortion occurs.
I could see them doing it for CDs as well - but the reasons for this would not be what you said, more like:
1. The Nyquist rate is only a theoretical limit and assumes ideal reconstruction of the output signal; in reality, even if the sampling perfectly captures every frequency up to 22 kHz, the reconstruction process is not ideal, and those 22 kHz frequencies may actually get folded down to 1 or 2 kHz.
2. Most people cannot hear above 20 kHz or even 18 kHz and the intensity of thermal noise is proportional to the bandwidth of the system; hence, filtering sounds above 20 kHz helps reduce the noise floor and makes the recording more accurate.
3. Filtering with a cutoff frequency of 20 kHz does not actually mean that all sounds above 20 kHz will be eliminated; in reality there is an interval between the cutoff (3-dB or half-power) frequency and the point where attenuation is high enough to make it equivalently "zero". When you filter sound, you have a choice between a short cutoff interval and high ripple in the passband, or a longer cutoff interval and a relatively flat passband - obviously for audio, you want the flat passband, so you have to settle for a bit of a "hill" at the end of your filter spectrum. Thus any filtering at 20 kHz would probably be used in order to be absolutely sure that no frequencies above 22 kHz or so are present in the spectrum. This is kind of like what you said, except that the filter isn't actually eliminating frequencies between 20 to 22 kHz - it's just attenuating them, and they will still be present in the final encoding, just at lower amplitudes.
As for canine frequencies, who fuckin' cares.
___________________
My party schedule:
2009-02-21 - DJ Attention @ I'm So Popular
2009-06-18 - DJ Annoying @ People Need To Know Where I'll Be
2012-11-32 - DJ Insufferable ɸ Or At Least the Stalkers I Complain About
2048-06-66 - Spastic & Whocares ¶ Although I'm Actually Flattered
9999-45-81 - Tweaker Gimp ☼ I Probably Won't Even Go To This But I Have To Make Sure I Fill Up All The Available Space Here
|