Closed Captioning: More Ingenious than You Know
[DTMF signals from the beginning of a Disney VHS tape] [if you remember those, you’re my kind of awesome] As you’ve gone through life as a high quality video connoisseur, you’ve no doubt seen the little toggle for turning captions, or subtitling, on and off. Here on YouTube, it’s right there. Or there. Depends on how you're using it. By hitting that button, your device will start displaying the text metadata embedded in the video stream that either YouTube has auto-generated, or that super awesome content creators like myself have put in so that you don’t have to simply watch a stream of unpunctuated words flow by like the world's largest run on sentence that never ends and which often contains incorrect words because the captoning system isn’t quite perfect and that while useful as a tool could really use… (jokes sometimes end up here, too!) But did you know that even the lowly VHS cassette often contains captions? And that a VCR as old as this dinosaur can make them work? That last statement is perhaps a bit disingenuous, but you’ll see what I mean shortly. Many VHS cassettes will have a little “CC” mark, or sometimes this mark. This tells you that the video was encoded with closed captioning. But, how? First, let’s go over a bit of terminology.
The reason why these are called “closed captions” is that they are not normally visible
While this terminology is a little antiquated at this point, “captions” implies that there is actual text baked onto the video, for everyone to see. Sometimes these are called open captions, though that’s likely a retronym. The trouble with these sorts of captions [sound of glass breaking]. So, closed captioning hides the captions until and unless they are requested by the viewer. Believe it or not, the first television program in the US to be captioned at all was "The French Chef”, broadcast on PBS in 1971. Those were open captions, but nonetheless that was the first time that a US television program was accessible to the deaf. This is really surprising to me, because you’d think that one of the great benefits of television over radio was that non-hearing people could get some benefits from it, and yet it took until 1971 for anyone to take advantage of the visual component for this purpose. My how we take our senses for granted. Anyway, according to the National Captioning Institute, the idea that what would become closed captioning in the US started with an experiment by the Nationals Bureau of Standards and ABC television. They were attempting to broadcast the precise time using non-visible parts of the television signal.
That idea never came to fruition, but it did lead to the idea of encoding text for the purpose of closed captioning
In the early 1970’s, more testing was done on various stations around the country, and in 1976 the FCC officially decided that Line 21 would be reserved for closed captioning. Eh, to explain what the means we need to go a little more low tech. Woah. Suddenly I fell kind of… interlaced. Great. This again. Well, at least I can say, “Look ma! I’m on TV!” An old-fashioned CRT television like this one is a pretty dumb device. There are no logic circuits in here, just a bunch of frequency oscillators, deflection coils, and other analog goodies. All around the part of the picture that you can see are blank sections that are used as triggers. I’ve made an earlier series on analog television that you can find if you’d like more information, but simply put, the image on the screen is made of one continuous line which is broken up and stacked 525 times per second.
Other countries and continents used different standards, but we’re talking about the Freedom Captioning System here so we’re sticking with good ‘ol NTSC
The TV knows when to break the line up because throughout the television signal, low-intensity pulses (called the horizontal blanking interval) cause it to reset the drawing process, pulling the line back to the beginning, and starting from there. But the TV also needs to know when to start the next stack of 525, so another trigger, the vertical blanking interval, is used. This trigger is simply a group of lines that have no information in them at all. To keep you from seeing these triggers, the picture tube is overscanned so they appear outside the borders of the screen.
Closed captioning simply uses one of the unused and unseen lines to digitally encode text
This worked because a TV doesn’t need that many blank lines to reset the vertical deflection--in reality it just needs a few. So sacrificing one of them wouldn’t hurt compatibility. The first demonstrations of this experimental system occured in 1972, and in the following year Washington DC’s public television station WETA started doing test broadcasts with the data transmitted on line 21. In 1976 the FCC decided to standardize that, and from then on Line 21 was reserved exclusively for closed captioning information.
