Why does closed captioning make these obvious mistakes?

[RivkahChaya]

@Aspenglow

I don’t completely understand it, but I can shed a little light.

Transcription for captioning is done by real time transcribers, like those used in courtrooms.

Usually, but some shows, like local shows, might use electronic transcribers.

And yes, sometimes the reception does get screwed up, although it used to be worse. I remember at Gallaudet, when I was there in 87-88, sometimes we’d get several seconds of black panels where the captions should be. So yeah, your TV could be screwing up and not differentiating between ’ and u for some reason.

@Keeve

If only it were so, then these spelling errors would be the only problem. And for live broadcasts, I suppose it is indeed the major problem. But from what I see, the captions for prerecorded shows and movies are taken directly from the script. If an actor goes slightly off script – use a synonym, changes the word order, adds a word or two – you don’t see it in the captions.

It’s worse the other way. Take the show Switched at Birth: from what I can tell, the Deaf actors do their own translations, and they are translating into ASL for the most part, but the script has been written without regard to the fact that some of it will be translated into another language, so the Deaf actors lines are full of metaphors and idioms that have to be interpreted or paraphrased rather than word-for-word translated. But the script as it was written is used in the subtitles for the hearing audience, and it is wonderfully clash-worthy. I have to deliberately avoid looking at the subtitles, or it interferes with my understanding the signs, because they are so different most of the time.

@wguy123

Not exactly related to the OP, but why do a lot of CC’s censor curse words? The show being captioned says the “bad” word but the captions change that word to a less offensive version. Are hard of hearing and deaf people more sensitive than the rest of the population? Or, are the folks doing the captioning delicate flowers that can’t handle typing such obscenities?

If it’s a word that is allowed (like “damn,” or “hell,”) then it goes through, because I’m pretty sure I’ve seen those. If it’s a word that a short delay would censor on a live broadcast, but it didn’t get caught on the audio, the captions may still be required to censor it, since they have a de facto delay, is all I can guess, because no, Deaf people are not more delicate, and if you watch an R-rated DVD with captions, all the f-words and s-words are there; I assume they are there on an HBO or Showtime movie or original program. Maybe someone with one of those channels can check and report back.

Interpreters can be fired if they get caught altering content to censor bad words (there was actually a case once where a religious nut interpreter claimed her freedom of religion was being interfered with when she was censoring a public school student’s classes by changing “damn” and “hell” in an English class discussion of The Scarlet Letter, and got told to stop it-- she lost), so I would think the same thing would happen to transcribers. If they are trained as court transcribers, they are probably already vetted for this kind of behavior (ie, as people who wouldn’t do it), because I assume that you can’t get away with this when you are transcribing a court case.

[Chronos]

A lot of people who use closed captioning aren’t completely deaf, but are hard of hearing, often as a result of old age, and old folks might be more sensitive to foul language…

except that you’d think that, if they were, they wouldn’t be watching the sorts of shows that use that language in the first place.

[bump]

So am I right in assuming that the captioning service isn’t in any way linked to the actual show/commercial being watched?

In other words, if I was to watch NCIS tonight, there’s not an official closed-caption stream being broadcast by CBS at the same time? It’s just some transcription mechanism or person doing it on the fly?

[wguy123]

@RivkahChaya

If it’s a word that is allowed (like “damn,” or “hell,”) then it goes through, because I’m pretty sure I’ve seen those. If it’s a word that a short delay would censor on a live broadcast, but it didn’t get caught on the audio, the captions may still be required to censor it, since they have a de facto delay, is all I can guess, because no, Deaf people are not more delicate, and if you watch an R-rated DVD with captions, all the f-words and s-words are there; I assume they are there on an HBO or Showtime movie or original program. Maybe someone with one of those channels can check and report back.

These are not live shows…they are filmed broadcast shows. I’ve noticed on several series watched over either Netflix or Hulu. If I remember correctly, Breaking Bad was a common one where they would put in asterisks for the curse words. Another show which I cannot recall would actually put in a substitute word which was comical (along the lines of substituting goshdarn for goddamn).

[Keeve]

@bump

So am I right in assuming that the captioning service isn’t in any way linked to the actual show/commercial being watched?

In other words, if I was to watch NCIS tonight, there’s not an official closed-caption stream being broadcast by CBS at the same time? It’s just some transcription mechanism or person doing it on the fly?

Think about it. It’s not coming from your TV, and it’s not coming from your cable company, and it’s not coming from your dish antenna. And it’s certainly not coming from Martians who have hacked into the transmission broadcast. It most definitely is coming from CBS. The question is what service is CBS (or, actually, the producers of the particular show or commercial that you’re watching) using to get the captions.

[sbunny8]

My spouse is trained as both a transcriber and a captionist. Those are two different jobs.

Transcribing is generally done in real time and uses software that allows the typist to type very quickly with shortcuts, something like 400 keystrokes per minute producing up to 160 words per minute. In many cases, the typist only types in the consonants and the software figures out what word they meant and fills in the vowels. Contrary to popular belief, the goal usually isn’t to produce a verbatim account of what was said but rather a meaning-for-meaning description of the conversation. But I believe the standard may be different for courtroom reporters. Transcribers in a classroom, for example, will provide a real time stream which appears on the laptop of a student who is sitting in one part of the classroom while the transcriber sits in another part of the classroom. In some cases, the transcriber will have a chance to go back over the file which was created and clean up the typos before submitting a final version.

