News:

"The phone is a remarkably complex, simple device,
and very rarely ever needs repairs, once you fix them." - Dan/Panther

Main Menu

Contentious and rambling Analog multiplexing for connecting PABX to Central Office discussion

Started by bellsystem, June 26, 2017, 05:18:03 PM

Previous topic - Next topic

Dominic_ContempraPhones

Quote from: bellsystem on June 28, 2017, 10:43:10 AM
So far, analog has proven to be far more reliable in disasters and emergencies than digital.

Digital 5ESS and DMS are way more reliable.  5E-XC is 6 9s and DMS-500 can do anything.

Dominic_ContempraPhones

Quote from: bellsystem on June 26, 2017, 05:18:03 PM
I asked the following question on Stack Exchange but was told it was off-topic: https://networkengineering.stackexchange.com/questions/42193/pbx-co-trunks-analog-multiplexing-analog-version-of-pri

It's sort of technical in nature which is I presume why it was put on hold.

Does anyone know if they can maintain an analog connection between the central office and their house/business using just 1 cable, like PRI,  but with analog channels instead of digital? Is there a name for this type of technology to distinguish it from PRI?

I want the call quality and reliability of having as many separate landlines coming into the building as I needed outside lines, but the convenience that one large cable coming in offers - in addition to one cable, Direct Inward Dialing and Direct Outward Dialing are my primary concerns.

Nobody on the site I posted to was helpful, so thanks in advance to anyone who can help me out with this!


===========================================
EDIT: text copied from "other site" and pasted here for posterity


"PBX CO Trunks: Analog multiplexing (analog version of PRI)? [on hold]

I've been doing more research into PBXs, specifically trunk connection methods. My understanding is that there are 3 primary options:

    Separate telephone lines for each trunk line
    Primary Rate Interface (PRI)
    Session Initiation Protocol (SIP)

I'm trying to figure out what the best way would be to connect an analog-only PBX to the central office. I've already ruled SIP/VoIP out as being considerably inferior, from a quality and reliability perspective especially. PRI is preferable to SIP, but my understanding is that PRI uses TDM/digital transmission which would not be acceptable.

- - - - - - - - - - - - - - - - - - - - - - - - - -
Premise:

    Transmission must be analog, not digital, for call quality and reliability - the call quality should be basically the same as that with a separate standalone landline connection.

    Digits should be dialed immediately. I know many systems "hold" digits and then analyze the digits dialed using a "dial plan". Internally, this might be of some used. But if "9" is dialed for an outside line, I want the PBX to grab an outside line for the station and step out of the picture. The digits should be sent to the central office then as they are dialed, not all at the end (again, as with a separate, standalone landline connection).

    Going along with my second point, all central office connections, whether they are individual cables or one large one with separate analog channels, would have to be identical. Because the circuit will be grabbed as soon as "9" is dialed, all "features" (i.e. ability to make Long Distance/International calls) would have to be the same for each line/channel, since it would be impossible for the actual intended number to be analyzed. Basically, the call should be being routed as it is being dialed, not after.

    There's extreme controversy it seems regarding 9-1-1 and 9 9-1-1 going around (the linked petition will be impossible to force for systems such as the one I desire). I know that 9-1-1 will not be doing anything because to the central office, it looks like 1-1 has been dialed which could have easily been a switchhook mishap (I assume this is why 1-1 is not rerouted to 9-1-1). The PBX will be in a home environment (not a public system) and since I, most of the time, will be its only user, I will know to dial 9911 instead of 911 anyways. 911 should NOT connect to anything.

    I don't believe nesting PBXs will be a problem. I am going to buy a PBX that will support at least 50 stations, but one of the "stations" will end up being a corded switchboard PBX (PMBX or Private Manual Branch Exchange), and other "stations" may be other sub-PBXs that are PABXs, like the main one. Is there any way to use features like Direct Inward Dial from telephone stations to PBXs nested further down in the tree?
- - - - - - - - - - - - - - - - - - - - - - - - - -

I'm trying to figure out which networking technologies would best meet most, if not all, of my requirements. My guess is it would be some technology that lies between separate regular landlines and PRI, if there is one. The PBXs that will actually be used are arbitrary and irrelevant, but they will all be analog and have no digital support.

