[bct] Re: Question about digital voice recorders

  • From: "Neal Ewers" <neal.ewers@xxxxxxxxxxxxxx>
  • To: <blindcooltech@xxxxxxxxxxxxx>
  • Date: Tue, 20 Dec 2005 11:17:40 -0600

Dan, nice explanation.


-----Original Message-----
From: blindcooltech-bounce@xxxxxxxxxxxxx
[mailto:blindcooltech-bounce@xxxxxxxxxxxxx] On Behalf Of The Scarlet
Sent: Tuesday, December 20, 2005 11:13 AM
To: blindcooltech@xxxxxxxxxxxxx
Subject: [bct] Re: Question about digital voice recorders

Mary, at the risk of repeating what others have probably said more 
coherently, allow me.

All sound is analog by nature, it slides continually from soft to loud, 
from low to high frequency, like sliding down a slide.  Now, imagine
slide becomes a set of stairs, that is what happens in the conversion of

the analog sound to a digital signal.

Sample size is the actual number of bits that are in each sample of the 
audio signal.  16 bits is pretty standard, 24 is better.  It is like 
touching corduroy fabric.  If it is a wide wale, your finger will only 
touch a few rows at a time.  If it is narrow, your finger will touch
rows.  If it is really superfine wale, you will touch many rows with
just a 

In the same way, the larger the sample and the faster the sample rate,
more of the sound will be captured exactly as it sounded originally.

Your cd collection is all done at 16 bits per sample and at a sample
of 44,100 per second.  There is a law, if you will, that says that the 
highest frequency you can record is one half the sample rate, it is
the Nyquist frequency.  That means that at a rate of 44,100 samples a 
second, the highest frequency sound you can record is just over 22,000 
hertz, which is adequate for most things.

The size of the sample determines how many different sound levels are 
possible.  A 16 bit sample has 2 to the 16th power different sound 
levels.  A 24 bit sample has many more.  This means, in practical terms,

that the larger the sample size, the lower the noise and more detail
be picked up.  For most purposes, a 16 bit sample is adequate.

The bit rate is different.  It has to do with the compressing that takes

place when turning a pcm file, oops, I mean a wav file, into a
format like mp3.  A bit rate is expressed in thousands of bits per
so 128 kbps means 128,000 bits per second.  While this may seem fast, it
about one tenth of what pure digitized audio at cd quality would consist

The higher the bit rate, the more lifelike the sound will be.  Music at
kbps, a common mp3 compression, sounds fairly good until you hear drums,

cymbals and other high frequency sounds.  They end up being injured by
compression that takes place.  MP3 compression is called a lossy 
compression, meaning that some data is lost forever when the audio is 
compressed.  This is why turning it back into a wav file does not make
sound any better.

So, the ds-2 recording at a sample rate of 44,100 samples per second and
sample size of 16 bits gives you a nearly cd quality master file.  When 
this is compressed to 128 kbps, you do lose some quality, but you also
a file that is only one tenth the space of the master file.

I do not know, but the ds-2 may do the mp3 encoding on the fly, or maybe

not.  In either case, given the specs, it is capable of producing a
sounding audio file.  I figure that a half decent external microphone
contribute to the quality as they probably did not put high end mikes in
inexpensive digital recorder.

Hope this helps.  If I simply confused things too much, you may have my 
ration of chocolate for an hour...well, maybe for half an hour.


Other related posts: