So, I read the V.61 spec and you got it pretty much spot-on. It's hybrid analog, with a modulation that looks like it's a slightly faster V.32bis (either 2800 or 3000 symbols/sec, compared to V.32b's 1800). When you're not using the voice feature, it's supposed to be able to use most of that capacity to give you a 19.2kbps data connection (though I wouldn't be surprised if something made it effectively 14.4 for you). When you turn on the voice mode, it keeps the same symbol rate but it downshifts from QAM256 (8 bits per symbol, 19200 usable bps) to QAM4 (2 bits per symbol, 4800 usable bps). This leaves so much "extra space" in the constellation that the modem can add a (filtered, processed, and decreased in amplitude) version of your voice to the signal and the other modem will still be able to decode all of the data bits. After it does that, the receiving modem subtracts the value corresponding to the data bit from the signal, and what's left is the processed voice signal. So it undoes that processing, and some slightly crunchy voice comes out of the speaker. 19,200bps (data-only) / 4800bps (voice+data mode) only requires 2400 symbols/sec, but I said there were either 2800 or 3000, so where do the extra ones go? They're used for a data "control channel" between the two modems that always runs at the low rate (800bps) and constantly exchanges information about the line conditions so that the modems can choose the optimal values of the audio processing filter values, and it's also used to switch between the data-only and voice+data modes basically instantly (25ms).
I *think* that the reason for offering two different symbol rates (even though they both give you the same data rate) is that the higher symbol rate also allows slightly wider audio filters (better voice quality), but the lower symbol rate is more likely to work if your phone lines are crap. The standard never comes out and says that, though. Oh, and most people pronounce QAM Q-A-M, just like AM or FM.
Thank you for this excellent explainer of how we were able to squeeze data and voice through a narrow pipeline!! Now, my next thought, is how this will all degrade with the switch from analog phone lines to digital VOIP.
@@setoth1234 anything that meets the standards that the telcos keep for digital voice (8kHz PCM with minimal jitter and no "advanced" compression -- like a T1, or G.711 over good solid VoIP) will handle this just fine. The long-distance infrastructure was already largely digital in the 90s, so modem designers took account of that kind of thing. Sending it over GSM or AMR or G.729 or Spex or Opus or Silk will murder it, but that's true for basically any modem, it's not specific to SVD.
Yelling into the phone until the connection dropped not only made me laugh out loud but also reminded me of that video where a guy screams at a bunch of hard drives in a datacenter and the performance drops noticably. Thanks for the lesson on technology i never got a chance to use
i found whistling loudly down the phone was the quickest way to get the modem to drop the call. came in useful when my sister was hogging the family computer and i wanted to use it :D
@@Nuskrad you could have tried to emulate a line reconnect, I had about as much crap as CRD (still have some of it) and I have managed to 'negotiate - connect' handshake with a modem several times using my voice. Yes I know how wasting time on shit like that is pointless, but there was a time in my life when I had a bunch of time and tech junk around to have nothing better to do.
We totally had Roger Wilco 1998! You could use it to voice chat and play a multiplayer game like Starsiege over dialup. It worked like a charm over my 26.4 modem.
My dad actually ""solved"" the problem of using a phone line for both data and voice in his youth by installing 2 completely different phone lines in the computer room. How he actually convinced his parents to do this I have absolutely no idea.
As one of those who was a teenager in the time of dialup, I can assure you that after picking up a phone so many times just to hear the squeal of a modem, eventually it wasn't a hard sell to get a second line.
when the phone line is in use nearly 24/7 with dialup, sometimes people need to make phone calls.... ugh, i remember those days "get off the phone, i need to make a call" x.x
Late 80's I did my fair share of BBSing. Eventually the idea of sharing the line peacefully with a teenager was painful to envision. Best part was it let me operate a BBS since I could accept calls without ringing throughout the house. Good times, indeed.
My family used a program/service called CallWave back when we had dial-up. Basically, anyone calling your phone number would get redirected to this voicemail service. Through a program running on your computer, it would alert you to the incoming call, and would let you listen to the message while still online to decide if you wanted to disconnect and call them back.
Ah. Should have read the comments before mentioning that myself. It worked using the call forwarding feature of your phone line, and you could set that up by dialing a special code before each call, same as you could for disabling call waiting. And the service was free--it just used advertising to support it. The app was always on screen, showing some ads. And I'm not sure, but I think they eventually would play a very short ad on the actual call. You know the kind that just says "Brought to you by . ."
Hello all! I was a big modem user from the days of a manual dial 300bps modem to the last v.everythimg Courrier, and I recall most of my modems (if not all) used a relay to cut off the attached phone on the passthrough jack. So when I transferred files with friends, we'd just leave the phone off the hook for when the modem hung up, so we could continue chatting by voice. sending 800k Apple IIgs disk images across a 2400bps modem took a very long time... It's possible since I was a heavy modem user for so long I always had somewhat fancy and expensive modems, so it's why this feature was there. I got DSL at the start of 1999 so that was the end of the line for my modem usage.
Adding to this, even my original manual 300bps modem cut off the attached phone when you switched it to answer or originate. I'd dial a BBS, wait for the carrier and switch to originate at which point the phone would cut off. To hear the modem carriers you'd need to pick up another phone elsewhere in the house.
@@CathodeRayDude The USRobotics Courier v.everything modem has a voice/data physical button on the front panel to even switch between the two modes. I can't recall when I got my Courier but it was the last modem I got because it was a fully firmware upgradable modem that could be flashed to support new standards. IIRC when I got it, it only supported 28.8k but then I upgraded it to 33.6k and then finally 56k at the end. It also supported the proprietary HST protocol which was a USRobotics 14.4k (or was it 19.2k) thing while everyone else was stuck at 9600bps. Some BBS's used HST modems so it was a nice speed boost. The Courier was costly compared to other modems, but it was the bees-knees. I still have an unopened v.everything and the branding on it says "MISSION CRITICAL CONNECTIVITY" in big red lettering. Fun fact: the modems inside all mobiles phones (to this day) use the same AT command set for everything. It's just hidden away by the mobile phone operating system. This command set was standardized by Hayes in their first smart modems. (Or at least I think that's the case with the AT command set.)
That was definitely not common -- at least on later models of less-than-Courier-class modems. I had a generic 2400-baud modem that came with the family 386, which we upgraded to a Zoom 28.8, and I later had a US Robotics internal 56K modem. None of those would stop you from picking up a phone attached to the second jack during a data call. I do seem to remember they all (?) had relays, but AFAIK, that was just a disconnect between the phone network and the computer. (Probably to help minimize the likelihood of damage from nastiness on the phone line.) I have quite a collection of modems now. Some day I'll have to look into whether any of them did this. I never knew that was a thing, because none of mine ever did, so it hasn't occurred to me to consider this in more recent years. Interesting.
@@CathodeRayDude Another tidbit now I'm thinking about things -- with modems with a relay, when you called a number (with the modem) and someone answered, you couldn't just pick up and talk to them. So, with the hayes protocol, simply pushing any key on the keyboard after issuing a ATD will cause the modem to instantly hang up (or switch back to the attached phone if you a holding the handset off hook.) I'm skimming the Courier HST manual and it works this way, so it had a relay. I still have my old 2400bps modem from 1989 (a Zoom modem) and I just opened it up, and sure enough the two phone jacks are just connected right to each other. So I guess the cost cutting had begun already!?
Your laugh at the yelling to disrupt the carrier reminded me... I was a member on a MajorBBS with some number of lines; I forget how many at this particular time. But the sysop had some kind of flow control or disconnect detection issue between the BBS and the modems wherein a line could drop, but the modem would remain connected to the BBS. After a user was dropped, when the next caller would hit the line for that modem, everyone would see that ghosted user sending "RING" "RING" "RING" over and over until the caller gave up. To keep from having to manually jump on the system and disconnect the errant modem, the sysop added an action word to the chat room that would disconnect the user if it saw the word ring on a line by itself. My girlfriend spent significant effort tricking people into typing it and kicking themselves offline.
Oh my gosh, if it were me, and I knew what was happening, I would have started typing modem commands in chat to see if I could get it to dial out! Or also send ATA to have it answer and the new caller would be logged in to the old user's account...
@@tkteun Modems don't respond to commands when they're in "data" mode. You have to either use one of the RS-232 side signals to switch to command mode (like DTR. DSR, I forgot which one) or send "+++" without sending CR, then wait until you get a command response. In this case, your modem is in data mode, but the other one is in command mode.
I used to work for Packard Bell. Their 33.6 (and above) modem/soundcards supported voice view. We even had a radish team to do support for customers that had those mpdems
Packard Bell is the *only* name that seems to come up when I search for this tech, I suspect they leaned into it more than anyone else. I'm really curious if you heard much about people actually using it, and what they used it for?
I think I get this now. "I want a faster horse" when in reality it's "I need a second phone line". This is how my parents solved it too. They wanted to be on-call at work because that gave them both mobile phones for free. So they both had Nextel phones, and we used the Walkie talkie feature all the time. This way, I could be left at home safely while online.
We here in Germany (and Europe) had something called ISDN,before DSL hit us in 1997. ISDN was like "digital dial-up" with two lines, both with 64kb/s. So we solved the simultaneously speaking and surfing problem. And we could bundle the lines and have 128kb/s internet speed. Btw, i like your in-depth videos a lot!
Same here in the UK. ISDN remained heavily in use for temporary high quality audio links for radio stations and is still around (but disappearing fast).
Yep, was marketed for the home user in the UK from about 1996 I think but was obviously in use in business for a long time prior to that. It came a little late (for the home) but for a year or two it was about the best home internet connection you could have, I remember it was a huge advantage for Quake players since it was very low latency compared to a modem.
I used to work for an ISP and there was this VHS vs Betamax period where there were two competing 56K standards (K56Flex (Lucent/Rockwell) and X2 (USRobotics)) meaning we had to invest in two types of access switches (banks-o-modems) with two different dial-up numbers, much confusion ensued with users using the wrong number and complaining of slow service, V.90 couldn't've come sooner!
What a fascinating video. I worked for AOL in 99-01, RadioShack after that and AT&T after that and I had no idea any of this existed. Thanks for all the hard work putting this together.
You're my favorite TH-camr. I get so excited every time you post a video because I know it's going to be exactly my niche of "who even cares about this" and you take a deep dive even if you know you're not going to get it 100%. Your presentation is amazing. I knew tons about how modems worked already and still learned something new.