I believe something along the lines of an early 1970s analog PABX would end up being used (as is described here), but I am interested in what transmission technologies concerning the trunk lines should be used.

Is there a sort of medium between separate analog lines and PRI? I'm looking for a medium that will retain an analog transmission between the PBX and the Central Office, but the switching technology itself can obviously be digital (as all PABXs are). Is there any way to maintain an analog connection between the central office and the PBX without running separate regular landlines for each trunk? Analog will be needed to support good call quality for regular analog rotary and pushbutton phones, fax machines, and modems (including dial-up). I want whatever trunk technology is used to function like a standalone line without actually being 4 or 6 separate standalone lines. Costs aside, my primary concern with having standalone lines will be direct inward dialing as well as direct outward dialing. Even if there are 50 stations, there are only a few numbers I'd want to have DID and DOD for - otherwise, I want only ONE telephone number - and NOT as many numbers as there are trunks (this might be appropriate for a Key Telephone System but not a PBX), since I want one number to be used for Caller ID from any non DOD phone, which I don't think standalone lines would support.

If my internal extensions range from 1000 to 1500, the numbers I need supported for DID/DOD would be 1000 as well as a few other numbers between 1000 and 1200. I don't want to have to purchase a whole block of 200 "numbers" if that can be avoided (unless there is no additional cost).

I am estimating between 4 and 6 trunks will be required. Costs are not a concern, although since this will be installed in a home environment with 1 bill-payer, ideally they should be low.

Summary:

    No digital transmission (i.e. packet switching)
    Analog multiplexing is possible so multiplexing is fine if each individual circuit is analog.
    Routing as calls are dialed, not after
    Compatibility with DID and DOD

CLARIFICATION: I'm not asking anything about a PBX at all. I'm inquiring about the protocols used, specifically PRI, and if an analog variant of PRI (possibly a predecessor to PRI) exists."


I'm sorry, but you've misinformed people here, and seem to have a basic lack of understanding of DID, DOD, progressive vs. common control, TDM vs. packetized.

There is no analog PRI.  PBX trunks (analog) ... there are a zillion ways.  I imagine you want DP inpulsing. and not MF or DTMF.  There's ADID (inward), E&M.  Read Bellcore.  When I worked on DMS there lots of ways to do it, but PRI was better -- no glare, very quick setup and teardown.  PRIs are expensive.  If your PBX is analog only trunkside, what is it?  You can have analog lineside and PRI trunkside.

I have analog rotary/TT, TCM digital, UNIStim IP, ISDN stimulus mode lineside, and analog loop start, analog DID, PRI, SIP, and H323 trunkside.

A CO that's digital is not packet switched, it's time switched.  Your best trunk under real world conditions is PRI.  Forget about analog trunkside if you can get a PRI -- it supports everything.  Analog lineside is fine.

Switches are not progressive anymore -- translation.  It's transparent to your ear.

bellsystem

I'm not confused.

As long as the distortion introduced into analog transmissions is less than the loss of quality that is inherently going to be produced when a digital sampling is taken of sound, which is INHERENTLY ANALOG... then analog will be superior.

Sounds like everyone here is dissing analog because of DISTORTION and DISTORTION only.

Maybe someone in the PSTN should consider using SHIELDED cabling to minimize distortion. Shielded analog would definitely be better than digital.

As it stands, digital was introduced for long-distance mostly because of cost. It was so much cheaper than to use microwave radio relay and digital cabling for long-distance calls so one cable could have many calls going across it.

If we ignore cost and focus on the fundamental qualities of the two transmission methods, sound is NATURALLY ANALOG and will be closest to its purest form if KEPT analog. When sound is converted to digital and then back to analog, the actual audio/sounds that were produced at the other end are NOT reproduced accurately.

unbeldi

Quote from: bellsystem on June 28, 2017, 05:31:32 PM
I'm not confused.

