Archived from groups: comp.dcom.voice-over-ip,comp.os.linux.networking (
More info?)
rdgentry1@cablelynx.com (P Gentry) wrote:
>floyd@barrow.com (Floyd L. Davidson) wrote:
>>
>> The sound card's output does not need to go to the modem,
>> just to the phone line. That is fairly easy to accomplish,
>> either using the modem, or not.
>
>I'll give you my phone # if you think it's that easy
Really, it is. Every sound card has a "line in" and a "line
out" jack. You want that to go to a telephone line? Put any
standard telephone hybrid network between the sound card and the
telephone line! Every phone patch, telephone set, modem,
etc. etc. has such a network in it. (The network from a
telephone set won't work though, only because it has a
"sidetone" path that is great for a telset, but bad for this
application.)
>> >Modems (_mo_dulator/_dem_odulator) use the analog _circuit_ facilities
>> >of the POTS (ie., PSTN) to make a connection to another DTE at the far
>> >end -- phones are not DTE's.
>>
>> Well, yeah, but that isn't significant.
>
>Just what, pray tell, will the modem "talk" to at the other end then?
>A modem modualtes/demodulates the "audio" signal into a byte stream.
>That's why they require settings like 8NP or 7NP or 7+P (# of
>bits/parity)
The modem need not talk to anything. It will quietly sit there
doing nothing if you don't enable the transmitter. The usual
configuration is that one modem is in "originate mode", which
does *not* send tones until triggered by reception of tones.
The other modem is in "answer mode", which immediately does send
tones. As long as the modem being used is in originate mode
there will be no tones unless you dial up a modem that is in
answer mode.
>> >Two modems connected to each other via POTS use a terminal emulator
>> >for human use. Thus all the terminal emulation options available in
>> >modem software packages.
>>
>> Eh?
>
>Surely you're kidding. Dial-in BBSes, VT100 and ANSI terminal menus,
>command sets, and keyboard layouts -- mostly of historical interest
>perhaps? Modems are a _byte_ stream _digital_ device -- and
>asynchronous at that. In this context, think telnet.
The first sentence of the original statement is wrong. It
flatly says that *all* modems connected via POTs for human use
necessarily use a terminal emulator. That is simply not true.
I've written dozens of small serial port programs that connect
two modems via POTS and talk to some remote device (everything
from telco Channel Banks to remote temperature sensing devices).
Most of those uses do *not* involve a terminal emulator.
Your followup isn't too great either! Modems are not "byte
stream" devices, nor are they asynchronous. The interface to
the computer is both of those; but the communications between
modems is almost always (as in, I don't know of a single modem
that is not) a bit oriented data stream. With v.32 and v.90 it
is synchronous too. (Actually, even Bell 103 modems are
synchronous at the modem to modem level, but they are not
packetized so that distinction is meaningless.)
>> >Modems can also be used to dial into a network (connecting to another
>> >DTE device) and have that far end "plug" you into the network. This
>> >is normally done with PPP -- point-to-point protocol. PPP is a
>>
>> Eh?
>
>RADIUS? AOL? MSN? Huh?
You should say "often", not "normally". Perhaps the use you are
familiar with is normally just that. Others may normally see
something very different...
....
>> >Point is that a modem is a (digital) _data_ device, not an analog
>> >transmitter despite the fact that it uses the analog (voice) circuit
>> >facilities of the "phone company".
>>
>> A modem is in fact *both* a digital device and an analog
>> "transmitter".
>
>Yes, that's why they are called mo-dems, but they have no ability to
>transmit an _un_modulated voice signal. The mo and the dem is how
>they "encode" and "decode" the digital bits.
A modem today is not solely a modulator and a demodulator. Most
modems include a pretty serious bit of computational capability,
not to mention signal processing. Some of them even seem to have
the ability to transmit "an _un_modulated voice signal"!
The simple "modulator/demodulator" was actually called a
"terminal unit", and the telephone industry used them to
multiplex Teletype channels onto a carrier voice channel. (A
43A1 TU used about 5 tubes if I remember right, would run at
speeds up to 75 baud, and could be stacked with as many as 16 to
a single voice channel because it used 180 Hz FSK!)