The design of those ASVD units is delightful! They coordinate perfectly with the case designs of the mid-90s, which so many hardware manufacturers ignored, and their form factor brings to mind the desktop intercoms whose functions they replicate! Just stellar work all around. 💕
Even having grown up around dialing my friends in the early/mid 1990s and seen my share of modems, dialup doom games, and tinkering with AT command sets, all of this SVD stuff is completely new to me. Even voice modems were a relatively rare novelty. I appreciate your nostalgia for all of this stuff, it was an era I don't miss very much at all but I love the historical deep dives.
When I started calling BBSes in about 1991, I had a 300 baud modem on my CoCo2. You could actually pick up the phone and tell the difference between just the carrier signal and when data was being transmitted. As a fourth grader, all of this blew my mind. Computing was so much fun back then! Great video.
You could also tell who was transmitting at the time. Your modem produced the lower tones, and the higher pitch was the modem on the other end. That, or you could flip between the two carrier tones?
as far as voicepath or v.61 is concerned, it works pretty much as you described it, it uses a simple 4 point qam constellation and then modulates the resultant qam points with the voice traffic limited to make sure the resultant output still fits in the constellation and decodes with adequate margins, then on the receiving end the data is extracted and then subtracted from the signal and what you're left with is the modulated voice data that can be extracted. it would be interesting to hear how it sounded when there is enough line noise to cause data errors and it subtracts out the wrong data, although a 4800bps it would probably take very loud input sound and lots of line noise to get it out of the margins and actually see data errors neat trick though and one that i actually saw implemented recently in an engineering school project to encode hidden extra data channels in the noise margin on optical fiber transmissions
didn't fall for the VoIP trick, but also lack the signals theory to figure out where you'd find a big enough hole in the spectrum to reliably stuff in a few kHz. thanks for the elaboration!
That has to be one of the most bizarre and slightly terrifying ways of modulating data I've heard of so far. Modulating the modulated data itself, that just feels... Wrong. Thank you for shedding light on this manner
@@MrRedwires if you want to feel this feeling even more, CuriousMarc has some videos about how the Apollo moon landing data, voice, and TV was modulated. At points there's phrases like "PCM on PSK on FM on PM"! And modern wifi goes even further with stacked modulations, it's pretty crazy. Of course if done wrong, one modulation might interfere with another one, but when they're picked carefully it's just like putting a box in a box in a box!
@@kaitlyn__L B&W backwards compatibility in the NTSC color standard quickly came to mind for me also. despite being a heinous design constraint, conceptually & mathematically i agree with the box-in-a-box description... i'm still somewhat in awe of this kind of thing happening decades ago, with comparatively primitive simulation or formal verification tooling to gain confidence in the design, and it pretty much just worked 🙃 (that could be confirmation bias though, i wonder if there's a good example of a broken attempt at backwards compatibility that ultimately steered a product line to a premature dead end)
Great video as always. When I first started watching your content I didn’t know your subscriber count, assuming you were in the millions. Your content is top notch and it’s just a matter of time until you grow. Keep doing what you do, and don’t change a thing!
Yeah, i subscribed pretty early on and it's been awesome to watch this channel grow. Every video is better than the last in my opinion and I think his style hits people at the most basic beginner enthusiast level and provides great info for the more hardcore. Love everything I've ever seen here.
Great episode! I remember how happy I was when I had a modem that could pause all activity on call waiting, let the phone ring and then resume everything after you've reconnected. There was even a little downloader program that allowed for a download to resume after you've taken a call and reconnected to the internet.
i've had people yell at me before for being a prescriptivist about it, and frankly, I see their point. *almost everyone* calls it DB9, and we aren't ordering these out of catalogs - _even some catalogs call it that._ and the distinction is meaningless: did they ever make a "DB" with 9 pins? absolutely not, so who cares if we make that an alias? DB15 is a different story, because they DID make that in both B and E sizes - but nobody really argues that point. honestly, we should probably chill on it, but, I'm not gonna lmao
Wow. I left speechless with the amount of efforts you put in this video. Thanks
ปีที่แล้ว +21
Amazing research on this again. I love the way you slowly ramp up the topic, so we get to know first every detail about the niche situation where these devices were useful. I've never heard of these devices. But then again, the whole dial up thing passed me by, at my parents there wasn't even a wired "landline", just something which was emulated over GSM via a Nokia Premicell terminal, and I didn't have the serial cable to use it's CSD capabilities (I wasn't even aware it had such feature back then).
Thank you so much - I put a lot of effort into making my stuff _highly accessible, without being dull_, and people rarely comment on that specifically so it's nice to hear that my efforts are worth it! I just figure, an awful lot of people probably know _more or less_ why these limitations exist, but not the details, and filling in the blanks helps the whole story come together. One thing I wish I hadn't missed was early data over GSM. There was a whole ecosystem of products based around it, but in the US, almost nobody used them except traveling businessmen, and by the time I became aware of them. all the systems were long gone.
My parents actually _never_ got a landline, so we never had a phone until cellphones became the norm in the early 2000s. I think I used CSD once, I was desperate due to some deadline, I think it might've been school recruitment or something like that, and my home broadband either failed at the time, or it was before we got one, I don't remember that clearly. Anyway, where I live, until the late 2000s it was conventional wisdom to _never_ use cellular Internet because it could blow your bill in minutes - especially CSD, but early GPRS wasn't much cheaper either. I remember people quite literally having the fear of clicking the Internet icon on the not-quite-smart cameraphones of the day by mistake. I felt like an early adopter among my group of friends, using Opera Mini via EDGE on my Nokia 3110 Classic ca. 2008 - that's more or less when mobile Internet started to become affordable-ish, likely thanks to the 3G and smartphone revolutions happening around that time. But most people I know were still clinging to the idea that it's prohibitively expensive to _ever_ turn on data for at least another year or two.
@@kFY514 yep, once time I tried GPRS and loading Google halfway used up the £10 of credit I'd just added. Then it gave me an error because I ran out of money. And then I realised it wouldn't be able to pick back up from where it was. Then I knew never to do it again! The first time I used mobile data was after I saved up for an iPhone, because I made sure to buy 500MB as well. On pay as you go, data you hadn't specifically bought ahead of time was still charged at a ludicrous rate. One time I forgot to get another 500MB after the previous month's had expired (there's a whole other "joy" to go into) and again ate through about £15 of credit accidentally. That's actually what got me to switch to a contract, at the same price as the 500MB I was buying each month, but with 1GB and a bunch of minutes and texts too.
@@CathodeRayDude You’re legitimately how I’m learning about this sort of tech, so you’re definitely doing a great job making it accessible.
ปีที่แล้ว
@@CathodeRayDude technically I think you could revive some of these stuff to be tested, eg. there are GSM base station emulators for SDR hardware which are getting more and more accessible for the average person. There are also some specific Motorola cellphones which can be hacked to function as a base station with something called osmocom. But running a GSM network (even if it's being done with very small power and for a short amount of time) is a slippery-slope at best as I've heard :D
I remember in the mid 90s, reading an article or two about the makers of the PGP encryption software in that they made software which would let you make secure, encrypted phone calls over standard modems. It was called PGPfone. I never heard of it being used, nor can I find much information about it. I wonder if the US government killed it off. It would be cool so see demonstrated though. Side note, an air fryer is an oven. A convection oven, more specifically. The one thing it isnt is a fryer.
I don't believe there's any conspiracy necessary. It's just non-trivial for the technology of circa 1995 to have the processing power to encrypt a phone call, decrypt the other side, encode and compress the voice, and decode and compress the voice all through a pretty-slow connection where you'd be lucky if you got 28kilobits/second. I recall hearing about this at the time, but I'd be surprised if it worked very well with mid 90s technology. By the time the technology caught up in circa 1999/2000, there were far better ways to talk via encrypted channels over the internet, and not over a modem.
@@stevesether Im not talking about a conspiracy. It wasnt a secret the US government wanted to prevent foreign countries from getting ahold of strong encryption. There were laws preventing the export of certain types of encryption to specific specific nations. I have no idea of the quality of the audio was or if there was latency. However, Im sure it wasnt advertised as something you wanted to call your mom with. It was for people who wanted to talk securely.
There are still legacy systems around that work like you describe at 11:00. I use some of them at work (logistics/supply chain related stuff), which means you are seeing data that isn't actually "live" but was exchanged with another computer very early in the morning before the workday started. Of course this actually works now with a bunch of custom made emulators and compatibility layers over a VPN on the regular internet, but fundamentally its the same 1980s code that in the past a minicomputer at a warehouse would run to make a phone call to a mainframe at 3 a.m. to exchange data.
@@user-fh2fm7vr4m if you think that's crazy, look up how the airlines still operate with booking flights and sharing data cross companies. There's a good video on it out there, probably Wendover or half as interesting or something.
My Dad used BBS for about half a year in the mid 90s before discovering the WWW, then with Dad/Mom both being gamers and Quake coming out, they decided it was only logical to get a second line. The interesting bit though a lot of people overlook is that they also used a service called MPlayer (not the media player). MPlayer was like the first "lobby" for video games as we know today. It would hook up Quake players with other Quake players unlike Gamespy which just showed you active servers. Ontop of that, it apparently had a VOIP-like voice chat. My father told me people would join MPlayer Quake lobbies *just* so they could have voice chat with their friends, and that in combination of running around a 3D world with your QuakeWorld skin (Avatar) made it like a prototype Second Life or VRChat. Of course, MPlayer died not too long after, Half-Life/Counter-Strike gave us VOIP in games a couple years later, and my parents dropped Dial Up as soon as they could sometime in '99-'01. That was the end of Dial Up for us, well except when I went to Grandma's. Took her till the later 2000s to upgrade heh.
Your explanations of the modulation techniques was dead on. I had to go read the spec out of curiosity, but you nailed it. Just some classic nested modulation madness.
Great video about technologies I didn't know existed. Growing up in South Africa, we were behind the curve technology-wise. I started using the Internet at university in 1996, but had dial-up at home from 2001 to about 2005. In 2005, my sister Skyped her boyfriend (audio only) in the USA through my Linux VIA EPIA system over a 56k modem and it worked well.
One of the first computer programs I ever wrote (back in the late 70s, and early 80s) was an address book for my TRS-80 Model 1 computer. I designed and made a little circuit that would use the cassette port to pulse-dial the telephone, at 20 PPS, and not the usual 10 PPS, that normal dial phones used. I kept all my phone numbers in that program and used it for many years! I even saved the database of telephone and addresses on a cassette tape, that I could load into the computer. Fun times!
Interesting. I wouldn't think most electromechanical telephone switches would handle fast dialing like that. I know the 5XB switch would throw trouble cards if you were out of spec. Maybe you happened to be on an early electronic switch.