As long as the distortion introduced into analog transmissions is less than the loss of quality that is inherently going to be produced when a digital sampling is taken of sound, which is INHERENTLY ANALOG... then analog will be superior.

Sounds like everyone here is dissing analog because of DISTORTION and DISTORTION only.

Maybe someone in the PSTN should consider using SHIELDED cabling to minimize distortion. Shielded analog would definitely be better than digital.

As it stands, digital was introduced for long-distance mostly because of cost. It was so much cheaper than to use microwave radio relay and digital cabling for long-distance calls so one cable could have many calls going across it.

If we ignore cost and focus on the fundamental qualities of the two transmission methods, sound is NATURALLY ANALOG and will be closest to its purest form if KEPT analog. When sound is converted to digital and then back to analog, the actual audio/sounds that were produced at the other end are NOT reproduced accurately.

This is rubbish.   

You might want to read the accounts of the technical problems that faced the engineers of the early transcontinental telephone lines, and study signal transmission theory, and the laws of digital information, of which the pioneers were also Bell System scientists.

Digital information is transmitted without alteration of information, otherwise facilities such as the World Wide Web via the Internet would hardly be possible.  Shannon's framework of digital information tells us that a signal can be completely and truthfully regenerated when it is sampled according to the sampling theorem.  This is implemented in digital telephony, which has a theoretical bandwidth of 4000 Hz.  But more recently, this has also been extended to provide much higher fidelity with HD voice.  In practice, even the more realistic band of ~3500 Hz pushes the capabilities of historical telephonic transmitters.  The response curves of various transducers are well known and available in Bell System records and other places.  Given theses limitations of the transducers, a telephone conversation cannot be transmitted with higher fidelity by analog means in practical circuits than with digital transmission.  It is the electro-mechanical transducers, receivers and transmitters, that limit the fidelity of reproduction, not digital transmission.  But digital transmission has other sources of distortion, but they can be fairly easily corrected, and is insignificant compared to analog signal distortions over distances typical in real-world situations.


bellsystem

Digital is not a truthful exact regeneration. By definition, as soon as something analog, like SOUND, is converted to DIGITAL, it is no longer the EXACT same sound wave (oh wait, it's not actually a sound wave at all).

In order for digital to be on par with analog when it comes to accuracy, distortions aside, the sampling rate would have to be INFINITY, which is, of course, impossible.
That's why analog is used. No such thing as sampling rates, since they're irrelevant. It's just a constant connection that literally transmits audio, not 1s and 0s of the audio.

Digital may sound better in the send in many cases but analog is a more ACCURATE reproduction of the original sound wave, since it IS that sound wave.

Dominic_ContempraPhones

Quote from: bellsystem on June 28, 2017, 05:31:32 PM
I'm not confused.

As long as the distortion introduced into analog transmissions is less than the loss of quality that is inherently going to be produced when a digital sampling is taken of sound, which is INHERENTLY ANALOG... then analog will be superior.

Sounds like everyone here is dissing analog because of DISTORTION and DISTORTION only.

Maybe someone in the PSTN should consider using SHIELDED cabling to minimize distortion. Shielded analog would definitely be better than digital.

As it stands, digital was introduced for long-distance mostly because of cost. It was so much cheaper than to use microwave radio relay and digital cabling for long-distance calls so one cable could have many calls going across it.

If we ignore cost and focus on the fundamental qualities of the two transmission methods, sound is NATURALLY ANALOG and will be closest to its purest form if KEPT analog. When sound is converted to digital and then back to analog, the actual audio/sounds that were produced at the other end are NOT reproduced accurately.

You are referring to quantization error, not distortion.  I've done it.

DMS and 5ESS use PCM mu-law.  So, in DMS, we take 8000 14-bit samples per second, but we quantize down to 8 based on an analysis of speech.  You are using a carbon microphone (OLD) that produces 20% distortion by default in the analog domain.  Compare that to electret microphones.  We did side by side tests on so many people in the 80s with ePhone.  Night and day.

