Wired telephone bandwidth

A number no one here seems to know or understand.

You're clearly clueless, AlwaysWrong.

BozoBin Now Invisible....

sidelight: When I worked as a Broadcast Engineer, I kept trying to use the POTS for occasional Remote Broadcast purposes.

What Worked:

- On local (not through multiplex systems) exchanges, the actual bandwidth was about 100 to 5000 Hz. I built couplers on and off the dialed network that supported that bandwidth and they sounded good. Built an automatic answerer on a dedicated incoming line and the prearranged daily "remote studio" calls popped up a light on the control board, and the DJ could use the regular "remote talkback" switch to talk to them and confirm. The remote users then heard "Cue audio" of the station until they heard their cue, the switch went from CUE to LIVE and they were on the air.

- On long-distance calls, the same system was used but it was obvious the bandwidth was limited. The most obvious result was loss of lows so the voices were missing their nice quality. But it was OK for Sports broadcasts.

Since Back Then:

- Some hardware exists which sends the 50-300 Hertz audio as a subcarrier in the 2500-3000 Hz range. Some loss of "Highs" in the audio, but the Lows are restored.

- Some broadcast stations used Modems over the POTS to send processed audio. Worked pretty well.

- These days, of course, remote broadcasts are routinely sent as VOIP (but sometimes with specific software with better low end and automatic gain control etc..) Anybody currently in Broadcast?? Please tell us about the current practices...

DSL was seemingly an Impossible Dream "Back Then"...

I worked on ISDN a bit at IBM. It turned out to fit it's internal acronym: "I Still Don't Need It"...

Regards, Terry King ..On the Red Sea at KAUST snipped-for-privacy@terryking.us

AFAIK that is the way it is supposed to work here in the US.

In Europe, the lowest level of digital trunking E1 (2.048 MHz) is implemented with 32 time slots of which 30 carried the actual audio signals, one handled the synchronization and one handled the signaling.

The US T1 1.536 Mbit/s signaling at 24x64 kHz tried to handle both the signaling (off hook) and voice data in the same stream.

At least this was the reason given to us by some US telcos when trying to get a 64 kbit/s data stream to our US offices in the 1980's.

Sounds like the same sort of scheme. BT like NASA speaks entirely in TLAs, ETLAs and SETLAs. DACS stands for Digital Access Carrier System I think mostly made by Telspec. Is the US pair gain the same technology?

I was an early adopter and involved in the testing of my ISPs server upgrades. I had automatic logging and statistical analysis of all connects, EQM and failure codes. Some of it was published in their FAQs to help others get 56Flex and then later V90 to work reliably.

Certain firmware versions and combinations were very much more reliable than others. They were interesting times (and like you I had control of some registers that are no longer user configurable).

ADSL is cheaper here too, but if the local copper won't support it then ISDN still gives you real 64k raw digital without compression and 128kb with both channels bonded. UK price is about 2.5x that of an ordinary phone line and it means you have a pair of adjacent phone numbers for fax/phone. There was a time when ISDN was ludicrously priced to protect their incredibly lucrative Kilostream fixed line rentals.

ISDN is good enough to carry almost broadcast quality digital sound.

Even Belgium manages to do ISDN at a sensible price although their ugly phone connectors look like they are intended to carry 4A mains power!

I am nominally at the range limit ~12 miles of ADSL in a rural area but still get almost 4M connection on old wires buried underground. I'd hazard a guess that it might work at 1M these days out to 20 miles.

They may be better served by mobile phone data services then. That is becoming affordable here on USB dongles (eg 1 GB @3Mb/s is £10-12 pcm, although 3Gb @ upto 7Mb/s for £15 is better value).

Regards, Martin Brown

I had to do a live remote on a US Army base with less than an hour's notice in the early '70s. I sent our spare audio board to the building and had the other engineer call me on our quietest phone line, then he plugged the board into the phone line. I did the same at the studio and got decent audio for the broadcast.

Under normal remote use I had to request a copper pair between the sites at least a month in advance and hope they didn't lose it.

AKA: 'Innovation Subscribers Don't Need'

Back then, a lot of stations used portable UHF STL equipment for live remotes. Marti made most of the systems I saw.

--
Greed is the root of all eBay.

"Pair gain" is used for any technology that "gains" a line by using what's already there.

Yes, that's why the modem developers let us use their instrumented code, so we could report back on real ISPs and lines.

ISDN has always been out of the question in the US. The tariff structure has always been dumb. V.90 put the nail in ISDNs coffin.

They tell us 17K' (3 miles). I'm right at that limit and recently had to downgrade to 768K (after 1-1/2 years working at 3Mb).

That's way too expensive for anything but mobile use.

Hahhahaah... wuss. I enjoy regular connections at 18Mb/s

Over ISDN, too! Everyone with a little dick brags about their size.

coded=20

=20

Try V.90 and V.92, Huge constellations, requires very good SNR. I may not have enough bits per symbol though. V.34 does not have quite=20 the bitrate of so called 56K modems. Of course, the transmitted=20 symbol stream is compressed then scrambled from the incoming data stream.

Huh?

V.90 is V.34-like uplink and 8 kbaud 8 bit downlink. V.92 is 8 kbaud 8 bit both ways.

Bla-bla-bla ?

VLV

Again?

--
Greed is the root of all eBay.

Trellis coded=20

symbol=20

quite=20

stream.

Not that i consider Wikipedia a proper reference source, but check the = stated speeds:

V.34 seems to be 28kpbs both ways, a bit short of V.90 56kbps down 33.6 = kbps up. Haven't had my hands on any of the actual standards for many years.

coded

speeds:

Just got a notice from Cox, new speeds (CHSI Premier)... 20Mbps download / 4Mbps upload, and I think the actual upload is actually faster than claimed. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
      The only thing bipartisan in this country is hypocrisy

The standard bandwidth of a telephone system outside USA is 300Hz to

3.4 kHz, as a telephone transmission technician like myself knows. Outside USA, the CCITT specifications are used, whereas within USA the usual system is the one defined by Bell. I was trained in New Zealand. In the olden days of frequency division modulation, we would send each telephone call at a different frequency down the long-distance line, and we would have a system of demodulators (like very sophisticated radios) at the other end, tuned to each signal. So if there was too much bandwidth transmitted with some particular signal it would interfere with nearby channels, degrading the network. For this reason it was necessary to use hard filtering. Nowadays most telephone traffic is digital, meaning that analogue signals are sampled 8000 times a second and the voltage measurement digitally sent to signal reconstruction equipment at the distant end. By reason of the sampling rate, the highest frequency that can be represented in the signal is less than half of 8kHz. In practice, we prefer much less (than half), and we limit the signal reaching the sampling equipment to the same 300 - 3.4kHz that we used in the analogue days.

Trellis coded=20

symbol=20

quite=20

stream.

stated speeds:

kbps up.

Wonderful for you. But those are seriously different modems and line = conditions.

coded

speeds:

up.

There was an important difference. On most lines you had a sporting chance with V.34 of getting the top speed. On a V.90 connection you were doing well to get more than 80% of top rated speed in either direction.

By the time 56k chipsets were around the analogue V.34+ (1996) was capable of 33k6 both ways on a good day. On a DACSd line V.34 was always better than 56k because it tolerated the signal abuse better.

On a good line the 56k modem would be something like 50k down and 28k up. So for uploading big files it was worth forcing a V34+ connection.

Regards, Martin Brown