You’re right on the lack of testing. i purchased 14.4 modem when it first came out in 1991 and there was a particular command function that was not working, I was a computer nerd that used procomm plus (they are still around). I called up tech support and they mailed me a pre-programmed EEPROM, and a note to please send the old one back. I swapped it out and it worked no problem. I wish I would have held on to the modem, it was an ISA board that came in two halves that came apart and it was chunky. Believe me there was no Rockwell modem on a chip on it.
Ah, man, the stuff I missed by being bleeding edge! By the time this stuff was around I was running ISDN, and BRI came with 2 lines. Thanks for an informative look at the stuff I missed. :) Fascinating as always.
I never had ISDN at home, but it was definitely a real solution to the problem at the time. It was just a little too late, because by the time ISDN became widely available, cable modems had come out to largely the same places. I know my family got a cable modem in 1998. In 2000-2001, I actually used ISDN, but in a full-out networked situation, which made it kind of worse than dial-up. We had it for Internet access at the local senior citizen's community center, which was where my high school CISCO networking class was held. There were like 12 of us sharing the ISDN line through Ethernet during the class. Before class started, we would all be trying to browse at the same time and it would bog down pretty hard, but if you snuck in some browsing while the teacher was lecturing, 128kbps was pretty nice at the time.
Man I just want to say that I love your work, I really appreciate the effort you put into these videos. I always find myself impatiently wondering when you're gonna post something. Greetings from Brazil!
This is fascinating, I had no idea such things existed, but I can imagine how many hours it took to figure out how to get all these demos going for us, so thanks for this ultra-level nerdiness! Sidenote: my school district had Dells of exactly the same model as yours until at least 2005.
Working with radio, i found out that modern RF modules have a weird solution for networking over UART. They simulate packets in software by either making each transmission the same length or adding a start and end code, and then add the "metadata" to the beginning of each packet so that the receiver always knows who sent what. This does mean that all data going in has to be parsed by the receiving computer, but it works.
In joining the chorus, I, too,, thoroughly enjoyed this walk down memory lane. I think I was 19 or so when those modems came out. I think that if I were able to afford the expansion slot version, I would have gotten it and just sat it beside my generic hayes-esque modem of the day. I was big into telecom for a long time, nice to see the coverage of a very old, but also very important, bit of history. Kudos, sir.
so im a telecom person in north america - he mentions CPC, which is something we use here in USA. im 30years old now, and i have only seen TWO isdn circuits my lifetime. I have many many more PRIs. I am also of greek decent, and do go to greece often, so im aware how ISDN works etc, i dont know why, but it seems like ATT was just a lazy company. *shrug*
I remember in the 90s. On of my friends had ISDN at his house. Not only did he get 64kbits download, but he also had a second line for the phone at the same time. This was absolutely groundbraking. He could call AND game at the same time!
We had 56k (X2) in 1997, and V.90 in 1998 as soon as USR (3Com) had the firmware to do it. The modems were DSP based, so all software; new standards were trivial to support. (in fact, X2 was a software upgrade) PRIOR to the USR modems, we had much older tech using Rockwell chips that had to be physically replaced to change modulation support. (which we did in 95-96 to go from 28.8 to 33.6.)
Alright, after spending *WAAAAY* too much time analyzing the V.61 ASVD spec, I'm _pretty_ sure it works as follows: (First, your QAM idea is mostly correct - although by 14.4, modems were using multiple frequencies.) First, the primer on modems: Models turn 1s and 0s into the screechy noises in the audible frequency range that phone lines transmit. Phone lines, in case you hadn't noticed, aren't exactly CD quality. One digital switching took over, analog phones lines allowed only a subset of the audio frequencies the human ear could hear - specifically the frequencies that the majority of human speech takes place in. So modems have to figure out a way to compress 1s and 0s into audio data. Note: Much older computers like Commodore 64 and Apple II tape ports worked much the same way. They store the data as analog audio. It's just that tapes have a higher dynamic range than phone lines, so they had "more room to work with." As time went on, engineers figured out better ways to compress data down and modulate it differently to allow for faster modem speeds. Up to 33.6 kbps is still 100% an analog "screechy noise in the audible human hearing range" compression. At the time of the V.61 spec, 14.4 kbps was the maximum speed. Yet for ASVD, you only get up to 4800 bps (4.8 kbps) That's because it used the compression _methods_ for 14.4kbps, but only used a tiny part of the audio "spectrum" available to it on the phone line. Now, for a side-bar on old vinyl records… Old vinyl records store audio in wiggles in the groove. But, because of the way it's stored, low frequencies carved into the groove at the same "volume" as high frequencies would be overwhelming to the needle, and a lot of the noise on a record is at higher frequencies, which would drown out the higher frequencies. So the RIAA developed "the RIAA curve" - sound is recorded with the bass lowered, and the high frequencies boosted. Then when played back, the amplifier boosts the bass, and lowers the high frequencies. This makes it sound like it should, while cancelling out (somewhat) the negative aspects of vinyl records. V.61 does similar to the voice portions of an ASVD call. Part of the voice frequencies are just outright cut out for use by the data (small gaps in the audible frequency range )others are compressed-and-expanded. You have even less dynamic range than a normal voice phone call, but it's enough to be listenable. if your call was local, it would probably be about the same quality as an international phone call over normal phone lines. The modem would compress the dynamic range of your audio stream (not digital compression, purely "squashing the signal") and send it analog over the line, then the receiving modem would decompress the dynamic range, introducing noise, but making it listenable. At the same time, it would cut out some audio frequency range completely from the "audio out" (speakerphone/headset jacks) because that portion was used for data, using standard bandpass filtering techniques. It's the same basic idea as a "DSL filter" would later use to get the "DSL frequencies" off your voice phones. Internally it's a splitter - one part goes to the modem, with a bandpass filter that can be turned on or off to hide the voice frequencies, the other part goes to the "audio out" portions, with a bandpass filter to turn off the data frequencies. If you did a "data only" call, the audio out portions would just be completely disabled, and the data modem would get the full use of the line; if you did an ASVD call, it would split it up thus the modem had to be much slower. This theoretically could have been done with older modem technology, too; but it would have been so slow as to be truly unusable with the older modem specs.
WHAAAAT? That's mind-blowingly cool! 🤯😎 Also @ 18:52 you literally killed a serial/networking connection by just yelling? Never thought I'd see that happen, LMAO!
It's really weird. Everything you show I lived through. I was extremely happy when every bit of technology you show was upgraded or replaced. Yet, your videos are hypnotic! Nice one!! I fix up 1930s to 1960s valve radios, so my parents think exactly the same about my hobbies.
26:00 v90 and v92 require compatibility with the equipment at the exchange (connection needs to be PCM not analog to achieve speeds higher than 33600 bps) which suggests there's some additional signaling required on the exchange side to make mixed voice and data calls possible.
i never searched for it because i never had the curiosity for it, but i finally learned the purpose of that sound. thank you so much for making these videos of old harware. love it so much.
32:30 I remembered that just today. I had looked up a 'phone number on my PC and just for a moment I remembered the good old days of being able to dial it up from the PC. Not sure how that was done now, but I remember it worked.
I love videos of solutions to problems in the 90s. My dad had a second line installed for his business. The sound of your voice speaking in ASVD the mic at the end would make a cool song mic audio.
When you shouted into the phone, as soon as I saw "No carrier" I burst out laughing before you even did. That was hilarious. Keep up the awesome work. 😂
We had two phone lines in the 90's and early 2000's so dialup wasn't a huge issue when we needed to make a phone call while dialed up on Prodigy or AOL (America On Line) lol
Perfect video! I can imagine the tons of research and the hours of fiddling with parameters to make different brands talk to each other... Even if I had owned such a modem, I would have needed your TH-cam channel to know how to use it! Congrats
14:31 claiming all modems are identical is very much misinformation. Yes, they all use the same modem chip, but the cheaper ones had worse support hardware and filtering on board. I was on a very marginal phone line, and fried a ton of modems. I assume this is due to our local carrier having issues with their ringer and pumping excess voltage throught the lines, but I can't test that now. Neeedless to say, I had modems dying after ~2 months, the longest lived were US robotics around 8k-10 months. I looked over my pile of dead modems, since I'd continued to save these for the relays on them, and the more expensive modems had much more extensive filtering circuitry., A more expensive modem had better filtering circuitry in front of it, which not only provided better connection stability but also helped insulate the modem from poor phone lines. Yes, all of these modems circa 98-03 use the same chip, but there were huge differences between them and a better quality modem would provide a better connection. The analog circuitry in front of the chip was the major differentiator here.
That was very informative! Loved it! I lived through that era. It was a plesant stroll down memory lane, hearing those modem noises! Wonderful! Thanks.
I ran a FIDO node in the 90s (bit late to the party; didn't bother me as a teen); you wouldn't believe the compatibility issues you'd have with even same-chip-different-oem modems at times 🙈 (US robotics and zyxel did well to me; ELSA was fine too, but a bit picky -- and quite pricey at the time) *edit* also, 16550 vs 16540 UART vs that cheapest of onboard serial controllers your big-box-store PC might've come with back then? quite a margin.
@@CathodeRayDude keeping that BBS running, with my dad going mental re: phone bills? a mess, it was, indeed :) (but also quite a learning curve: when IP, broadband, "se internet" became a thing, I already had a clue what that decentralized networking thingie might be about. served me well for years.)
I was lucky that my parents got ISDN with four channels very early, because my dad often worked from home. He was a very modern man. That was in Germany during the mid 90's and that was really really expensive. Each ISDN channel is basically a digital phone line. They also could be combined. In 2000 my parents got T-DSL with around 700-800 kbit/s. I used that as an opportunity to download pirated movies and music CDs, burned them on CD's and DVD's and solt them on the schoolyard. Made a lot of money for a 14 year old. My parents were upper middle class, but they were pretty strict not to spoil me with money, so i worked for it. When i was 16, i had enough earnings to afford a beast of a PC. A brand new Pentium III 800E system with four DVD Burners and massive hard disk. I believe i had two 40 gigs and external firewire ones. What a time that was. Nowadays i still own a business, but with more legal methods ha ha.
We didn't get ISDN until 1999, but having the option to either use both lines to get a whopping 128 k connection or use a 64 k connection while still using the other line for regular phone calls felt incredible at the time. Of course that was soon surpassed by DSL (we got 512 k HDSL in 2001), which could use full speed regardless if you were on the phone or not. Come to think of it, our connection speed seemed to double every couple of years back then. The speed internet technology progressed was remarkable.