In theory yes, but, all analog transmission will pick up noise, like a vinyl record, along copper.  Analog modulation on fiberoptic lines would be immune however.  The problem is how do you switch.  You can't without degrading the signal.  Digital is immune because of error correction and interpolation  There are only two discrete states.  And, on long distance, you had to YELL, LOUD, when the system was all analog.  It was bad.  The switches couldn't do anything.  VoIP I don't like unless it's on a voice-only LAN and carefully done.

Digital Class 5 DMS came about not because of noise but because those SxS machines kept breaking down and jamming with all those moving parts, and were a maintenance nightmare -- 5XB wasn't quite as bad, but still a pain.  Crosstalk, clacking relays -- come on.  Now, DMS and 5ESS come along and there are almost no moving parts, and then D evolved to become a single stage time switch -- non-blocking.  100% of the switching fabric could be used, vs. 10% on SxS.

All DMS did was read out numbers in a different order than they were read in, and the position determined the trunk, and so on, through a big glob of memory.  It was a sorting machine that blew people away.

We did a lot of work on analog.  The last mile is analog because of distance.   If you want all analog, you can do it.  I convert digital to analog, IP to analog, the other way around, but you need incoming only and outgoing only without E&M, and you don't have DISA DID.

The PSTN isn't even TDM anymore -- the periphery is, but the core is optical Ethernet.

Dennis Markham

Guys, this conversation is way over most of our heads.  It's certainly over mine.  This thread is not something that I would normally follow.  I don't have the time to read every thread on the forum.  I was alerted by a moderator that this was "heating up" so I have picked my way through it.  Let's keep it civil please.

You all feel free to continue.  But please keep personal remarks and other possible inflammatory comments off the forum.

Thank you.

~Dennis

bellsystem

I agree with Dennis,

Most of the replies I have been receiving are "technical" explanations about why digital is so great and analog sucks or personal attacks on my telephone preferences,

To be clear, what I want to know is purely academic. I never said I would plan to implement a system. I simply want to know if such a system exists, how it works, and how it would differ from PRI and standalone analog landlines.

Rather than inserting your own opinions into answers to the point where I can't distinguish facts from opinion, I would appreciate it if I could get a facts-only answers that just lays it on thick. If nobody knows, that's fine. I'm only half-sure of what I'm asking myself.

Also, to Dominick: So, in DMS, we take 8000 14-bit samples per second, but we quantize down to 8 based on an analysis of speech.
What if you're not sending speech? How about music? That's likely to lose some of its richness when sent digitally.

And if I may ask, if carbon transmitters are so primitive, then why do they sound better than pretty much any cell phone or VoIP phone? I haven't used a digital phone that isn't VoIP, so I can't compare. But I do own 7 telephones right now - 5 of which have carbon transmitters. The two that do not are relatively new and are electronic phones - an AT&T 100 and a Panasonic programming unit. The Panasonic phone sucks. Tons of static, totally unusable, except for Caller ID. The AT&T phone is so soft, you can't hear the dial tone unless you jam it into your head. The voice quality on one of my Western Electric 500s is ten or twenty times better, easily, than my non carbon-transmitter phones. Then again, these are all connected to Panasonic electronic modular switching systems, which are digital. But transmission is analog.

Also, calls on Panasonic PBXs ARE routed as they are dialed. If I dial 9, I hear another dial tone. On VoIP phones or probably most other PBXs today, VoIP or not, you don't hear another dial tone after dialing 9. If someone could explain how NOT routing the call as it is dialed is supposed to make the call route faster, that would be much appreciated. I have 2 PBXs tied to each other, rather than to any Central Office lines, and if I keep going back and forth again, the delay tends to get longer and longer (i.e. ringing, stop ringing). Not sure if that is related to this somehow.


Finally, can people have their own Central Office, either for collecting's sake or to do things the way they want? How would that all work, as far as connecting it with the rest of the PSTN?
If you have your own Central Office, can it be private? Or does it have to be available for public use?


Thanks all!