(Speaking of making modems do things the manufacture never meant
them to... I used to use 43A1 Terminal Units as a tube tester!
They used WECO 408 tubes, as did many other pieces of equipment,
X type SF signaling units being one example. But the SF Unit
didn't need a particularly high gain 408 tube, and the 43A1 did.
So I set up a row of about 48 sockets with the only filaments
wired up, and would plug 4 dozen tubes in and let them burn in
for a few days to stablize them. Then I'd take them one at a
time an plug them into a spare 43A1, with an AC VTVM probe on
the output pin jack. Tube went back into boxes, sorted by which
would be used for SF Units, other equipment, or 43A1's. Low
gain, medium gain, and high gain... Eventually that was no
longer necessary because WECO started doing essentially the same
thing, and sold a 408A/B tube specifically for the 43A1 units.)
Whatever, when "modem" concept first moved to the customer
location the speed and bandwidth were slightly increased (135
baud!), and an external dialing unit was required. But shortly
the two were combined, and a "modem" as we know them was born...
and has *never* since then been restricted to just a modulator
and a demodulator.
A modern modem can probably handle voice. It can also do a
significant bit of digital signal processing (of the analog
signal) because every v.32, v.34, v.90 or v.92 modem contains a
digital echo cancellor. A pretty fancy little device, actually.
>> >The voice _facilities_ of the phone companies would be available with
>> >an interface card that speaks FSO/FXO.
>>
>> That is totally bogus too. FXS/FXO merely means that instead of
>> connecting a wire line loop to an interface designed for a
>> wireline loop, there is a non-wireline carrier system in
>> between. The FXS provides an interface to the real wireline
>> loop (e.g., a telephone set) and the FXO provides what looks
>> like a wireline loop to the interface (e.g., a line card in a
>> telephone switch). The carrier facility in between can be
>> almost anything (fiber, microwave, digital carrier, analog
>> carrier, whatever).
>
>Was only included to indicate the "reality" that a FXS/FXO capable
>card is the only reasonable ($) hope of carrying on a "relayed"
>connection.
That just is not true. An FXS/FXO setup is merely one particular
type of "relayed" connection. It doesn't specifically apply to
the OP's scenario because the OP has no need to convert a wireline
loop interface into something that will emulate one for the distant
end.
There is some similarity though, and if you remove the "only
reasonable" bit, it does make a good example of what is needed:
2-wire 4-wire 2-wire
+-----+ wireline +-----+ carrier +-----+ wireline +--------+
| Tel | loop | |--->>>---| | loop | telco |
| Set |-----//---| FXS | | FXO |----//----| switch |
| | | |---<<<---| | | |
+-----+ +-----+ +-----+ +--------+
The reason for this is because the "4-wire carrier" cannot
transmit DC pulses or loop current. It might be a microwave or
a fiber optic system, for example, and the telset may be located
hundreds of miles from the telco switch. So the DC pulses and
loop current are converted to some other form of signaling.
With analog systems it was almost always SF (2600 Hz "Single
Frequency"). With digital systems it almost always uses Robbed
Bit Signaling (the AB bits in a the SuperFrame of a DS1 rate
digital line).
The similarity with the OP's scenario is that if we chop off the
entire left hand side, and provide something like this,
4-wire 2-wire
+------------+ +-----+ wireline +--------+
| computer |---->>>----| | loop | telco |
| voice card | | FXO |----//----| switch |
| |----<<<----| | | |
+------------+ +-----+ +--------+
We have the same thing, except the computer has to synthesize
whatever it is that the FXO expects to see on it's 4-wire voice
channel side for signaling and supervision. That isn't
technically difficult, and there no doubt are off the self units
where the FXO card plugs right into a PCI slot in the computer.