Excellent video! Loved when you started talking about voice coding towards the end. I find it particularly fascinating on land mobile radios. Something to note, error correction comes in a few flavors, you can use ARQ (Which requires a retransmission like you mentioned), or you can do forward error correction (FEC). You add redundancy to your message to make it more tolerant to noise, whatever source that may be. Now that being said, that slows data rates as well. But with voice packets, you usually don't care which you also mentioned. So in slighlty more modern systems, your best bet is to use some type of datagram scheme and just do FEC on it to the best of your abilities after characterization of the channel and more often than not, the voice sounds intelligible enough even in the harshest of conditions. The fun part is when you get into the wireless sphere and start analyzing energy per bit to noise power spectral density and how the FEC impacts it as RSSI decreases. Fun stuff 😀
I'm going to start saying QAM like wham from now on. It is normally pronounced like "kwom" rhyming with ROM. But I like yours better. Anyway, these modulation schemes are used all the time in rf engineering, so that is where you go to learn about that sort of stuff.
I have vague memories of SVD being a thing back in the modem days, but I never actually tried to use it. Watching this, I'm glad I didn't. Also, I find it amusing that the modem being called tiny is larger than the 33.6K modem we had, which was small enough that it had a Mini-DIN port instead of DB-9 for its serial connection.
For that mic input issue, sounds like there may be a bad capacitor in the unit causing those issues. Would probably be a fairly easy repair if you want to tackle it.
@@CathodeRayDude bad electrolytic caps were notorious right around when that unit would have been made. Should be easy to spot if the unit can be opened easily.
Man, normally I get so bored with long-form videos.. but I totally can just sit and watch your shit for hoooouuurs. So much interesting shit. A lot of which i already knew, but it's just so cool to see it all in action and stuff.
The intro to this video made me extremely nostalgic. When I was younger I my parents would routinely take me to our farmhouse to get away from the city for a couple days, much to my dismay as a tech enthusiast in the early 2000s, and the only technology I had access to was my SNES that I had to beg to bring over every time. One day I was playing Super Mario World and reached the furthest point in the game I had been to at the time: the ghost castle where you're supposed to solve a puzzle to progress (IIRC you have to enter a specific P button, otherwise the stage loops indefinitely). As a kid I had no idea how to solve it, no internet to look the solution up and no telephone lines to ask my city friends. The solution I found was to walk 45 minutes over to a friend's house, ask them how to solve it (as well as how to perform the super moves in Final Fight 3), and walk back. That specific day always stuck to me as a funny point in time where internet was already a thing, but not commonplace enough to be convenient. Very retrofuturistic in a sense.
When I saw the intro, I thought it was a gimmick and wanted to chime in and say our (monopoly) telco had a tiny application that could indicate a call waiting if someone would call you (I think even with caller ID), but it's obviously far more inferior to the solutions you display; I'm jealous at your setup :) I do remember vividly the day I wasn't allowed onto the internet since my brother was paying for it, and blowing into the handset (instead of screaming) to disconnect while not triggering any suspicion. I also ran into the problem you mentioned at 19:00 and I never got the game to connect, with spite up to this day. Thanks for making this awesome video!
i loved this video so much (i mean i love all your videos ngl), its so cool to see how chaotic the landscape of anything was in that era, i like deep dives into all the weird shenanigans of a "standard", i now feel like i understand how modems work and what could have been and i know i'll rant about this several times in the future lol so thank you for this
Just a caveat on Hayes commands when picking up an ongoing call: You can issue an ATO comman to start negotiating the carrier as with the ATX0 + ATDT method, just easier and without messing around your modem's dial settings.
I worked for a Dutch telco at the time. We were working on using DSVD for doing actual normal phone calls during a modem call. Our proof of concept forwarded your normal phone number to the modem on our end that would put the phonecall inside the DSVD session if you got called, and if you picked up the phone you would have gotten a normal dialtone. This was linked to the PPP/RADIUS login, so if you’d dial up with someone else’s user you would be able to use their phone number, which was a funny side-effect. ISDN and DSL happened shortly after, there was more money in both of them. So it was cancelled. Also, around 90% of the market wasn’t using our internet service but the ones of dedicated ISP’s, not the phone companies one, so management saw no money in it, which made sense.
I'm at minute 6:00 and so this may be premature to share, but there used to be a service called CallWave which used the "call forward on busy" feature that almost all phone companies had available for free to forward calls to their toll-free number. When this happened, I guess the phone company would also forward the original destination number and then an app would pop up on your PC, tell you who is calling, let you screen the call like Google Voice does, and press a button to answer it if you wanted. If you did this, it would immediately hang up - this was pre-V.92 - and then their toll free number would call you with the original call. They even had some adapter that would make your actual phone ring (we never bought it). It was only $4.99/mo or something like that and I found it life-changing. I could download something that would take days with the download manager I think had the word Gator in its name and not worry about missing a call and getting in trouble. I'm sure that the hour I have left will reveal something absolutely insane that I never knew about, but I wanted to share CallWave in case that's the thing and I'm not the only person on the planet to have ever used it :).
And I'm done. Definitely never heard of either of those technologies, but as Tay Zonday mentioned, I did use Roger Wilco as well as MPlayer (which apparently is from the same people that made XBAND, but I never knew back in the day). CallWave was the thing that solved the mom yelling at you for playing Quake online all day while she missed calls problem for me. And it's such a simple approach that I'm still surprised that I'm among the 5 people that remember it. You'd think that if you were a major ISP like AOL or AT&T WorldNet that you'd add this technology to your service just so you can one-up the competition, but it never happened. They had call forward on busy, they had software-controlled phone switches, they had all the call metadata. But as far as I know, only CallWave did anything with any of this. I should also add - they could detect if the incoming call was a fax call or not and if it was, it would just auto-answer and let you download JPEGs and TIFFs of the fax pages once it's received and processed. Like all stuff that AT&T could do for essentially free, they just needed the system tray application.
I'm heartened to see that I'm not the only one who remembers the promise of DSVD. I was working at Radio Shack in 95-96 and remember reading about it, probably in Nuts & Volts, or a similar magazine. I figured that the problem with needing two identical modems would be resolved through broader software adoption later down the line as the tech became the standard. As for me, I just barely got into the 56k era when Road Runner came to town in 1998, which I was the first in town to have.
I'm happy that my connection in 2002 was fast enough to handle Teamspeak while I was gaming, I lowered the bitrate though to improve my latency. This is next level, I would have loved this before the year 2000.
Your laughing scene summed up every parent/guardian ever (at least on the inside) in the 90s who deliberately did such things to disconnect the modems so that they could use the phone 🤣🤣
OK DSVD that is pretty cool. I knew that there was a bunch of other features floating around at the time but I just struggled in the era to get a modem with voice capability that actually routed the voice capability through the sound card not out to separate jacks. I am myself enjoyed the speakerphone ( and answering machine ) software that was built into my Compaq. And Willow phone a neat little under a megabyte Windows software that made any voice capable modem a speakerphone and only worked on like three computers I've ever had. And the way the internal ones actually use the phone with a ring signal is pretty darn clever if only that was made the standard. Most of the modems I have used interrupt the secondary device that seems like a really neat use for it couple it with some sort of car waiting and someways I wish we had another 10 years of standardizing and the modem Field where we could get to one chip that supported everything.
FunFACT: Television Close Caption line 21 encoders and decoders, still use dial-up modems to caption TV shows and News every single day... They even still use 1200 BAUD, 7 bits, Odd parity, and 1 stop bit. That is STILL the industry standard for captioning systems!
![PASULOL รามเกียรติ์ ตอนที่ 4 นนทกพบรัก [Ramakien Ep.4: Nonthok in love]](http://i.ytimg.com/vi/tMUQgmZzf_A/mqdefault.jpg)
So, I read the V.61 spec and you got it pretty much spot-on.
It's hybrid analog, with a modulation that looks like it's a slightly faster V.32bis (either 2800 or 3000 symbols/sec, compared to V.32b's 1800). When you're not using the voice feature, it's supposed to be able to use most of that capacity to give you a 19.2kbps data connection (though I wouldn't be surprised if something made it effectively 14.4 for you).
When you turn on the voice mode, it keeps the same symbol rate but it downshifts from QAM256 (8 bits per symbol, 19200 usable bps) to QAM4 (2 bits per symbol, 4800 usable bps). This leaves so much "extra space" in the constellation that the modem can add a (filtered, processed, and decreased in amplitude) version of your voice to the signal and the other modem will still be able to decode all of the data bits. After it does that, the receiving modem subtracts the value corresponding to the data bit from the signal, and what's left is the processed voice signal. So it undoes that processing, and some slightly crunchy voice comes out of the speaker.
19,200bps (data-only) / 4800bps (voice+data mode) only requires 2400 symbols/sec, but I said there were either 2800 or 3000, so where do the extra ones go? They're used for a data "control channel" between the two modems that always runs at the low rate (800bps) and constantly exchanges information about the line conditions so that the modems can choose the optimal values of the audio processing filter values, and it's also used to switch between the data-only and voice+data modes basically instantly (25ms).
I *think* that the reason for offering two different symbol rates (even though they both give you the same data rate) is that the higher symbol rate also allows slightly wider audio filters (better voice quality), but the lower symbol rate is more likely to work if your phone lines are crap. The standard never comes out and says that, though.
Oh, and most people pronounce QAM Q-A-M, just like AM or FM.
Thank you for this excellent explainer of how we were able to squeeze data and voice through a narrow pipeline!! Now, my next thought, is how this will all degrade with the switch from analog phone lines to digital VOIP.
@@setoth1234 anything that meets the standards that the telcos keep for digital voice (8kHz PCM with minimal jitter and no "advanced" compression -- like a T1, or G.711 over good solid VoIP) will handle this just fine. The long-distance infrastructure was already largely digital in the 90s, so modem designers took account of that kind of thing.
Sending it over GSM or AMR or G.729 or Spex or Opus or Silk will murder it, but that's true for basically any modem, it's not specific to SVD.
Telecommunications Act 1997 (Cth) | ALRC for John Linguis
Why not use ISDN? That what the rich Kids used back then.
Yelling into the phone until the connection dropped not only made me laugh out loud but also reminded me of that video where a guy screams at a bunch of hard drives in a datacenter and the performance drops noticably. Thanks for the lesson on technology i never got a chance to use
i found whistling loudly down the phone was the quickest way to get the modem to drop the call. came in useful when my sister was hogging the family computer and i wanted to use it :D
lol I watch that video all the time lmao
IIRC there was also a particular model of hard drive that would crash if a certain song was played.
@@DrewWalton Janet Jackson's Rhythm Nation.
@@Nuskrad you could have tried to emulate a line reconnect, I had about as much crap as CRD (still have some of it) and I have managed to 'negotiate - connect' handshake with a modem several times using my voice. Yes I know how wasting time on shit like that is pointless, but there was a time in my life when I had a bunch of time and tech junk around to have nothing better to do.