Victor Laszlo

The traditional common battery central office arrangement has been replaced, over the last 30 years or so, in rural areas that are becoming suburbanized, with pair-gain equipment, powered by commercial juice, with back-up batteries. There are probably fewer continuous copper lines in service right now in RBOC's as there are lines that are not directly powered by the battery in a central office.

AL_as_needed

To echo Dennis: this is a topic well over my understanding, Im simply a collector and not an engineer by any means....BUT... just my two cents from a simple mind....

I understand the debate between digital/landline (analog) systems regarding transmission etc, however I think we are missing a few key points (but hey, I could be wrong....)

Cell phones generally have many microphones arranged around the edge of the device to better pick up voice. Lets face it, these "smart phones" are less than ergonomic. A G-3 handset is self aligning for optimal voice transmission, i-phones, not so much. So the designers took the liberty to improve the odds in their favor. As a result you tend to pick up a lot of ambient sounds and the like. Calling is also almost seen as a secondary or tertiary function of these devices. "Browsing", imaging, and apps are likely the largest responsibility of a modern smart phone vs calling (ask a high-schooler how much they actually call their friends vs snapchat or other apps). So a network to handle all that sort of data simply needs to function differently than the days of Ma Bell.

Moving on.... transmit/receive characteristics for cell vs a landline aside, at some point they all end up hitting the airwaves (AT&T Long Lines?) and or a satellite blinking overhead. IMO a purely copper system has not existed for some time, especially concerning long distance calls. These moves to digital (cordless) have been in place since the 30s with "portable" radio phones.

You can create your own network however, its more of a PBX system though in all actuality. I wont even try to explain all that, there are experts in that dept all over CRPF. Additionally you can also look into C-net. Yes, it uses the internet as a means of transmission, but it comes fairly close to replicating an analog system in terms of the switching gear and so on (could also be wrong about that...)

Landlines and payphones are dying off quickly and it is sad to see. However (and it pains me to see it), such was the fate of steam locomotives, VHS, etc. Money drives the world, faster, cheaper, "better" is always going to be out the old reliable vanguards of the prior generations.

Well thats my rant.... Im going to go hug my WE 302 thats on my desk to cheer myself up....
 
TWinbrook7

Dominic_ContempraPhones

BellSystem,

The PSTN was never designed to transmit CD quality audio when it was analog.

If DMS will accept a candlestick with no dial and you can just say who you want, why would you not like that?  It could any kind of ringing, was very forgiving on slow dials.  We had rotary phones that dialed at 20 pulses per second.  We could dial by saying "Sara, get me Doc Stewart".  This was speaker independent speech activated intelligent dialing.  Implicit SAID.  It would even remove dial tone and you would get that 1920 experience.  Your SxS can't do that.  We used to set it up so that we had every phone from a dial-less candlestick to rotary to touch tone to ISDN (EKTS) to Centrex EBS or ISDN to Centrex IP phones.  It was magical -- all these phones from different eras could call each other.  Toss in 1A2, sure.

What is the PSTN?

The PSTN was Aggregated Digital Time Division Multiplexing via
Synchronous Optical Networking (SONET - Optical Carriers OC-3 thru OC-192), and later
Asynchronous Transfer Mode (ATM)

Now, the core is Optical Ethernet (Dense Wave Division Multiplexing) while the periphery is still TDM in many places.

5ESS and DMS are TDM.  4ESS down to 1A and 1ESS no longer exist, sadly.  DMS took on the role of 4ESS.

5ESS, 5E-XC, DMS-10, 100, 200, 250, 300, 500, CS1500, CS2000, CS2100 (Military) and all the remotes 10 years ago were there.  It's changed quite a bit.

Now it's mostly MPLS/IP.  Everything keeps changing.

Didn't you know?  It's not analog except on the last mile and it changed from circuit to packet gradually over the last 10 years.

DMS has DWS (Dialable Wideband Service).  Others call it Switched Fractional or Multi-Rate ISDN.

This allows us to take a PRI which is 24 channels, each channel capable of 64 kilobits/sec,
and bonding 2 to 24 of them for any call.