However, another way to to it is to use out of band signaling,
and a voice modem,
4-wire 2-wire
+------------+ +-----+ wireline +--------+
| computer |---->>>----| | loop | telco |
| voice card | |modem|----//----| switch |
| |----<<<----| | | |
+------------+ +-----+ +--------+
| |
| RS-232 supervision |
+---------------------+
That might seem odd to some, but consider this very common
telco equipment configuration for a FXS/FXO,
2-wire 4-wire 4-wire
+-----+ wireline +-----+ carrier +-----+ trunk +--------+
| Tel | loop | |--->>>---| |----->>>----| telco |
| Set |-----//---| FXS | | FXO | | switch |
| | | |---<<<---| |-----<<<----| |
+-----+ +-----+ +-----+ +--------+
| |
| E&M DC Supervision |
+--------------------+
The difference is that the telco interfaces to a 4-wire _trunk_
card rather than to a 2-wire _line_ card.
Clearly the OP can make a choice between in-band and out-of-band
signaling, and that amounts to choosing which hardware to buy.
>Plus these are the kind of cards OP will most likely see
>referenced. Have not kept up over the past 5 years, so if you know of
>something better I would be interested to hear about it. A link?
I'm sure a search on google would turn up a list of interesting
links. It is also possible that posting a query to
comp.dcom.modems and crossposting it to alt.dcom.telecom and
comp.dcom.telecom.tech might be productive.
>> >boring details (and their many variations) -- just know that modems
>> >cannot utilize analog _voice_ facilities.
>>
>> Bullshit. That is exactly what they are designed to do. And
>> it's been a *long* time since the dialer (ACU) and the
>> modulator/demodulator portions of a modem were separate units.
>
>Splitting tech (short)hairs -- and who said _anything_ about any kind
>of separate units. (Though I did toy with the idea of "remembering"
>the first modems that were cradles _for_ a handset!)
See the above discussion.
>If you can come up with a better phrase to indicate what the average
>joe thinks of when when speaking about "voice" and "phones" I'm more
>than willing to accept -- never cared for this (my) phrase used here,
>but wasn't going to spend more than ten mins thinking of a better
Perhaps, but that is not the point. The point is that the modem has
both mod/demod capability *and* supervision capability. Using only
the supervision capability isn't exactly unheard of.
>expression. Just call me slow witted ;-) BTW, don't try your method
>with a trunked PBX -- they acquire dial tones in a differnet manner.
Oh, whoop dee do. Some (but not all) PBX's use "ground start"
rather than "loop start" on wireline trunks. Big deal. We
aren't talking about ground start lines (and there *are* both
telephones and modems that can use ground start too). Moreover,
in addition to those two interfaces which are commonly used for
lines, there is the myriad list of trunk interfaces too, which
many PBX's also use. Who cares, since that was not what the OP
specified, nor is it necessary for the OP to even consider
anything other than interfacing to a standard POTS line (unless
of course he arbitrarily wants to).
>> >Faxes are another hardware interface that uses the _circuit_
>> >facilities of POTS. Faxes _pre-date_ computers -- sometime in the
>> >1930's IIRC and using the grandfarther of today's hi-res drum
>> >scanners. That's why the term "fax/modem" -- a modem is not a fax
>> >interface device without extra hardware.
>>
>> A fax includes a modem. A modem isn't a "FAX" until you add
>> a printer...
>
>Well, here I'll say quite definitively that you are simply too "wired
>into" current technology packages and that about faxes and fax
>machines (old and current) you have this "backwards" -- have installed
>100's of fax machines _and_ cards that were modem-less.
You have *never* installed a fax machine that did not have a modem
in it!
>They are
>different technologies both using the "phone lines". Neither one
>_requires_ the other, despite their "similar" use of the phone lines.
That is absolutely wrong. *Every* FAX by definition has a modem
as an essential part, by definition. That modem may not use
v.32, v.34, v.90, or v.92 modem protocols, but then again it
might not use v.26, v.27 v.29 or v.33 modem protocols either
(and neither would the modem *you* use). Regardless, those are
all "modem protocols", and FAX modems, for example, use v.29
modem protocols for 4800, 7200, and 9600 bps transmissions.
FAX modems might also use v.17, v.27 or v.34 modem protocols.
>They _can_ share certain abilities/requirements, like acquiring a dial
>tone ;-) For establishing a session and transmitting data they are
>completely different.