We totally had Roger Wilco 1998! You could use it to voice chat and play a multiplayer game like Starsiege over dialup. It worked like a charm over my 26.4 modem.
I make my modem hold the call to breathe in
CHOCOLATE RAIN!
as a fellow baritone, i solute you Tayzonday
Holy crap! I can't believe I FORGOT about Roger Wilco!!!
I came here to say that . I think I was even using the beta service in 97 with a friend.
I appreciate you using the classic Door stuck clip to demonstrate CS 1.6 voice chat.
DOOR STUCKKKKKK
@@CullenCraft PLEAASE, I BEG YOU
My dad actually ""solved"" the problem of using a phone line for both data and voice in his youth by installing 2 completely different phone lines in the computer room. How he actually convinced his parents to do this I have absolutely no idea.
As one of those who was a teenager in the time of dialup, I can assure you that after picking up a phone so many times just to hear the squeal of a modem, eventually it wasn't a hard sell to get a second line.
I still remember my father yelling "Enough is enough! We are getting ISDN!" when he tried to use the telephone while I was online again. :-)
when the phone line is in use nearly 24/7 with dialup, sometimes people need to make phone calls....
ugh, i remember those days "get off the phone, i need to make a call" x.x
It was easy for me. I paid for the second phone line.
Late 80's I did my fair share of BBSing. Eventually the idea of sharing the line peacefully with a teenager was painful to envision. Best part was it let me operate a BBS since I could accept calls without ringing throughout the house.
Good times, indeed.
My family used a program/service called CallWave back when we had dial-up. Basically, anyone calling your phone number would get redirected to this voicemail service. Through a program running on your computer, it would alert you to the incoming call, and would let you listen to the message while still online to decide if you wanted to disconnect and call them back.
Was trying to remember what that service was - callwave. Thx
AOL had their own free version of this that we used iirc
My parents refused to pay for services, so we used K-Mart's Blue Light free dial-up and CallWave. CallWave was a game-changer in my house.
Ah. Should have read the comments before mentioning that myself.
It worked using the call forwarding feature of your phone line, and you could set that up by dialing a special code before each call, same as you could for disabling call waiting.
And the service was free--it just used advertising to support it. The app was always on screen, showing some ads. And I'm not sure, but I think they eventually would play a very short ad on the actual call. You know the kind that just says "Brought to you by . ."
Grandma had this, but couldn't remember the name!
Hello all! I was a big modem user from the days of a manual dial 300bps modem to the last v.everythimg Courrier, and I recall most of my modems (if not all) used a relay to cut off the attached phone on the passthrough jack. So when I transferred files with friends, we'd just leave the phone off the hook for when the modem hung up, so we could continue chatting by voice. sending 800k Apple IIgs disk images across a 2400bps modem took a very long time... It's possible since I was a heavy modem user for so long I always had somewhat fancy and expensive modems, so it's why this feature was there. I got DSL at the start of 1999 so that was the end of the line for my modem usage.
that's interesting - the 300 baud modems DID cut off the phone jack, i forgot to mention that, but none of the newer ones I've ever used did it.
Adding to this, even my original manual 300bps modem cut off the attached phone when you switched it to answer or originate. I'd dial a BBS, wait for the carrier and switch to originate at which point the phone would cut off. To hear the modem carriers you'd need to pick up another phone elsewhere in the house.
@@CathodeRayDude The USRobotics Courier v.everything modem has a voice/data physical button on the front panel to even switch between the two modes. I can't recall when I got my Courier but it was the last modem I got because it was a fully firmware upgradable modem that could be flashed to support new standards. IIRC when I got it, it only supported 28.8k but then I upgraded it to 33.6k and then finally 56k at the end. It also supported the proprietary HST protocol which was a USRobotics 14.4k (or was it 19.2k) thing while everyone else was stuck at 9600bps. Some BBS's used HST modems so it was a nice speed boost. The Courier was costly compared to other modems, but it was the bees-knees. I still have an unopened v.everything and the branding on it says "MISSION CRITICAL CONNECTIVITY" in big red lettering.
Fun fact: the modems inside all mobiles phones (to this day) use the same AT command set for everything. It's just hidden away by the mobile phone operating system. This command set was standardized by Hayes in their first smart modems. (Or at least I think that's the case with the AT command set.)
That was definitely not common -- at least on later models of less-than-Courier-class modems. I had a generic 2400-baud modem that came with the family 386, which we upgraded to a Zoom 28.8, and I later had a US Robotics internal 56K modem. None of those would stop you from picking up a phone attached to the second jack during a data call.
I do seem to remember they all (?) had relays, but AFAIK, that was just a disconnect between the phone network and the computer. (Probably to help minimize the likelihood of damage from nastiness on the phone line.)
I have quite a collection of modems now. Some day I'll have to look into whether any of them did this. I never knew that was a thing, because none of mine ever did, so it hasn't occurred to me to consider this in more recent years. Interesting.
@@CathodeRayDude Another tidbit now I'm thinking about things -- with modems with a relay, when you called a number (with the modem) and someone answered, you couldn't just pick up and talk to them. So, with the hayes protocol, simply pushing any key on the keyboard after issuing a ATD will cause the modem to instantly hang up (or switch back to the attached phone if you a holding the handset off hook.) I'm skimming the Courier HST manual and it works this way, so it had a relay.
I still have my old 2400bps modem from 1989 (a Zoom modem) and I just opened it up, and sure enough the two phone jacks are just connected right to each other. So I guess the cost cutting had begun already!?
Your laugh at the yelling to disrupt the carrier reminded me... I was a member on a MajorBBS with some number of lines; I forget how many at this particular time. But the sysop had some kind of flow control or disconnect detection issue between the BBS and the modems wherein a line could drop, but the modem would remain connected to the BBS. After a user was dropped, when the next caller would hit the line for that modem, everyone would see that ghosted user sending "RING" "RING" "RING" over and over until the caller gave up. To keep from having to manually jump on the system and disconnect the errant modem, the sysop added an action word to the chat room that would disconnect the user if it saw the word ring on a line by itself. My girlfriend spent significant effort tricking people into typing it and kicking themselves offline.
Oh my gosh, if it were me, and I knew what was happening, I would have started typing modem commands in chat to see if I could get it to dial out! Or also send ATA to have it answer and the new caller would be logged in to the old user's account...
@@PinkPandaKatie Convincing the BBS to send AT commands without having your own modem respond, sounds like the recent log4shell/log4j vulnerability
@@tkteun Modems don't respond to commands when they're in "data" mode. You have to either use one of the RS-232 side signals to switch to command mode (like DTR. DSR, I forgot which one) or send "+++" without sending CR, then wait until you get a command response. In this case, your modem is in data mode, but the other one is in command mode.
@@PinkPandaKatie Thought that was more of a PPP thing than a BBS :)
@@tkteun Well, the modem doesn't know or care whether you're using PPP or not
I used to work for Packard Bell. Their 33.6 (and above) modem/soundcards supported voice view. We even had a radish team to do support for customers that had those mpdems
Packard Bell is the *only* name that seems to come up when I search for this tech, I suspect they leaned into it more than anyone else. I'm really curious if you heard much about people actually using it, and what they used it for?
What is a "radish team" ?
@@allalphazerobeta8643 Radish is the name of one of the companies making this stuff. Its in the video.
packard bell had those crappy modems that were built into the crappy sound card they had that you could never find drivers for 😒
@@jessihawkins9116 the ones with an actual yamaha opl chip on em are pretty decent. not all of them were crap lol
I think I get this now. "I want a faster horse" when in reality it's "I need a second phone line". This is how my parents solved it too. They wanted to be on-call at work because that gave them both mobile phones for free. So they both had Nextel phones, and we used the Walkie talkie feature all the time. This way, I could be left at home safely while online.
We here in Germany (and Europe) had something called ISDN,before DSL hit us in 1997. ISDN was like "digital dial-up" with two lines, both with 64kb/s. So we solved the simultaneously speaking and surfing problem. And we could bundle the lines and have 128kb/s internet speed. Btw, i like your in-depth videos a lot!
I was looking for that comment. 😉
Same here in the UK. ISDN remained heavily in use for temporary high quality audio links for radio stations and is still around (but disappearing fast).
We had ISDN in the US too. I had it at home from 1998-2000. Before that, I always had a second phone line, all the way back to 1991.
Yep, was marketed for the home user in the UK from about 1996 I think but was obviously in use in business for a long time prior to that.
It came a little late (for the home) but for a year or two it was about the best home internet connection you could have, I remember it was a huge advantage for Quake players since it was very low latency compared to a modem.
LOL we had ISDN here too. It was just pretty expensive, what with it needing to multiplex several individual phone lines.
I used to work for an ISP and there was this VHS vs Betamax period where there were two competing 56K standards (K56Flex (Lucent/Rockwell) and X2 (USRobotics)) meaning we had to invest in two types of access switches (banks-o-modems) with two different dial-up numbers, much confusion ensued with users using the wrong number and complaining of slow service, V.90 couldn't've come sooner!
I totally forgot about K56Flex! Thank you for the reminder!
ISTR zyxel had their own standard around that time as well. HST and Telebit were still kicking around, too. fun times.
My USR X2 was firmware updated to V.90 I believe after some time.
I was just lamenting the barren landscape of my TH-cam subscriptions. And then this showed up! My day is saved!
What a fascinating video. I worked for AOL in 99-01, RadioShack after that and AT&T after that and I had no idea any of this existed. Thanks for all the hard work putting this together.
You're my favorite TH-camr. I get so excited every time you post a video because I know it's going to be exactly my niche of "who even cares about this" and you take a deep dive even if you know you're not going to get it 100%. Your presentation is amazing. I knew tons about how modems worked already and still learned something new.
The design of those ASVD units is delightful! They coordinate perfectly with the case designs of the mid-90s, which so many hardware manufacturers ignored, and their form factor brings to mind the desktop intercoms whose functions they replicate! Just stellar work all around. 💕
Even having grown up around dialing my friends in the early/mid 1990s and seen my share of modems, dialup doom games, and tinkering with AT command sets, all of this SVD stuff is completely new to me. Even voice modems were a relatively rare novelty. I appreciate your nostalgia for all of this stuff, it was an era I don't miss very much at all but I love the historical deep dives.
When I started calling BBSes in about 1991, I had a 300 baud modem on my CoCo2. You could actually pick up the phone and tell the difference between just the carrier signal and when data was being transmitted. As a fourth grader, all of this blew my mind. Computing was so much fun back then! Great video.