Two calls using 12 bonded channels would do it.  How?

A compact disc stereo channel is 44,100 samples per second x 16 bits/sample = 705,600 bits per second

1/2 a PRI (12 channels) x 64 x 1024 = 786,432 bits per second will carry STEREO LEFT
1/2 a PRI (12 channels) x 64 x 1024 = 786,432 bits per second will carry STEREO RIGHT

Each call only needs 705,600 bits per second of the 786,432 available.

Analog systems cannot transmit with this fidelity.  You couldn't do this with an ordinary telephone.  Special equipment is required, similar to the stuff for video broadcast over the PSTN.  Video used this all the time before 1MMS came out.  It's wasn't packet -- it's circuit.  You dialed a phone number, but we had special equipment for it.

DMS is carrying the data RAW in DWS, so CPEs on either end handle the rest, because a telephone's microphone cannot transmit frequencies as high as 20,000 Hz.

Carbon microphones produce distortion -- we measured it at 20%, whereas electret was almost nothing.  This is the actual telephone.  Then we put the digital converters in the phones, and we used something called TCM ping pong.  Each caller transmits at twice the normal the speed but only half the time.

There's no problem with SONET or ATM or T1 or whatever.  I don't understand what your concern is, because SONET and ATM are dead, and have been replaced by Dense Wave Division Muliplexing Optical Ethernet.

Are you not aware the PSTN is fiber -- my brother and I can show you.  It's all optical trunking between switches.  You have copper twisted pair going in, and that can be analog or T1, or PRI, or ISDN BRI.  There are packet gateways too, cellular as well.

AT&T's backbone was 40 gigabit IP/MPLS several years ago.  Dunno what it is now.

If you want your own circuit-switched central office with battery, I would suggest an Option 11 PBX with analog line cards and it has battery backup, will accept rotary dial, digitone, Autovon, ISDN ksets, and Meridian phones.  It will satisfy your needs.  Very nice small switch.  You can do 7 digit dialing, and you can use the Universal Analog Trunk Card, although I recommend the PRI.

You'll have fun.  I did, because the like the old and the new.  I don't like what's happening now, because our Bible was not to desupport rotary.  We still had 10 party service coding, 4 party, 2 party ... any scenario you wanted was possible.

I can't think of any analog electromechanical switches that can act like a CO that aren't huge.  SF 1 is still big.  Long gone.  There you have it.

I'm having fun with you but everything above is true.  You've just been misinformed.

If people don't understand this, maybe they should learn, or complain to the FCC and CRTC.

TDM was fully backward compatible and even offered 10 party line service.  Geez, you could use a candlestick phone with no dial on it.  How can you not like that?

DMS would even take away the dial tone, you just had to say "Sara, get me Doc. Stewart".  You'd have to record that as voice print in advance.  We called it SAID.  The cutest thing.

It was a modern machine that supported every phone ever built.  I loved that thing

Dominic_ContempraPhones

BellSystem ... listen I don't want to argue but when we switched to electret people were used to the distortion of carbon and it sounded weird to them.  It's not richer, trust me.  It's what you're used to, and analog created a lot of problems -- crosstalk, and it didn't do well over long distances.

I understand why you feel the way you do, but I don't like VoIP, because it is inferior, but digital is not at that sampling because we only had 16 bit chips back then, and mics where #### anyway.  The processor was only 36 MHz.  We went electret in 1983 ... yeah, it was then.

bellsystem

QuoteIf you want your own circuit-switched central office with battery, I would suggest an Option 11 PBX with analog line cards and it has battery backup, will accept rotary dial, digitone, Autovon, ISDN ksets, and Meridian phones.  It will satisfy your needs.  Very nice small switch.  You can do 7 digit dialing, and you can use the Universal Analog Trunk Card, although I recommend the PRI.

So what's the difference between Universal Analog Trunk Card and PRI as far as features?

Can Analog support Direct Inward Dialing? I know PRI can.

