Virtual Modem for VoIP

Yes, of course. A modem modulates and demodulates data between analog and digital representations and an ATA...uh...wait...tell me the difference again?

--kyler

I for one will not argue with the kook, ok?

(I still have one that also does 110 baud...)

That's nothing, I used to have a Creed 444 on 50 baud RTTY :-)

Ivor

The DSP would be the source (or emulated modem).

--through VoIP-- Talking to a real modem on the destination.

A Shiva netmodem would be great as an OUTDIAL device hooked to your PBX with a TELNET port into it.

however...

The problem is that I want to use a program to call out using my existing soundcard and say connect to a remote system. Be it a ISP indial, a VPN indial, a BBS indial, a FAX machine indial or even reverse it so that when someone calls my VoIP POTS number, it connects to my modem and provides modem/fax telephony without the hardware hit (meaning an ATA).

Tests have proven that VoIP can tolerate speeds up to 14.4 better than

80%. Technically I don't require that speed. 2400 is more than sufficient.

MWave and DSP's are the shortest route to the solution just because they were able to emulate the sounds and belches that a modem makes through software. A modern day SoundBlaster Audigy (or similar) has

100x the DSP computing capability than that DSP had.

Heh... Personally I'm thinking in terms of backwards compatibility but alright...

Compared to POTS telephone lines, seems pretty fair to me.

Plus it would give more incentive to digitally minded people to setup Asterisk servers and link their telephone lines up.

Unfortunately, my cash poor ass can't set one up right now. However, if I do, I'll do it the right way and hook up a cell phone base to the FXO so it can call out on my "any-time" minutes or unlimited nights and weekends. :)

Too bad the pessimistic has reigned.

Do you understand how much loss there is in that extra D/A and A/D? It's just not worth it.

Not much of the world is interested in highly erratic 9600bps data service, though.

There were companies doing what you propose in the past (though they did it the right way, by doing the modulation/demodulation at the phone-company-interface end, not at the originating end). To the best of my knowledge they have all shut down due to lack of consumer interest.

There are already endless fax delivery services, again, all of them doing it the smart way, not the backwards and inefficient way you are proposing.

There's even a free one with decent coverage:

miguel

Where are you calling? Minutes in the USA are about 1 cent. How much effort will you put into this, and how could it possibly be worthwhile?

Let's say you put 40 hours into sorting your scheme out, and you value your time at $50/hour. In order to break even you'd need to use 200,000 minutes of connectivity time. That's 4 hours a day, 7 days a week, for over two years. It's just a losing proposition from start to finish.

miguel

To do what with?

These are much easier problems to solve, and if people cared, they would have.

I really don't understand who would want to.

To be honest, the only people I have ever heard of using dialup modems for anything other than the internet in this day and age fall into a few categories, none of whom would be better served by a low-speed, high-maintenance-overhead national outdial network:

1) alarm services - almost always local calls. 2) devices like TiVO - they contract with indial networks so they're local calls for almost everyone. Surely the reach of those networks is broader than anything you could create given the limited demand for an inferior version with low data rates and bad connection quality. 3) telesensing - usually local calls, otherwise the tiny cost of long-distance is easily justified, and in any case reliability is more important than cost. 4) fax - sending a fax long-distance costs about 5 cents per page or less. Anyone who sends in enough volume to care will have contracted with a nationwide fax delivery service.

You don't need one ATA per extension (if you're talking about extensions in the sense that people normally think about in a home), just one per line.

You can get a free fax number from efax.com.

miguel

And there's no compelling argument for them to sacrifice security, reliability, and speed for the sake of saving 1 cent per transaction.

You're talking to the wrong guy. I've got a collection of acoustic couplers going back to when they were the size of a big-city yellow pages. I fondly remember the days before CRTs were affordable, when upgrading from a round-key teletype to a DECwriter was like stepping onto the set of Star Trek. I use text-only tools for mail and news (wouldn't catch me posting from MSN) and you couldn't pry me away from the command line for 90% of the tasks I do.

But I do recognize when something changes for the better, and fungible data transit over IP is so much better than point-to-point D-A-D in so many ways for almost everything that it's not even funny.

It's anathema to reliability. You are adding extra layers of translation, conversion, bits of hardware, delay, and entropy, for no conceivable gain. What starts as data should stay as data. To convert it to noise, convert that noise back to bits, convert those bits back to noise, and convert that noise back to bits one last time is the height of pointless friction.

miguel

Considering that the data is in fact digital going between MARK and SPACE in a modulated dance of data within your internal system (no noise) into a SIP phone (no noise again) to an outgoing VSP where upon it might gain line noise on the POTS end, seems pretty digital to me all the way through.

IP has WAY too much overhead. Nevermind layers, this is all LAYER 3 stuff anyways, the audio packet rides on top of an UDP packet therefore may introduce noise from dropped packets it still seems to me that there is a viable need for it.

Doesn't matter if you came from the day of accoustic couplers (believe me, I've had mine as well), the most you probably did was talk to the university's PDP. I'm talking about real users from the home computer revolution that connected to BBS's and the like to transfer information. Oh sure, there were biffed characters in the data stream. That's why protocols for file transfer were created (ala XMODEM, KERMIT and Zmodem). But the fact is that these systems were able to carry data long before IP was commonplace.

I remember when I was in college, all three of our major universities were tied into a single 56k data line. 10 years later, all three of the universities are now OC-48. Sure technology changes, but that's no excuse for wantonist and wasteful use of data resources. I see better applications for bandwidth like IP-Television. ;)

At this point we'll just agree to disagree. I still see quite the viable need for an Analogous Virtual Modem that travels over VoIP. When the architecture is truly unobtrusive and ubiquous, when my fridge can stay stocked by itself via online shopping, when I can connect to any system in the world (on the Internet or not) without changing transport medium, then it's a new world. Till then, we need to keep the wrenches turning.

Ah.... yet another application.

We cannot forget about TDD protocols :P At least this one is already existing!

[snip]

I'm not sure I want my fridge doing my shopping for me..!

Ivor

On Mon, 14 Mar 2005 22:16:31 -0000, "Ivor Jones" put finger to keyboard and composed:

Sometimes low tech is faster than high tech.

Wi-Fly and TCP (Transmission by Carrier Pigeons):

- Franc Zabkar