You could also tell who was transmitting at the time. Your modem produced the lower tones, and the higher pitch was the modem on the other end. That, or you could flip between the two carrier tones?
Computing is still lots of fun today, it's just very different :)
as far as voicepath or v.61 is concerned, it works pretty much as you described it, it uses a simple 4 point qam constellation and then modulates the resultant qam points with the voice traffic limited to make sure the resultant output still fits in the constellation and decodes with adequate margins, then on the receiving end the data is extracted and then subtracted from the signal and what you're left with is the modulated voice data that can be extracted. it would be interesting to hear how it sounded when there is enough line noise to cause data errors and it subtracts out the wrong data, although a 4800bps it would probably take very loud input sound and lots of line noise to get it out of the margins and actually see data errors
neat trick though and one that i actually saw implemented recently in an engineering school project to encode hidden extra data channels in the noise margin on optical fiber transmissions
didn't fall for the VoIP trick, but also lack the signals theory to figure out where you'd find a big enough hole in the spectrum to reliably stuff in a few kHz. thanks for the elaboration!
That has to be one of the most bizarre and slightly terrifying ways of modulating data I've heard of so far. Modulating the modulated data itself, that just feels... Wrong.
Thank you for shedding light on this manner
@@MrRedwires if you want to feel this feeling even more, CuriousMarc has some videos about how the Apollo moon landing data, voice, and TV was modulated. At points there's phrases like "PCM on PSK on FM on PM"! And modern wifi goes even further with stacked modulations, it's pretty crazy.
Of course if done wrong, one modulation might interfere with another one, but when they're picked carefully it's just like putting a box in a box in a box!
@@kaitlyn__L B&W backwards compatibility in the NTSC color standard quickly came to mind for me also. despite being a heinous design constraint, conceptually & mathematically i agree with the box-in-a-box description... i'm still somewhat in awe of this kind of thing happening decades ago, with comparatively primitive simulation or formal verification tooling to gain confidence in the design, and it pretty much just worked 🙃 (that could be confirmation bias though, i wonder if there's a good example of a broken attempt at backwards compatibility that ultimately steered a product line to a premature dead end)
Great video as always. When I first started watching your content I didn’t know your subscriber count, assuming you were in the millions. Your content is top notch and it’s just a matter of time until you grow. Keep doing what you do, and don’t change a thing!
Yeah, i subscribed pretty early on and it's been awesome to watch this channel grow. Every video is better than the last in my opinion and I think his style hits people at the most basic beginner enthusiast level and provides great info for the more hardcore. Love everything I've ever seen here.
Great episode! I remember how happy I was when I had a modem that could pause all activity on call waiting, let the phone ring and then resume everything after you've reconnected. There was even a little downloader program that allowed for a download to resume after you've taken a call and reconnected to the internet.
As a somewhat pedantic fan of D-sub, I so appreciate the distinction of DE-9 vs "DB-9" at 31:46. Thank you.
i've had people yell at me before for being a prescriptivist about it, and frankly, I see their point. *almost everyone* calls it DB9, and we aren't ordering these out of catalogs - _even some catalogs call it that._ and the distinction is meaningless: did they ever make a "DB" with 9 pins? absolutely not, so who cares if we make that an alias? DB15 is a different story, because they DID make that in both B and E sizes - but nobody really argues that point. honestly, we should probably chill on it, but, I'm not gonna lmao
Wow. I left speechless with the amount of efforts you put in this video.
Thanks
Amazing research on this again. I love the way you slowly ramp up the topic, so we get to know first every detail about the niche situation where these devices were useful. I've never heard of these devices. But then again, the whole dial up thing passed me by, at my parents there wasn't even a wired "landline", just something which was emulated over GSM via a Nokia Premicell terminal, and I didn't have the serial cable to use it's CSD capabilities (I wasn't even aware it had such feature back then).
Thank you so much - I put a lot of effort into making my stuff _highly accessible, without being dull_, and people rarely comment on that specifically so it's nice to hear that my efforts are worth it! I just figure, an awful lot of people probably know _more or less_ why these limitations exist, but not the details, and filling in the blanks helps the whole story come together.
One thing I wish I hadn't missed was early data over GSM. There was a whole ecosystem of products based around it, but in the US, almost nobody used them except traveling businessmen, and by the time I became aware of them. all the systems were long gone.
My parents actually _never_ got a landline, so we never had a phone until cellphones became the norm in the early 2000s. I think I used CSD once, I was desperate due to some deadline, I think it might've been school recruitment or something like that, and my home broadband either failed at the time, or it was before we got one, I don't remember that clearly. Anyway, where I live, until the late 2000s it was conventional wisdom to _never_ use cellular Internet because it could blow your bill in minutes - especially CSD, but early GPRS wasn't much cheaper either. I remember people quite literally having the fear of clicking the Internet icon on the not-quite-smart cameraphones of the day by mistake.
I felt like an early adopter among my group of friends, using Opera Mini via EDGE on my Nokia 3110 Classic ca. 2008 - that's more or less when mobile Internet started to become affordable-ish, likely thanks to the 3G and smartphone revolutions happening around that time. But most people I know were still clinging to the idea that it's prohibitively expensive to _ever_ turn on data for at least another year or two.
@@kFY514 yep, once time I tried GPRS and loading Google halfway used up the £10 of credit I'd just added. Then it gave me an error because I ran out of money. And then I realised it wouldn't be able to pick back up from where it was. Then I knew never to do it again!
The first time I used mobile data was after I saved up for an iPhone, because I made sure to buy 500MB as well. On pay as you go, data you hadn't specifically bought ahead of time was still charged at a ludicrous rate. One time I forgot to get another 500MB after the previous month's had expired (there's a whole other "joy" to go into) and again ate through about £15 of credit accidentally.
That's actually what got me to switch to a contract, at the same price as the 500MB I was buying each month, but with 1GB and a bunch of minutes and texts too.
@@CathodeRayDude You’re legitimately how I’m learning about this sort of tech, so you’re definitely doing a great job making it accessible.
@@CathodeRayDude technically I think you could revive some of these stuff to be tested, eg. there are GSM base station emulators for SDR hardware which are getting more and more accessible for the average person. There are also some specific Motorola cellphones which can be hacked to function as a base station with something called osmocom. But running a GSM network (even if it's being done with very small power and for a short amount of time) is a slippery-slope at best as I've heard :D
I remember in the mid 90s, reading an article or two about the makers of the PGP encryption software in that they made software which would let you make secure, encrypted phone calls over standard modems. It was called PGPfone. I never heard of it being used, nor can I find much information about it. I wonder if the US government killed it off. It would be cool so see demonstrated though.
Side note, an air fryer is an oven. A convection oven, more specifically. The one thing it isnt is a fryer.
I don't believe there's any conspiracy necessary. It's just non-trivial for the technology of circa 1995 to have the processing power to encrypt a phone call, decrypt the other side, encode and compress the voice, and decode and compress the voice all through a pretty-slow connection where you'd be lucky if you got 28kilobits/second. I recall hearing about this at the time, but I'd be surprised if it worked very well with mid 90s technology.
By the time the technology caught up in circa 1999/2000, there were far better ways to talk via encrypted channels over the internet, and not over a modem.
@@stevesether Im not talking about a conspiracy. It wasnt a secret the US government wanted to prevent foreign countries from getting ahold of strong encryption. There were laws preventing the export of certain types of encryption to specific specific nations. I have no idea of the quality of the audio was or if there was latency. However, Im sure it wasnt advertised as something you wanted to call your mom with. It was for people who wanted to talk securely.
There are still legacy systems around that work like you describe at 11:00. I use some of them at work (logistics/supply chain related stuff), which means you are seeing data that isn't actually "live" but was exchanged with another computer very early in the morning before the workday started. Of course this actually works now with a bunch of custom made emulators and compatibility layers over a VPN on the regular internet, but fundamentally its the same 1980s code that in the past a minicomputer at a warehouse would run to make a phone call to a mainframe at 3 a.m. to exchange data.
terrifying, but I completely believe it. I still work with a machine that doesn't support lowercase.
Literal tech witchcraft
I guess you're talking about UUCP. There are TCP/IP implementations of that also.
@@user-fh2fm7vr4m if you think that's crazy, look up how the airlines still operate with booking flights and sharing data cross companies. There's a good video on it out there, probably Wendover or half as interesting or something.
My Dad used BBS for about half a year in the mid 90s before discovering the WWW, then with Dad/Mom both being gamers and Quake coming out, they decided it was only logical to get a second line. The interesting bit though a lot of people overlook is that they also used a service called MPlayer (not the media player). MPlayer was like the first "lobby" for video games as we know today. It would hook up Quake players with other Quake players unlike Gamespy which just showed you active servers. Ontop of that, it apparently had a VOIP-like voice chat. My father told me people would join MPlayer Quake lobbies *just* so they could have voice chat with their friends, and that in combination of running around a 3D world with your QuakeWorld skin (Avatar) made it like a prototype Second Life or VRChat. Of course, MPlayer died not too long after, Half-Life/Counter-Strike gave us VOIP in games a couple years later, and my parents dropped Dial Up as soon as they could sometime in '99-'01. That was the end of Dial Up for us, well except when I went to Grandma's. Took her till the later 2000s to upgrade heh.
Your explanations of the modulation techniques was dead on. I had to go read the spec out of curiosity, but you nailed it. Just some classic nested modulation madness.
I feel old when I see you describing how modems work in contrast to modern network technologies and I already knew all of that. D:
Who doesn't know how call-waiting works, right guys? haha...
Wait, seriously??
@@nickwallette6201 whats a call?
@@frogz who was phone
My ISP provides a VOIP function that provides a land-line. Call waiting works exactly as you would expect.
@@nickwallette6201 everyone when I put them on hold on my cellphone
Great video about technologies I didn't know existed. Growing up in South Africa, we were behind the curve technology-wise. I started using the Internet at university in 1996, but had dial-up at home from 2001 to about 2005. In 2005, my sister Skyped her boyfriend (audio only) in the USA through my Linux VIA EPIA system over a 56k modem and it worked well.
One of the first computer programs I ever wrote (back in the late 70s, and early 80s) was an address book for my TRS-80 Model 1 computer.
I designed and made a little circuit that would use the cassette port to pulse-dial the telephone, at 20 PPS, and not the usual 10 PPS, that normal dial phones used.
I kept all my phone numbers in that program and used it for many years! I even saved the database of telephone and addresses on a cassette tape, that I could load into the computer. Fun times!
Interesting. I wouldn't think most electromechanical telephone switches would handle fast dialing like that. I know the 5XB switch would throw trouble cards if you were out of spec. Maybe you happened to be on an early electronic switch.
