[music-dsp] What is so special about 44100 Hz ?
sandeep at uisdl.com
Tue Jun 19 07:43:37 EDT 2001
Thanx for ur explanation.But I have doubt regarding the 44.1kHz.Why
perticularly 44.1kHz is used when u can use 44k,44.5kand so on ?
16 bits can represent up to 64 k. Even why Cd quality is 44.1kHz ?
----- Original Message -----
From: "Veli-Pekka Tätilä" <vepeta00 at oppi.edu.ouka.fi>
To: <music-dsp at shoko.calarts.edu>
Sent: Wednesday, June 26, 2002 2:26 PM
Subject: Re: [music-dsp] What is so special about 44100 Hz ?
> ----- Original Message -----
> >From: "sandeep" <sandeep at uisdl.com>
> To: <music-dsp at shoko.calarts.edu>
> Sent: Tuesday, June 19, 2001 10:14 AM
> Subject: [music-dsp] What is so special about 44100 Hz ?
> > hi,
> I am a newbie here but I think I know the ansers to your questions
(somebody please correct me if I am wrong).
> > Why is sampling frequency confirming to Nyquist condition chosen as
> > when any frequency >= 40 K Hz ( Twice of radio frequency range 20KHz)
> > suffice the condition ?
> (use small k to represent kilo that is *10^3)
> Nope, the sampling rate of 40 kHz wouldn't be good enough. If the signal
we are recording (at the sammpling rate of 40 kHz) contains any higher
frequencies than 20 kHz, these frequencies cannot be represented by our
sampling rate and they would produce aliasing (a frequency of x Hz above the
Nyquist will appear to be x Hz below the Nyquist in the actual recording).
So what do we do to get rid of the aliasing? The process is called
anti-aliasing and it could be done with a lowpass filter which we use to
filter out unwanted, too high, frequencies from the signal.
> Now the slope of our anti-aliasing filter cannot be indefinitely sharp.
The slope measures how quickly the amplitude of a signal drops for instance
18 dB/octave means that amplitude drops 18 dB when frequency is doubled (and
we are in a frequency range at which filtering takes place, in a lowpass
filter usually above the cutoff). As the filter always gradually attenuates
frequencies and doesn¨t remove them completely, some of the "material above
20 kHz will always pass through just with lower amplitude. For this reason
we use a slightly higher sampling rate of 44.1 kHz so we can more
succesfully filter out stuff above 20 kHz to get rid of some of the
> There's another reason why not tu use a sampling rate of 20 kHz. IMagine
recording a 20 kHz sine wave with the sampling rate of 40 kHz. What you get
is a square wave as only two samples are taken per one cycle of the sine
wave. YOu'd need much higher sampling rate to come even close to a "proper"
22 kHz sine-wave. With the sampling rate of 80 kHz you would still get only
4 samples of the sine-wave I mentioned. That's why (I think) some audio
programs let you choose the sampling rate of up to 192 kHz (though, I think
this is wasting of disc space).
> I can understand 24 bits and 48 kHz well but when we are talking about 64
bits (a far greater dynamic range can be represented than what we can hear)
and 192 kHz that's wasting of discspace in most cases!
> > Why always CD quality is measured and mentioned at 44100 Hz in most of
> > practical applications(Atleast in MP3)?
> Because audio CD:s use the following parameters: sampling rate is 44100 Hz
and 16 bits are used to represent one sample. That's why it's referd as
being CD quality.
> dupswapdrop -- the music-dsp mailing list and website: subscription info,
> FAQ, source code archive, list archive, book reviews, dsp links
dupswapdrop -- the music-dsp mailing list and website: subscription info,
FAQ, source code archive, list archive, book reviews, dsp links
More information about the music-dsp