Developments to the captioning technology continued to be made throughout the 1970’s and early 1980’s
Much of the work was done by engineers at PBS, specifically at WGBH in Boston, with an important development being the editing consoles used for inserting closed captioning data into pre-recorded material
By the end of the 1970’s, closed-captioning advocates decided that they should form a nonprofit group, with the goal of creating the standards for the system and encouraging larger broadcasters to get onboard
So, the National Captioning Institute was formed in 1979, and in 1980 the first television series to be fully closed captioned was broadcast, and this also appears to be the year that the first closed captioned home video releases made it onto the market.
This symbol here is actually a service mark of the National Captioning Institute
I always thought it was just a generic mark indicating closed captioning, but apparently it meant that the NCI is who actually created the captions for the program. So that’s pretty neat. This mark, designed by Jack Foley of WGBH, was placed in the public domain and thus became fairly standard as well. Now that we know a bit more about its history, let’s take a closer look at how it works. But first, I gotta get out of this thing. There’s no place like 4K! There’s no place like 4K! There’s no place like 4K! If I mess with the vertical synchronization of this television, the image will roll and the vertical blanking interval can be seen. On this line, which happens to be line 21, you’ll see that occasionally, it bursts into a black and white scramble of dots.
This is where the closed captioning is
We’re seeing this visually, as after all the picture tube is still gonna react to it as if it were visual information. But don’t think that it’s being read literally like a barcode--though it looks like one, what this really is is a tiny portion of the signal going high-low-high-low. In other words, it’s a digital datastream. But how do we read that and turn it into captions? Great question! We need a CLOSED CAPTIONING DECODER! Luckily I have one right here. This is the National Captioning Institute’s very own TeleCaption 3000. Glad they used 3000 because it still sounds futuristic! These were actually built by Sanyo, and they weren’t cheap. Costing around $200, they cost more than some basic televisions. There’s a great article from the Chicago Tribune written at the time of this unit’s introduction which has lots of cool info, and you can check it out in the description if you’d like to learn more. Although they were expensive, they did at least include some nice features, as we’ll see shortly. This device is able to read the data from line 21, and generate text on-screen. Remember, that data looks like a bar-code, but it’s processing that as a datastream because… OK now that I think about it… all barcodes kind of work that way, at least with the laser-type scanner thing with the spinning mirror--but, anyway! It’s reading that data. Alright. That data encodes text and timing information, and after determining what it has to do, the device draws blocks of text on top of the screen, providing a black background to ensure it’s visible. Now, this may seem pretty simple, but this device actually has a pretty sophisticated job to do. See, it can only only look at one channel at a time, so it needs to have its own television tuner built-in. Then, it has to spit out the same picture to the TV, while altering it to contain the captioning. And it has to do that in real time with technology of the 1970’s. Of course, this device isn’t nearly that old--it was made in 1989 and by that time the electronics weren’t so crazy-advanced--but the way the NCI designed this device is awesome. Take a look on the back and you’ll find a coaxial input from the antenna, and another output to go to the TV. You’ve got your standard channel 3 or 4 selector (used in case your local market had a station broadcasting on either one of those channels), but what’s really clever is this switched outlet on the back. Since they had to integrate a television tuner and it would be outputting the altered signal as if it were, say, a VCR, they designed it to give it a little bit of extra functionality. If you plug your TV into the outlet on the back, then you can use the TeleCaption 3000 to give it a remote control! See, the TeleCaption has its own remote, and when you turn it off, it kills the power to that outlet. So, if you leave your TV turned on, then you could simply turn your TeleCaption on and your TV would automatically come to life. Better still, they added a volume control, so if you left your TV on, set to channel 3, and with the volume on high, you could control everything--the channel, the volume, the TV’s power, and of course the captions--with this remote. I really appreciate this thinking on the part of NCI, as even really old televisions with rotary knobs could now be operated completely via remote control. Nice going, NCI! One slightly strange thing about this device is the inclusion of adjustments for the background as well as the text. I wasn’t really sure what to expect from these, but apparently they allow you to make the background lighter and the text darker. I’m not really sure why you’d want to do this, but I suppose a grey background is a little less intrusive. And maybe some old sets would be affected by the black box and skew the image geometry a bit. No matter the reason, options are welcome. Buying one of these decoders wasn’t your only option, however. Some high-end VCRs were available with built-in closed captioning decoder circuitry, but apparently this wasn’t well promoted and they never sold well. In any case, this machine will work with a standard VCR, and given that it creates a value-add in the form of a television upgrade, I think it was among the best options. So let’s look a bit more at the captions themselves.