You’re right on the lack of testing. i purchased 14.4 modem when it first came out in 1991 and there was a particular command function that was not working,
I was a computer nerd that used procomm plus (they are still around). I called up tech support and they mailed me a pre-programmed EEPROM, and a note to please send the old one back. I swapped it out and it worked no problem.
I wish I would have held on to the modem, it was an ISA board that came in two halves that came apart and it was chunky. Believe me there was no Rockwell modem on a chip on it.
Ah, man, the stuff I missed by being bleeding edge! By the time this stuff was around I was running ISDN, and BRI came with 2 lines. Thanks for an informative look at the stuff I missed. :) Fascinating as always.
I never had ISDN at home, but it was definitely a real solution to the problem at the time. It was just a little too late, because by the time ISDN became widely available, cable modems had come out to largely the same places. I know my family got a cable modem in 1998. In 2000-2001, I actually used ISDN, but in a full-out networked situation, which made it kind of worse than dial-up. We had it for Internet access at the local senior citizen's community center, which was where my high school CISCO networking class was held. There were like 12 of us sharing the ISDN line through Ethernet during the class. Before class started, we would all be trying to browse at the same time and it would bog down pretty hard, but if you snuck in some browsing while the teacher was lecturing, 128kbps was pretty nice at the time.
Man I just want to say that I love your work, I really appreciate the effort you put into these videos.
I always find myself impatiently wondering when you're gonna post something.
Greetings from Brazil!
This is fascinating, I had no idea such things existed, but I can imagine how many hours it took to figure out how to get all these demos going for us, so thanks for this ultra-level nerdiness!
Sidenote: my school district had Dells of exactly the same model as yours until at least 2005.
CRD - I really enjoy your videos. You're a great educator.
Working with radio, i found out that modern RF modules have a weird solution for networking over UART. They simulate packets in software by either making each transmission the same length or adding a start and end code, and then add the "metadata" to the beginning of each packet so that the receiver always knows who sent what. This does mean that all data going in has to be parsed by the receiving computer, but it works.
This has been one of your best videos ever. I love your channel but this content is just perfect 👌👌 Thank you for taking the time to make it! 😊
In joining the chorus, I, too,, thoroughly enjoyed this walk down memory lane. I think I was 19 or so when those modems came out. I think that if I were able to afford the expansion slot version, I would have gotten it and just sat it beside my generic hayes-esque modem of the day.
I was big into telecom for a long time, nice to see the coverage of a very old, but also very important, bit of history.
Kudos, sir.
ISDN sure was a nice thing to have back in the day
Yup we had it aswell. Was pricey though.
so im a telecom person in north america - he mentions CPC, which is something we use here in USA. im 30years old now, and i have only seen TWO isdn circuits my lifetime. I have many many more PRIs. I am also of greek decent, and do go to greece often, so im aware how ISDN works etc, i dont know why, but it seems like ATT was just a lazy company. *shrug*
I remember in the 90s. On of my friends had ISDN at his house.
Not only did he get 64kbits download, but he also had a second line for the phone at the same time.
This was absolutely groundbraking. He could call AND game at the same time!
We had 56k (X2) in 1997, and V.90 in 1998 as soon as USR (3Com) had the firmware to do it. The modems were DSP based, so all software; new standards were trivial to support. (in fact, X2 was a software upgrade) PRIOR to the USR modems, we had much older tech using Rockwell chips that had to be physically replaced to change modulation support. (which we did in 95-96 to go from 28.8 to 33.6.)
It's 330am here. Im omw to downtown Chi to work and struggle bus and the modem yelling made me smile.
Woo hoo! A CRD video to enjoy! Thank you! 🍻🌎❤️🌮
18:51 I love how the modem starts screaming back, gives up and just buzzes after that
Alright, after spending *WAAAAY* too much time analyzing the V.61 ASVD spec, I'm _pretty_ sure it works as follows:
(First, your QAM idea is mostly correct - although by 14.4, modems were using multiple frequencies.)
First, the primer on modems: Models turn 1s and 0s into the screechy noises in the audible frequency range that phone lines transmit. Phone lines, in case you hadn't noticed, aren't exactly CD quality. One digital switching took over, analog phones lines allowed only a subset of the audio frequencies the human ear could hear - specifically the frequencies that the majority of human speech takes place in. So modems have to figure out a way to compress 1s and 0s into audio data. Note: Much older computers like Commodore 64 and Apple II tape ports worked much the same way. They store the data as analog audio. It's just that tapes have a higher dynamic range than phone lines, so they had "more room to work with."
As time went on, engineers figured out better ways to compress data down and modulate it differently to allow for faster modem speeds. Up to 33.6 kbps is still 100% an analog "screechy noise in the audible human hearing range" compression. At the time of the V.61 spec, 14.4 kbps was the maximum speed. Yet for ASVD, you only get up to 4800 bps (4.8 kbps) That's because it used the compression _methods_ for 14.4kbps, but only used a tiny part of the audio "spectrum" available to it on the phone line.
Now, for a side-bar on old vinyl records…
Old vinyl records store audio in wiggles in the groove. But, because of the way it's stored, low frequencies carved into the groove at the same "volume" as high frequencies would be overwhelming to the needle, and a lot of the noise on a record is at higher frequencies, which would drown out the higher frequencies. So the RIAA developed "the RIAA curve" - sound is recorded with the bass lowered, and the high frequencies boosted. Then when played back, the amplifier boosts the bass, and lowers the high frequencies. This makes it sound like it should, while cancelling out (somewhat) the negative aspects of vinyl records.
V.61 does similar to the voice portions of an ASVD call. Part of the voice frequencies are just outright cut out for use by the data (small gaps in the audible frequency range )others are compressed-and-expanded. You have even less dynamic range than a normal voice phone call, but it's enough to be listenable. if your call was local, it would probably be about the same quality as an international phone call over normal phone lines. The modem would compress the dynamic range of your audio stream (not digital compression, purely "squashing the signal") and send it analog over the line, then the receiving modem would decompress the dynamic range, introducing noise, but making it listenable.
At the same time, it would cut out some audio frequency range completely from the "audio out" (speakerphone/headset jacks) because that portion was used for data, using standard bandpass filtering techniques. It's the same basic idea as a "DSL filter" would later use to get the "DSL frequencies" off your voice phones.
Internally it's a splitter - one part goes to the modem, with a bandpass filter that can be turned on or off to hide the voice frequencies, the other part goes to the "audio out" portions, with a bandpass filter to turn off the data frequencies. If you did a "data only" call, the audio out portions would just be completely disabled, and the data modem would get the full use of the line; if you did an ASVD call, it would split it up thus the modem had to be much slower.
This theoretically could have been done with older modem technology, too; but it would have been so slow as to be truly unusable with the older modem specs.
Imagine a commodore c64 with 56k6 loading speed from tape. That would be faster than a C64 diskdrive.
WHAAAAT?
That's mind-blowingly cool! 🤯😎
Also @ 18:52 you literally killed a serial/networking connection by just yelling?
Never thought I'd see that happen, LMAO!
I would love to find someone who talks about the components of civil aircrafts like you do about 90's/2000 technology stuff.
avionics has a lot of Easter eggs
It's really weird. Everything you show I lived through. I was extremely happy when every bit of technology you show was upgraded or replaced. Yet, your videos are hypnotic! Nice one!!
I fix up 1930s to 1960s valve radios, so my parents think exactly the same about my hobbies.
26:00 v90 and v92 require compatibility with the equipment at the exchange (connection needs to be PCM not analog to achieve speeds higher than 33600 bps) which suggests there's some additional signaling required on the exchange side to make mixed voice and data calls possible.
Very cool episode. Love the touch with picking up the phone sound during testing. 🙂
I got lost in your funny words magic technology man but I'm impressed with your musical taste on display at 53:30
18:55 I love this. Your reaction is delightful.
I would love it if you did a series of videos where you tour the Seattle Telecommunications Museum :)
i never searched for it because i never had the curiosity for it, but i finally learned the purpose of that sound. thank you so much for making these videos of old harware. love it so much.
32:30 I remembered that just today. I had looked up a 'phone number on my PC and just for a moment I remembered the good old days of being able to dial it up from the PC. Not sure how that was done now, but I remember it worked.
I love videos of solutions to problems in the 90s. My dad had a second line installed for his business. The sound of your voice speaking in ASVD the mic at the end would make a cool song mic audio.
Oh boy hope this covers Mplayer! Was way ahead of its time
When you shouted into the phone, as soon as I saw "No carrier" I burst out laughing before you even did. That was hilarious. Keep up the awesome work. 😂
We had two phone lines in the 90's and early 2000's so dialup wasn't a huge issue when we needed to make a phone call while dialed up on Prodigy or AOL (America On Line) lol
Perfect video! I can imagine the tons of research and the hours of fiddling with parameters to make different brands talk to each other...
Even if I had owned such a modem, I would have needed your TH-cam channel to know how to use it! Congrats
Wow, the first sound I heard in this video really was DSPOSIT1 Zombie sight #1 Sound Ideas Series 6000 "Animal, Creature - Large Animal Growl 02"
Where's The Lie
I have to say, I was queuing up video to watch , I dropped everything else when I saw that this was out. I cant get enough CRD!!!
14:31 claiming all modems are identical is very much misinformation. Yes, they all use the same modem chip, but the cheaper ones had worse support hardware and filtering on board. I was on a very marginal phone line, and fried a ton of modems. I assume this is due to our local carrier having issues with their ringer and pumping excess voltage throught the lines, but I can't test that now. Neeedless to say, I had modems dying after ~2 months, the longest lived were US robotics around 8k-10 months. I looked over my pile of dead modems, since I'd continued to save these for the relays on them, and the more expensive modems had much more extensive filtering circuitry., A more expensive modem had better filtering circuitry in front of it, which not only provided better connection stability but also helped insulate the modem from poor phone lines.
Yes, all of these modems circa 98-03 use the same chip, but there were huge differences between them and a better quality modem would provide a better connection. The analog circuitry in front of the chip was the major differentiator here.
That was very informative! Loved it! I lived through that era. It was a plesant stroll down memory lane, hearing those modem noises! Wonderful! Thanks.
I ran a FIDO node in the 90s (bit late to the party; didn't bother me as a teen); you wouldn't believe the compatibility issues you'd have with even same-chip-different-oem modems at times 🙈 (US robotics and zyxel did well to me; ELSA was fine too, but a bit picky -- and quite pricey at the time)
*edit* also, 16550 vs 16540 UART vs that cheapest of onboard serial controllers your big-box-store PC might've come with back then? quite a margin.
oh i can't even imagine, hahahaha, gosh it had to have been a mess
@@CathodeRayDude keeping that BBS running, with my dad going mental re: phone bills? a mess, it was, indeed :)
(but also quite a learning curve: when IP, broadband, "se internet" became a thing, I already had a clue what that decentralized networking thingie might be about. served me well for years.)