Captioning, OTOH, is generally done from video, in a much more deliberate fashion. The captionist may take an hour to properly caption just 10 minutes of video, playing it back over and over at varying speeds, trying to make sense of what they are hearing. It is fiendishly difficult to caption an entire hour of video without making at least one mistake. Sometimes captionists make typos. And sometimes the people in the video talk so fast that what’s being said literally won’t fit on the screen. And sometimes the captionist simply can’t figure out what the people are saying. Amongst the reasons for not being able to hear what someone on the video is saying are such diverse elements as: use of technical jargon, hard to understand accents, variations in volume, and background noise. The sad part is that, once the captions have been submitted to the producer of the video, the mistakes are pretty much set in stone and it is darned difficult to ever correct the mistakes.

Then there’s captioning for live TV shows such as news reports. These are generally done in real time by a human typist (or speech-to-text software that makes about 3x as many mistakes as the human would). I feel sorry for them. They work at breakneck speed, no time to ever go back and fix any mistakes, and thousands of people see it when they mess up and once in a while their efforts get immortalized on the internet.

[Bijou_Drains]

I’ve noticed whole sentences that are not on the captions. Why is that?

[Keeve]

@sbunny8

Captioning, OTOH, is generally done from video, in a much more deliberate fashion. The captionist may take an hour to properly caption just 10 minutes of video, playing it back over and over at varying speeds, trying to make sense of what they are hearing. It is fiendishly difficult to caption an entire hour of video without making at least one mistake. Sometimes captionists make typos. And sometimes the people in the video talk so fast that what’s being said literally won’t fit on the screen. And sometimes the captionist simply can’t figure out what the people are saying. Amongst the reasons for not being able to hear what someone on the video is saying are such diverse elements as: use of technical jargon, hard to understand accents, variations in volume, and background noise.

You can’t possibly be speaking about a live broadcast, so I presume you’re speaking about a non-scripted show such as have become so popular in recent years, or a game show or talk show. But I would imagine that for a scripted show, they’d start with the script, and merely look to correct the changes, right? Wouldn’t that be faster and easier?

@Bijou_Drains

I’ve noticed whole sentences that are not on the captions. Why is that?

My guess is that they made those captions directly from the script, without bothering to check for ad libs and other changes inserted by the actors.

[sbunny8]

@Keeve

You can’t possibly be speaking about a live broadcast, so I presume you’re speaking about a non-scripted show such as have become so popular in recent years, or a game show or talk show. But I would imagine that for a scripted show, they’d start with the script, and merely look to correct the changes, right? Wouldn’t that be faster and easier?My guess is that they made those captions directly from the script, without bothering to check for ad libs and other changes inserted by the actors.

In some cases, the captionist is given a copy of the script and then they only have to look for deviations, add descriptions of noises, and format the text so it fits the screen (all of which still takes more time than you might think). But many times, the captionist doesn’t have access to the script.

I’ll give you an example, a made-for-TV movie about Viet Nam filmed in 1981, so it didn’t have captions when it was released. Now a professor wants to show the movie in class so they can discuss issues surrounding the Viet Nam war. The college has captionists on staff, two of whom is assigned to make captions for this video. The captionists doesn’t have a copy of the script. Much of the dialogue is hard to hear since much of the movie was filmed outdoors, sometimes with several people talking over each other. One captionist does the first half, the other does the second half. Each spends about 4 hours. The file is attached to the video and sent to the professor a few days before the lecture.

[Student_Driver]

@sbunny8

Transcribing is generally done in real time and uses software that allows the typist to type very quickly with shortcuts, something like 400 keystrokes per minute producing up to 160 words per minute. In many cases, the typist only types in the consonants and the software figures out what word they meant and fills in the vowels. Contrary to popular belief, the goal usually isn’t to produce a verbatim account of what was said but rather a meaning-for-meaning description of the conversation. But I believe the standard may be different for courtroom reporters.

My mom used to do clean-up transcribing work on court transcriptions. She transcribed transcriptions, essentially. She had a special word processing program that would read the data recorded by the in-court transcriber, and used AI to render it into somewhat readable text. She was also given tape recordings of the actual testimony/etc., as well as documents and photographs that would allow her to correct spellings of names and such. She’d read the raw transcription while listening to the tapes on a dictation machine, stop when she needed to reference names, and so on, while also doing basic editing. (She had an extensive background in medical and psychiatric transcribing, so she was a popular go-to for court cases that involved professional medical witnesses or evidence regarding mental health issues or names of drugs.) Seemed kind of convoluted, but she could turn around a proper transcription in just a few hours for a given case. I did get pressed into service as another opinion when court transcription was garbled and the audio was inconclusive. I remember one: the program rendered the word “bassoon,” the tape was garbled, and it took in-context understanding to figure out the word was “basement.”