Bullshit. FAX machines use a modem to 1) acquire a dial tone,
2) establish a session, and 3) transmit data. That might even
be done with *exactly* the same hardware and modem protocol
(v.34) that is used for computer data. In fact, any high speed
FAX machine that works at a speed greater than 14.4Kbps is going
to use v.34 protocols. The slower protocols are still more
common, but they are different only in detail, not in
functionality or overall purpose from the various modem
protocols used for computer data.
>> >"Voice" modems have minimal ability to process a voice signal --
>> >namely record or playback a "voice stream". With software they can be
>> >used to build suprisingly sophisticated (and large) voice mail
>> >systems. In both cases this is a dial-in capability -- not dial-out.
>> >Well, not entirely true, as these modems could be used for messaging
>> >services, like _sending_ pager text.
>>
>> I'm not really familiar with voice modems, as I've never owned
>> or used one. But what you just said is in effect that a voice
>> modem will do *exactly* what the OP wants!
>
>One would think so. The dial-in capability (voice mail) requires
>_both_ hardware _and_ software additions to a basic modem -- and if
>"extended" you might think this would allow a modem to act as a
>handset. (Not sure why this was never carried out -- technical, $,
>regulations, opposition, public acceptance?) The dial-out capability
>was a means of "plugging into" services provided by the phone company.
> Without the CO facilities it would not work -- nothing to work with.
Well, gee... a POTS line that won't work without a CO is not exactly
a revelation! ;-) Your regular telephone set requires that too!
I'm not sure what the significance is that it requires more
software and hardware than just a modem. So what? Computer
data transfer requires more software too. It also requires
significant hardware, and the only difference is that the
hardware required is so universally needed that it is provided
by default. But a serial port is *not* necessary for a computer
to function otherwise...
>> >For all this, modems still cannot establish and carry on a full duplex
>> >_voice_ connection like a simple handset! But then they weren't
>> >designed to do so.
>>
>> I don't see why not! As I said, I've never used a voice modem,
>> so I'm not sure they actually have the ability to do what you
>> described above. But if they do, then clearly they can carry on
>> a full duplex voice connection. (Incidentally, a "simple
>> handset" cannot! It requires a slightly more complex "telset",
>> which means there has to be a hybrid network in order to have
>> full duplex.)
>
>You misunderstand "handset" -- I'm speaking of the RJ11 connected
>phone most "hard wired" phone users pick up to dial out or answer a
>call.
A handset plugs into a telset. The telset (not a handset) plugs
into the telephone line.
Typical residential users rarely see either a handset or a
telset without seeing the other, and since the introduction of
the "Princess" telephone way back when it is not uncommon (e.g.,
cell phones) to see the two devices combined as a single unit.
However, they are in fact two very separate items. Commonly the
handset on many telsets has a cord that can be replaced (with a
longer one, for example). But in some places you'll find the
two are *very* distinct. Operator switchboards, for example.
Each operator has a *personal* handset, that nobody else ever
uses. They sit down and plug it into a telset, which is perhaps
one of several, but all of them would commonly be used by
whoever happens to be physically where that telset is located.
Your misuse of the terminology is not a reflection on my
understanding of it... ;-)
>> I see no reason at all that you cannot arrange to dial your
>> neighbor and chatter away. The modem doesn't necessarily *have*
>> to produce carrier tones! In fact, you might notice that it
>> dials up a connection and if you tell the modem to dial your
>> neighbor and pick up your an extension line, you can talk to the
>> neighbor just fine and the modem will not make a peep.
>
>You _have_ to pick up your "handset" to chatter away _because_ a modem
>can't provide that capability. In fact, early modems wouldn't even
That is simply not true. Voice modems provide exactly that
functionality.
>allow you to "interrupt" a modem session (they wouldn't "release")
>this way. Last time I actually used this "method" of dialing out, you
>also had to pick up the handset within a limited amount of time. And
>it won't work with most (any?) modems without a pass-through jack.
But *clearly* you are being limited by your ability to configure
the software. The hardware is quite capable of doing exactly as
described, and there is a myriad of software available to use
that functionality in a variety of ways.
>> If the "voice modem" has the ability to send and receive audio
>> over the phone line (which means the modem must have both an
>> input and an output jack for audio), ...
