home *** CD-ROM | disk | FTP | other *** search
- Path: polarnet.com!floyd
- From: floyd@polarnet.com (Floyd Davidson)
- Newsgroups: comp.dcom.modems
- Subject: Re: ? Max speed for voice line
- Date: 2 Feb 1996 06:03:59 GMT
- Organization: __________
- Message-ID: <4es9gf$7hp@zippy.cais.net>
- References: <1996Jan31.112321@miavx1> <4erqau$6hb@forged.passport.ca>
- Reply-To: floyd@tanana.polarnet.com
- NNTP-Posting-Host: tanana.polarnet.com
-
- Bill Antic <vasant@passport.ca> wrote:
- >odchetveriko@miavx1.acs.muohio.edu (Oleg Chetverikov) wrote:
-
- >>I am taking a networking class and my prof said that the highest speed
- >>that can be achieved over the voice grade line (telephone line) is
- >>25,900 bps. I suppose he was talking about physical transmission, without
-
- Your prof is correct if he is saying that an average voice grade
- line might get about 26Kbps, but he is dead wrong if he says that
- 25.9Kbps is the absolute max.
-
- However, the discussion provided by Bill Antic, vasant@passport.ca,
- is _grossly_ in error.
-
- >I am sure that you are aware that the bandwidth of a voice grade line
- >if approx. 3300 Hz, which in turn results in a severe restriction in
- >the signaling speed, (refer to Shannon's theorem). Signaling speed,
- >therefore, is limited to a maximum of approx. 2400 baud (signal
- >transitions per second) due to the bandwidth of the voice grade lines.
-
- If you refer to Shannon's theorem it will be discovered that it
- says nothing about the maximum signaling speed, and relates only
- to maximum data transfer rates. If you refer to v.34
- specifications it might also be noted that only one of the possible
- combinations (one with the maximum bit rate that can be used,
- 28.8Kbps) requires 3300Hz or more bandwidth (244-3674Hz is
- required) and uses a signaling rate of 3429 (considerably more
- than 2400). Other combinations use 3200, 3000, 2800 and 2743
- symbol rates for data rates from 28.8 to 21.6Kbps.
-
- >The data transfer rate, on the other hand, depends on the type of
- >modulation used, and that is why we must differentiate between the
- >terms "baud" and "bits/sec". For example, if we employ a quad-bit
- >modulation, also known as quadrature phase-shift keying (QPSK), each
- >LINE transition would represent four bits of DATA, hence a 2400 baud
- >line would have a data transmission rate of 9600 bps.
-
- The above is correct.
-
- > This appears to
- >be the practical limit of data transfer WITHOUT DATA COMPRESSION.
-
- This is obviously incorrect. 28.8Kbps modems get 28.8Kbps with
- no compression. 33.6Kbps modems get 33.6Kbps without compression.
-
-
- > The extent to which data can be compressed depends on the type of
- >data, but generally accepted ratio is 4:1, more or less. If we take
- >the above example of 9600 bps and multiply it by the compression ratio
- >of 4, we get the maximum DATA transfer rate of 38,400 bps. This
- >therefore means that the DATA transfer rate over a voice grade line
- >could be as high as 38,400 bps while the signaling speed at the same
- >time is only 2400 baud. Of course you must realize that this is a
- >theoretical limit.
-
- The theory is relatively correct, but the numbers are wrong.
-
- > In practice we have to take into account the line
- >noise, type of data which will affect compression ratio, and some
- >other factors, resulting in a typical rating values of 28.8k or 33.6k
-
- Line noise does not affect the compression ratio. The type of data
- does.
-
- >for modems. You must keep in mind, however, that although a modem may
- >be rated at, say 28,800 bps, the actual transmission rate is usually
- >less, often 24,600 or even 20,000 bps. It depends on the condition
- >of the line, type of data, etc., etc.
-
- The type of data will not affect the actual transmission rate. The
- condition of the line (noise, bandwidth, etc.) will.
-
- >So your prof appears to be correct!
- >
- >I hope this will help.
-
- Oh?
-
- Floyd
- --
- Floyd L. Davidson Salcha, Alaska floyd@tanana.polarnet.com
-