Wired telephone bandwidth

POTS is limited at the upper end by the DACs. They use 8k samples per second, so that's ideally a 4kHz bandwidth. I'm not sure what the low end limiter is, but it's supposed the spec is 300Hz.

No.

Ask Floyd L. Davidson up in alt.engineering.electrical

Bandwidth is specified as 300Hz - 3000Hz. 2700Hz width. Not

4k. They may provide more, but my more recent tests suggest they hard limit it pretty close these days. Back when (1950's), it was much more.

I've got a copy of the official specifications on the shelf, if anyone needs specific citations. They are designed to cover different phone and switching systems, but they all agree on this point.

There was a ruckus some years after when tape players made it into business places after WW II (german invention discovered in captured tanks, I think) where some folks would tape at normal speed and play fast for international calls to short the time (and expense), allowing the other side to record fast and then play back slower. I think that was part of the first "why" that AT&T started figuring it was worth some trouble to start sticking low pass filters in the line.

That's the story I heard and I'm sticking to it. But I have no real idea, at all. Sounded good when I heard it, maybe 30 years back, though.

Anyway, the easiest way to tell is just sit at a piano, hit keys, and let the other side tell you when it goes "clunk, clunk" instead of a nice tone.

Jon

Originally one pair of wires sent one long-distance call. Then, to save wire, a bunch of calls were sent SSB over one pair, over close-spaced carriers. The Western Electric "J" system used 4 KHz channel spacings, so the baseband audio was sharply filtered between

300 and 2600 Hz to avoid channel-channel crosstalk.

Later digital systems sampled at 8 KHz, so an antialiasing filter preceded the digitizer, usually about 2700 Hz.

The tape recorder thing doesn't make much sense. International phone lines never had enough bandwidth to allow that trick.

John

There remains the tariffs for POTS which state 300Hz to

3000Hz as the required bandpass. I believe I must have read them more than a few times, by now.

It was one of those tales I heard from someone working at the phone company, back then. I frankly had (and have) no idea.

Jon

I think the bandwidth is supposed to be 3KHz, but do mot know if that is the 3dB point and do not know the rolloff rate. However, traditionally (ie up to 5 years ago and maybe later) one could *reliably* get 48Kbaud data rate with a modem. But (some??) telcos have deliberately, in some unknown manner, throttled that so the best one can now do is 28.8Kbaud. And the response from then as well as from PUCs is "all we are required to support is VOICE QUALITY". BARF! AFAIK there is not a goddamn thing anyone can do about that. I say it will get worse as the greed of telcos will allow them to decrease that to maybe a max of 1Kbaud. Why? so that dial-up "customers" will get so fed up with slow internst that they will pay $$$ for "high speed" (which seems to never be as fast as advertised - and definitely not as reliable as dial-p).

Please comment as to WTF telcos did to throttle dial-up from 48Kbaud to 28.8Kbaud.

"The" telephone channel is usually specified as 300..3400 Hz. This is what is guaranteed to be transmitted to anywhere on the world. The bandwidth of the transmission line between you and the first electronic gadget it meets is from dc to something that is specific to each case.

Pere

Robert Baer wrote in news:- aSdnakxDbTj6RLWnZ2dnUVZ snipped-for-privacy@posted.localnet:

I'm still on dialup,and I usually log on at 48K,according to the W98SE taskbar indicator.Sometimes,it logs on at 49K,but it's never stable. ISTR that the max limit was 52K,but was usually limited by coils in the lines.

Century Link in central Florida is my telco.

--
Jim Yanik
jyanik
at
localnet
dot com

NO. NONE of you get those speeds. Those speeds are IF the handshake can pipe multiple streams into your NOISE FREE line. The line is NEVER noise free, and the actual hardware speed is NEVER more than 32k bits/per second, and even that is derived from stacking several 9600 baud streams together and that only happens on the best, most noise free connections, which are closest to their switch.

It is limited by the integrity of the link between you and your first digital switch point, which is usually not all that great.

jfgi

r second,

miter is,

ud

Google "concentrator" a widget that pipes more lines down existing copper pairs..

My beef is that cell uses so much compression, you cant use a modem with cell.

And I'm stuck in a area with no wimax, so its $$$$ cable or death, because the previous idiot renter of the apartment signed a paper allowing removal of the copper pair...

Steve

second,

is,

In the UK they call them DACS. Ugly prehistoric technology devices that totally destroy bandwidth to share one real copper line. eg

There used to be a site with a spectrum analyser on that could be dialed into on a modem and would return a 3 bands per octave spectrum.

OTOH you can use cellular for bulk datastreams and get reasonable price performance for modest amounts of data. No good for video on demand but fine for Usenet and casual web browsing.

Can't you get the telco to give you a real copper wire and pinch it off some old dear who just has a basic handset? That is what they do round here when the real copper is running out. Applying for an ADSL service from your supplier should get you out of the speed rut.

Regards, Martin Brown

2.6kHz bandlimiting significantly deteriorates the clarity of speech.

????

3.6 kHz that is.

There was an initiative for 16kHz sampling utilizing the same 56/64kbps carrier bandwidth (G.722 and such). That makes the quality significantly better; however it didn't receive wide acceptance.

The tape recoder idea originates from spy novels.

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

300Hz - 3600Hz

Nope.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------

Yup, that's better. My refs on the old FDM systems are vague, but suggest about a 300-3500 Hz signal band, maybe 3200 Hz net bandwidth, which fits into 4 KHz SSB spacing if you do the mod/demod well.

John

Not Kbaud, Kbits/sec. Baud rate is *considerably* lower (1200, or is it 2400?). Bit rates are obtained by using fancy modulation schemes yielding many bits per baud. IIRC, usually based on quadrature amplitude modulation.

Shannon sets the theoretical limit.

Probably down to SNR (Shannon again).

All they were ever required to support.

--
"Electricity is of two kinds, positive and negative. The difference
is, I presume, that one comes a little more expensive, but is more
durable; the other is a cheaper thing, but the moths get into it."
                                             (Stephen Leacock)

Probably not much...

What basically happened, was that more and more, there were 2 or more subscribers added to a single cable pair, with carrier equipment to share that pair between them. The carrier equipment didn't have the bandwidth needed to support the higher modem speeds (that basically, relied on the fact that the digital CO was digitizing right at the cable head, so why go back to analog at the ISP...) so people got frustrated. To the telcos, they were just selling you a voice line, guaranteed to be usable by a telephone. If you really needed digital transmission, then you should have to pay for a specially conditioned line that gave you that much throughput! BTW, the data line usually cost more than a DSL line does today...

Charlie

more that 1 wpm and less than 35 wpm.

-- --- .-. ... .

--
Adrian C