One of the great things about the closed captioning protocol is that the text doesn’t just get plastered anywhere on the screen
Its location could be defined. This meant that the text could shift left and right to give an indication of which character is talking. Which is pretty handy. And the text could even be placed up high so that it wouldn’t obscure other visual information on screen. It’s pretty impressive that this encoding was thought out for a captioning system with roots in the 1970’s, and it’s also frustrating that YouTube captions don’t allow you to do this. That’s right YouTube, this old VCR and this box are more sophisticated than your captioning system. Maybe work on that. Now, this may look like it’s limited to all-caps, but in fact the captions do support mixed-case. It just tended to be used for descriptions rather than dialogue. Not sure why, and there are exceptions to that rule, but the default seems to be all-caps. Of course with today’s context, it seems like everyone is shouting. One of the more surprising aspects of this system was how robust it was. This is a pretty old VHS cassette with a recording made from live TV, done with not the greatest incoming reception, and recorded on the low-quality SLP speed. The captions still come through fine, and with few errors.
Even in areas far from the transmitters, the closed captioning system would still work reliably
Of course, for many years, the likelihood that a program would actually contain captioning would be pretty slim.
It took a lot of effort from the NCI and other organizations to persuade broadcasters to start using it
But over time, it began to spread. Every home video format began to support it. You can even find CEDs with closed captioning, like this copy of Robin Hood bearing the NCI’s service mark. CEDs? You mean those things from RCA? That really failed format that was video on a vinyl disc? Is he gonna do a video about it sometime? Yes.
Another big development was the real-time captioning console, which allowed for captioning of live and unscripted events
Stenographers were hired to furiously spit out captions of newscasts, sporting events, and other unscripted programs
Of course, accuracy wasn’t exactly perfect, and one of the downsides was that the captions would often appear with a fair delay, but hey, it was better than nothing. But the real windfall came in early 1991 when Congress finally got around to passing the Television Circuitry Decoder Act of 1990.
This required all televisions of screen size 13 inches or greater to include the ability to display closed captioning
The law went into effect on July 1st, 1993, and from that point on the important accessibility feature became… more accessible. This TV was made in 1994, and sure enough--closed captioning can be turned on via a few button presses on the remote. When DVD rolled along in 1996, suddenly closed captioning got a whole lot more interesting. And for a number of reasons. First, DVD players natively support subtitles. Subtitles in a DVD work very differently from the Line 21 closed captioning, as they are stored on the DVD as series of images. When you toggle them on, the DVD player is stamping semi-transparent bitmaps on top of the video, and because of this many languages could be supported without needing to have support for their character sets. This also explains why the subtitles can look so different from DVD to DVD. But--DVD still supported line 21 closed captioning. That’s pretty remarkable once you realize that in order to do that, the DVD player has to do its own line 21 encoding from metadata on the disc.
The video files on the disc don’t contain anything outside the frame, so the DVD player itself has to generate the vertical blanking interval and shove that captioning data into line 21
This just happened in the background, enabling those who relied on captions to use their existing equipment if they preferred. This was also useful in a scenario like a hotel, where a central DVD player might be broadcasting the same program to many televisions. Then only those who wanted captions would view them. Of course, now I needed to know if Blu-Ray supported line 21 captions. It looks like it doesn’t given that over HDMI there is no vertical blanking interval, and being an HD format what’s the point? BUT! The PlayStation 3 has a composite output, so let’s see if anything happens. Oh. And just as another test--would the PS3 send Line 21 captions out with DVDs? Looks like it does! Look at Sony being all backward compatible. Wait… Did you know that closed captioning systems exist for movie theaters? I’ve only ever seen these systems in person at theme park shows, but the system is very clever. At the rear of the theater is an LED dot matrix display. This display is showing subtitles for the film...backwards. The idea is that, if you need captions, the theater will provide you with a rectangular piece of plexiglass on a stand, and by placing this in front of you, you can aim it to reflect an image of that display into your eyes, either below the screen or on top of it if you choose. And since the reflection reverses the image, the previously backwards text appears forwards.