I've seen a video of a guy who has his own small SxS central office. Can I just build my own central offices, connect my PBX to it however the heck I want to, and then connect the Central Office to other Central Offices? Could it be a private CO, or would I have to connect subscribers who wish to be connected?

unbeldi

Quote from: bellsystem on June 28, 2017, 07:13:36 PM
Digital is not a truthful exact regeneration. By definition, as soon as something analog, like SOUND, is converted to DIGITAL, it is no longer the EXACT same sound wave (oh wait, it's not actually a sound wave at all).

Why don't you show us where your 'definition' is written.  Show us that Shannon's theory of information transmission is wrong, and backup your story with sources.

Shannon showed us that ALL communication is essentially digital. This is a fundamental precept of modern communication. It is something that nobody has argued much anymore for close to 70 years.   Your level of belief is that of the 1940s at best, and it is only belief, not science.  Without Shannon's recognitions, communication would have been at a standstill, without T1, TDM, fiber optics, all the technologies Dominic showered you with, and without the Internet.

Quote
In order for digital to be on par with analog when it comes to accuracy, distortions aside, the sampling rate would have to be INFINITY, which is, of course, impossible.
That's why analog is used. No such thing as sampling rates, since they're irrelevant. It's just a constant connection that literally transmits audio, not 1s and 0s of the audio.

Digital may sound better in the send in many cases but analog is a more ACCURATE reproduction of the original sound wave, since it IS that sound wave.
Isn't that a circular contradiction?  How can something sound better but be less accurate when it is the message that is important.

Analog transmission is in fact NOT used as a rule.  The fact that the last-mile network is still analog is purely historical and reflects more local economics and politics than science and technology. It certainly has nothing to do with quality, fidelity, or similar notions.  The last mile is about the best we can do with a modern level of quality in analog transmission.

EVERY communication system has to make compromises between physics, economics, and policy or politics.  Your ideas about infinite sampling rates to achieve analog quality is simply wrong, because no transmission channel exists that has infinite bandwidth, that can transmit messages without loss or distortion.  A perfect channel is exactly as impossible as an infinite sampling rate.

If that weren't the case, then we never needed telephony, and in 1915, Alexander Graham Bell in New York City could have simply spoken to Watson in San Francisco directly, without the first transcontinental line and no telephone.  But as we know, this method already fails across a busy mid-town NYC avenue. Despite human's considerable intelligence, the brain cannot discriminate against all that noise. The same happens in an analog communication channel, no matter what its bandwidth is.  In fact, we do better with less bandwidth because we know that the crucial message extends only from 300 to 3000 Hz, and predominantly even less. Thus we can simply sample that message at a rate of twice its highest frequency and truthfully reproduce it at the other end of a digital communication channel.


I think it would behoove you to take in the advice you get here on the Forum and take it a starting point for your own research and not contradict immediately with ill-founded notions.  You have already heard from people with exceptional expertise in telephony and its history.  Some of the members here have built and maintained central offices professionally, some have built entire offices in their spare time for fun, and many have equipment in their collections that rivals or exceeds any technology museum.  If this many people contradict your views in such a short time, you might want to rethink your positions rather than dig in, and do some studying on your own to familiarize yourself, or to find well-founded arguments against those contributions.  Telecommunications is an immense field and not everyone is expert in all areas, but contributions here are always intended to help, rarely for self-promotion.


Alex G. Bell

Quote from: unbeldi on June 29, 2017, 10:09:46 AMEVERY communication system has to make compromises between physics, economics, and policy or politics.  Your ideas about infinite sampling rates to achieve analog quality is simply wrong, because no transmission channel exists that has infinite bandwidth, that can transmit messages without loss or distortion.  A perfect channel is exactly as impossible as an infinite sampling rate.
OTOH, even if infinite sampling rates could be accomplished it would be pointless since human hearing is far from having infinite bandwidth either.  It is pointless to transmit signals which cannot be perceived at the receiving end.

Motion pictures or videos of which B-S is so fond are sampled data streams too but they are perceived as continuous motion as long as the frame rate exceeds to persistence of the eye and central nervous system.