>
>Nope, it "looks" like just a regular plug-in or serial modem -- no
>separate "audio" cords -- though some provide external jacks and some
>require/use your sound card. When someone calls in to _leave_ a voice
>mail, the handset they are using is effectivey the "mic" -- in fact,
>many "voice modems" require you to use the pass-through handset as a
>mic to record "intro" or "direction" messages for the voice mail
>system. (All the ones that I used on-site were this way.)
You are contradicting yourself. A non-voice modem does not
connect to the sound card. A "voice modem" does, and uses the
"some provide external jacks and some require/use your sound
card" functionality. If someone calls to leave a message, the
audio is passed from the telephone line to the sound card input,
and the sound card converts it to PCM or whatever type of
digital encoding and stores it on disk. The outgoing message to
the calling party was previously stored on disk, and the sound
card converts it from PCM, or whatever, and outputs that on the
"line out" jack that connects to the "line in" on the modem and
that is sent to the telephone line.
Precisely what the OP wants to do, except he wants real time
digital i/o rather than using store and forward audio from a
disk file.
>> ... then it is definitely
>> possible to control the modem via the rs-232 connection and use
>> the computer's sound card(s) for the audio interface.
>
>This was (is) done sometimes to a limited extent but due to the
>proprietary nature of most (all?) sound cards at the time (circa
>'93-95) it was "less that reliable".
Eh? You've been saying it can't be done, and now point out that
it has been being done for a decade. And I won't agree at all
with the idea that is was seriously limited by the '93-95 time
frame, though I'd say that was indeed true a decade before that.
Of course, we are talking about a technology that has been
commercially available for 20 years now! It may not be fully
mature, but this isn't exactly bleeding edge stuff any more!
>> ... Note too
>> that if the modem cannot do that, it is technically blindingly
>> simple to provide that functionality (it is basically the same
>> thing as a "phone patch" used by CB or Ham operators). ...
>
>And as above, I'm not sure why it was never followed up.
Lots of call centers use exactly that functionality, so it
really isn't correct to say it was never followed up on. It
just doesn't have much use in the consumer modem market, that's
all.
>> ... The modem can be used for dialing and for hook-switch control.
>
>Unfortunately, that's about all a modem can be used for at this level.
> Beyond this and the byte stream encoding/decoding (mo - dem) starts
>up.
No it doesn't. There is no requirement that the modulator be
enabled at all.
And supervision is not the only potential use either. Virtually
every modem contains a very significant capability for digital
signal processing, and every one of them implements an echo
cancellor. That has benefits for voice calls too. Likewise it
would take very little additional programming to allow the modem
to do just about anything you'd like to the telephone line (from
test signals to voice recognition). All that's needed is a
*reason* for someone to add extra functionality to the modem.
In many cases it would require no extra hardware, and could be
added by merely flashing the modem's memory.
>> Of course, when all of this is put together what it amounts to
>> is an expensive, hard to maintain, piece of junk. And it is
>> going to _sound_ exactly like what it is...
>
>Which is why I suggested that OP look at providers that offer internet
>VoIP services or internet phone services -- whichever he is most
>interested in.
I don't see much difference... VoIP hasn't quite gotten to the
point where it is much more than "an expensive, hard to
maintain, piece of junk" itself! However, it is right on the
verge... and when it really does arrive it *will* cause an
entire paradigm shift that will forever change the telecom voice
message business.
>BTW, I hope the folks around here that need some RS232 advice (among
>other things) know what a resource you provide. I did _one_ measly
>and "easy" project collecting data from a bar code scanner that
>downloaded library spine codes. On top of that, you regularly watch
>the posts and provide accurate help.
You sure make it hard to pick on you. If you'd just be a
"normal" Usenet guy, and snarl a little, we could call each
other names and all kinds of things, eh?
(And I really do appreciate that you are NOT a "normal Usenet
guy"! And while you may be saying a few things that are quite
technically correct, that is of little significance compared to
the interesting information that the resulting discussion
produces. I may know a bunch of technical details, but you are
the one with enough imagination to make this all interesting!)
--
FloydL. Davidson <http://web.newsguy.com/floyd_davidson>
Ukpeagvik (Barrow, Alaska) floyd@barrow.com