The most common of these systems is the Rear Window Captioning System, also created in part by WGBH Boston
Jeez, those folks are nice. Anyway, if you’ve ever looked behind you and saw a display board with backwards subtitles--now you know what that’s for.
Online streaming services seem to handle closed captioning differently
One thing I’ve noticed with the way Netflix handles it, is that the captions can be forced on. I’m pretty sure Blu-Ray can do this, too--and maybe even DVDs. The practical upshot of doing this is that any on-screen text, such as can be automatically translated into a different language if a different audio track is selected. That’s pretty handy.
One thing that the hearing impaired can look forward to is real-time captioning
I’m sure this is already in development, but this would be a socially acceptable use of something like Google Glass. With speech-recognition technology as advanced as it already is, I suspect it will not be long before a simple glasses-like device enables real-time captioning. Oh wait, of course it’s already a thing and you can find some links in the description. But before I go--did you notice that there were three modes on this closed captioning decoder? There’s TV, which does nothing. There’s Caption which does… captions. And then, there’s Text. What happens if I turn that mode on? Interesting. A black box covers almost all of the screen. What could that be for? That’s right James. You’ve waited so long. It’s coming. Soon, we’ll talk about Teletext, the much more advanced sibling to closed captioning that much of Europe enjoyed, but which never caught on here in the States. So, uh, stay tuned for that. Thanks for watching, and I hope you enjoyed the video! I think that closed captions are now something we very much take for granted, and the hearing impaired are very thankful that we do. Though it took an act of Congress, I’m glad that we finally decided to give the hearing impaired the same access to television that the rest of us have. As always, thank you to everyone who supports the channel on Patreon, especially the fine folks that are scrolling up your screen. With the support of people just like you, Technology Connections has gone from my hobby to, well, this! And I’m very thankful for that. If you would like to support the channel and get perks like early video access, behind-the-scenes stuff, as well as other Patreon-exclusive content, please check out my Patreon page. Thank you for your consideration, and I’ll see you next time! ♫ insufferably smooth jazz ♫ I accidentally discovered something while shooting B-roll for this video. Earlier I made a video on a little something called Macrovision, an analog copy protection scheme.
This deliberately screwed with the vertical blanking interval to confuse VCRs, and indeed it created a few problems for my capturing process this time
To help make the caption data easier to see, I tried playing a Laserdisc which is known for not having Macrovision. And now, I think I know why. Laserdiscs are chock full of other barcode-like things in the blanking interval, and judging by how they are jittering around, I’m thinking this is what the player looks for when skipping tracks or even just looking at the timecode. And, assuming this is true, this would explain why Macrovision never made it to Laserdisc. Laserdisc had already done its own screwing around with the vertical blanking interval--in its case to create the control system for the format--and trying to put Macrovision in there would probably break it. And while we’re on the subject of closed captioning, I’ve finally turned on community subtitle submissions for the channel.
If you speak another language and are willing to help out by translating my videos into that second language, please check out the link below with instructions on how to do so
This would be really helpful to me to give exposure to the channel in other countries, as well as just being pretty neat. If you decide to do it, I promise you’ll get your very own Unofficial Official Imaginary Badge of Complete Awesomeness. You’ll know it’s true so long as you believe. Hey, for everyone who used these captions--I'm sorry they were on top of other captions so frequently. I REALLY REALLY wish that YouTube would let me move them around like.. you know... like we could in 1980. That'd be swell.
Tags: youtube live captioning
Read more help:
Watch, read, educate! © 2021