I remember voice chat in the late 90s was "being in the same room".
I was lucky that my parents got ISDN with four channels very early, because my dad often worked from home. He was a very modern man. That was in Germany during the mid 90's and that was really really expensive. Each ISDN channel is basically a digital phone line. They also could be combined. In 2000 my parents got T-DSL with around 700-800 kbit/s. I used that as an opportunity to download pirated movies and music CDs, burned them on CD's and DVD's and solt them on the schoolyard. Made a lot of money for a 14 year old. My parents were upper middle class, but they were pretty strict not to spoil me with money, so i worked for it. When i was 16, i had enough earnings to afford a beast of a PC. A brand new Pentium III 800E system with four DVD Burners and massive hard disk. I believe i had two 40 gigs and external firewire ones. What a time that was. Nowadays i still own a business, but with more legal methods ha ha.
We didn't get ISDN until 1999, but having the option to either use both lines to get a whopping 128 k connection or use a 64 k connection while still using the other line for regular phone calls felt incredible at the time. Of course that was soon surpassed by DSL (we got 512 k HDSL in 2001), which could use full speed regardless if you were on the phone or not. Come to think of it, our connection speed seemed to double every couple of years back then. The speed internet technology progressed was remarkable.
Excellent video!
Loved when you started talking about voice coding towards the end. I find it particularly fascinating on land mobile radios.
Something to note, error correction comes in a few flavors, you can use ARQ (Which requires a retransmission like you mentioned), or you can do forward error correction (FEC). You add redundancy to your message to make it more tolerant to noise, whatever source that may be. Now that being said, that slows data rates as well. But with voice packets, you usually don't care which you also mentioned. So in slighlty more modern systems, your best bet is to use some type of datagram scheme and just do FEC on it to the best of your abilities after characterization of the channel and more often than not, the voice sounds intelligible enough even in the harshest of conditions. The fun part is when you get into the wireless sphere and start analyzing energy per bit to noise power spectral density and how the FEC impacts it as RSSI decreases. Fun stuff 😀
I'm going to start saying QAM like wham from now on. It is normally pronounced like "kwom" rhyming with ROM. But I like yours better. Anyway, these modulation schemes are used all the time in rf engineering, so that is where you go to learn about that sort of stuff.
I have vague memories of SVD being a thing back in the modem days, but I never actually tried to use it. Watching this, I'm glad I didn't.
Also, I find it amusing that the modem being called tiny is larger than the 33.6K modem we had, which was small enough that it had a Mini-DIN port instead of DB-9 for its serial connection.
I got a phone call halfway through this video, and it shows how much I take modern multitaksing for granted when youtube just shrunk to a little box
Hey Gravis! Thanks. This was very good.
For that mic input issue, sounds like there may be a bad capacitor in the unit causing those issues. Would probably be a fairly easy repair if you want to tackle it.
that's a good point, it didn't occur to me that it could be tired old parts
@@CathodeRayDude bad electrolytic caps were notorious right around when that unit would have been made. Should be easy to spot if the unit can be opened easily.
I love the nice touch of "recording" the phone call to play in the video
DOOR STUCK! DOOR STUCK!
I BEG YOU
Man, normally I get so bored with long-form videos.. but I totally can just sit and watch your shit for hoooouuurs. So much interesting shit. A lot of which i already knew, but it's just so cool to see it all in action and stuff.
erratum: at 14:00 you say "US Robotics" when holding a Hayes branded modem... AFAIK the former never acquired the latter. ;)
lmao yeah i completely missed that. it was a plain old flub, wasn't in the script
Absolutely incredible stuff! So grateful I don't have to support any of this stuff in a work environment, but so neat to learn about.
The intro to this video made me extremely nostalgic. When I was younger I my parents would routinely take me to our farmhouse to get away from the city for a couple days, much to my dismay as a tech enthusiast in the early 2000s, and the only technology I had access to was my SNES that I had to beg to bring over every time. One day I was playing Super Mario World and reached the furthest point in the game I had been to at the time: the ghost castle where you're supposed to solve a puzzle to progress (IIRC you have to enter a specific P button, otherwise the stage loops indefinitely). As a kid I had no idea how to solve it, no internet to look the solution up and no telephone lines to ask my city friends. The solution I found was to walk 45 minutes over to a friend's house, ask them how to solve it (as well as how to perform the super moves in Final Fight 3), and walk back. That specific day always stuck to me as a funny point in time where internet was already a thing, but not commonplace enough to be convenient. Very retrofuturistic in a sense.
When I saw the intro, I thought it was a gimmick and wanted to chime in and say our (monopoly) telco had a tiny application that could indicate a call waiting if someone would call you (I think even with caller ID), but it's obviously far more inferior to the solutions you display; I'm jealous at your setup :)
I do remember vividly the day I wasn't allowed onto the internet since my brother was paying for it, and blowing into the handset (instead of screaming) to disconnect while not triggering any suspicion. I also ran into the problem you mentioned at 19:00 and I never got the game to connect, with spite up to this day.
Thanks for making this awesome video!
i loved this video so much (i mean i love all your videos ngl), its so cool to see how chaotic the landscape of anything was in that era, i like deep dives into all the weird shenanigans of a "standard", i now feel like i understand how modems work and what could have been and i know i'll rant about this several times in the future lol so thank you for this
18:50 Bahaha that was amazing ! You’re really amazing 😂 screaming in the phone making the connection dropped it’s super funny!
Just a caveat on Hayes commands when picking up an ongoing call:
You can issue an ATO comman to start negotiating the carrier as with the ATX0 + ATDT method, just easier and without messing around your modem's dial settings.
I worked for a Dutch telco at the time. We were working on using DSVD for doing actual normal phone calls during a modem call. Our proof of concept forwarded your normal phone number to the modem on our end that would put the phonecall inside the DSVD session if you got called, and if you picked up the phone you would have gotten a normal dialtone.
This was linked to the PPP/RADIUS login, so if you’d dial up with someone else’s user you would be able to use their phone number, which was a funny side-effect.
ISDN and DSL happened shortly after, there was more money in both of them. So it was cancelled. Also, around 90% of the market wasn’t using our internet service but the ones of dedicated ISP’s, not the phone companies one, so management saw no money in it, which made sense.
“Two of them” has become my favorite meme because of this channel. I love seeing it in CRD videos :)
The beauty of Fourier series! Summed signals make our world possible, it’s beautiful
Haha, i just checked this channel about an hour ago to see if there was a new episode, now im home and see the blue dot! What a coincidence!
I'm at minute 6:00 and so this may be premature to share, but there used to be a service called CallWave which used the "call forward on busy" feature that almost all phone companies had available for free to forward calls to their toll-free number. When this happened, I guess the phone company would also forward the original destination number and then an app would pop up on your PC, tell you who is calling, let you screen the call like Google Voice does, and press a button to answer it if you wanted. If you did this, it would immediately hang up - this was pre-V.92 - and then their toll free number would call you with the original call. They even had some adapter that would make your actual phone ring (we never bought it). It was only $4.99/mo or something like that and I found it life-changing. I could download something that would take days with the download manager I think had the word Gator in its name and not worry about missing a call and getting in trouble. I'm sure that the hour I have left will reveal something absolutely insane that I never knew about, but I wanted to share CallWave in case that's the thing and I'm not the only person on the planet to have ever used it :).
And I'm done. Definitely never heard of either of those technologies, but as Tay Zonday mentioned, I did use Roger Wilco as well as MPlayer (which apparently is from the same people that made XBAND, but I never knew back in the day).
CallWave was the thing that solved the mom yelling at you for playing Quake online all day while she missed calls problem for me. And it's such a simple approach that I'm still surprised that I'm among the 5 people that remember it. You'd think that if you were a major ISP like AOL or AT&T WorldNet that you'd add this technology to your service just so you can one-up the competition, but it never happened. They had call forward on busy, they had software-controlled phone switches, they had all the call metadata. But as far as I know, only CallWave did anything with any of this. I should also add - they could detect if the incoming call was a fax call or not and if it was, it would just auto-answer and let you download JPEGs and TIFFs of the fax pages once it's received and processed. Like all stuff that AT&T could do for essentially free, they just needed the system tray application.
I'm heartened to see that I'm not the only one who remembers the promise of DSVD. I was working at Radio Shack in 95-96 and remember reading about it, probably in Nuts & Volts, or a similar magazine. I figured that the problem with needing two identical modems would be resolved through broader software adoption later down the line as the tech became the standard.
As for me, I just barely got into the 56k era when Road Runner came to town in 1998, which I was the first in town to have.
I'm happy that my connection in 2002 was fast enough to handle Teamspeak while I was gaming, I lowered the bitrate though to improve my latency.
This is next level, I would have loved this before the year 2000.
comprehensive presentation with generous amounts of humor and pop references.
The usage of DOOR STUCK as the visual example when mentioning in-game voice chats as VoIP put a stupid grin on my face.
Your laughing scene summed up every parent/guardian ever (at least on the inside) in the 90s who deliberately did such things to disconnect the modems so that they could use the phone 🤣🤣
That brown and beige phone. You’ve got my sisters first phone. It’s in my laundry room now.
OK DSVD that is pretty cool. I knew that there was a bunch of other features floating around at the time but I just struggled in the era to get a modem with voice capability that actually routed the voice capability through the sound card not out to separate jacks. I am myself enjoyed the speakerphone ( and answering machine ) software that was built into my Compaq. And Willow phone a neat little under a megabyte Windows software that made any voice capable modem a speakerphone and only worked on like three computers I've ever had.
And the way the internal ones actually use the phone with a ring signal is pretty darn clever if only that was made the standard. Most of the modems I have used interrupt the secondary device that seems like a really neat use for it couple it with some sort of car waiting and someways I wish we had another 10 years of standardizing and the modem Field where we could get to one chip that supported everything.
When you pull out an empty shell after the Cirrus Logic chip, hilarious. Thanks for including that :P
I really really appreciate your videos. They scratch just the right mental itch.
FunFACT: Television Close Caption line 21 encoders and decoders, still use dial-up modems to caption TV shows and News every single day...
They even still use 1200 BAUD, 7 bits, Odd parity, and 1 stop bit. That is STILL the industry standard for captioning systems!
Fantastic presentation as always, you do a wonderful job of getting information across. Thank you.