From: Ofir Carny <Ofir_at_vidius.co.il>
Date: Thu, 10 Apr 2003 14:49:42 +0200

Hi,
Coming from the area of telephony I always assumed that CD PCM (as in music CDs) was non linear too, is it completely
linear? I can't see any reason to do anything that way. Am I missing something here?
 
-----Original Message-----
From: Eric Woudenberg, Minidisc.org Editor
[mailto:minidisc_at_zoot.minidisc.org]
Sent: Thursday, April 10, 2003 3:14 PM
To: rockbox_at_cool.haxx.se
Subject: Re: Frequency as it relates to recording quality


Hi Ofir,

>AFAIR and according to
>http://www.wikipedia.org/wiki/Sampling_frequency the ph one uses
>8Khz, if the phone is good enough for you (and don't consider noise,
>etc., that's unrelated), it should be good enough for voice. My
>experience is similar, as long as I am rec ording voice, I can
>downsample as far as I can (at least with the original arc hos FW)
>and hear no difference, I can also compress a lot (I think I used the
>lowest level once). AFAIR phones uses 64kbps PCM - and that means 8
>bit depth (I ran across 32kbps ADPCM too),

US telephones use 8bit u-law (Europe uses 8bit a-law), which are
non-linear encoding schemes roughly equivalent in dynamic range to
12bit PCM (but with more noise).

>and that usualy seems OK to me, voice confere ncing and VoIP usually
>go so low that it's scary and so did early cell phones.

Telephones are not perfect for voice -- try differentiating an "S" and
"F" spoken alone.

Rick

**********************************************************************
This email and attachments have been scanned for
potential proprietary or sensitive information leakage.
Vidius, Inc. Protecting Your Information from the Inside Out.
www.vidius.com
**********************************************************************
Received on 2003-04-10