Live TV Captioning: Human or Machine Transcription?

  • Thread starter Thread starter Gear300
  • Start date Start date
  • Tags Tags
    Machines
AI Thread Summary
The discussion revolves around the capabilities of captioning technologies used in live broadcasts, such as NFL and FIFA events. There is a debate on whether these systems are advanced machines or skilled human stenographers. Observations indicate that captions often lag behind spoken words in news programs, while in studio settings, the opposite occurs. The use of captions is prevalent, especially for accessibility, but certain formats like Jeopardy can be problematic due to spoilers. Technological advancements, such as those in the Google Pixel 6A, allow for real-time captioning during phone calls and podcasts, suggesting that local machine processing is involved. While machines are increasingly utilized for transcription, especially in live settings, human involvement remains crucial for accuracy, particularly in high-stakes environments like news broadcasts or legal proceedings. Stenotype keyboards are commonly used, and the need for precise transcription is emphasized, as errors can lead to significant consequences in various contexts, including business negotiations and courtroom settings.
Gear300
Messages
1,209
Reaction score
9
Are they really caption machines, or really well-trained stenographers? Like in an NFL or FIFA or other live broadcast. I took it that they might have been caption machines with highly advanced grammar compilers, but looking around, I guess they can just as well be very fast typists/keyboard-players.

When somebody asked two years ago which language would dominate the future (because English seems to hold most of the Internet), I thought that ironically NLP and fast-translating machines would preserve pluralism in language, mostly because in a hundred or so years, we would have fast-translating or quick-witted grammar compilers.
 
Computer science news on Phys.org
Gear300 said:
Are they really caption machines, or really well-trained stenographers? Like in an NFL or FIFA or other live broadcast.
I think machines.

One thing I have noticed.
  1. For news programs, the captions lag behind the spoken words.
  2. For studio programs, the spoken words lag behind the captions.
I use captions on TV almost 100% of the time. But I learned to never use captions when watching Jeopardy because the captions reveal the answer before either I or the contestant have time to answer.

I just got a new Google Pixel 6A phone. It has the feature of showing captions for live phone calls or podcasts. It works for downloaded podcasts even when in airplane mode. Obviously, the captions must be generated locally in the phone's machine. I suspect that I may be able to select the language of the phone's caption machine independent of the language of the speech.
 
  • Informative
  • Like
Likes berkeman, jedishrfu and FactChecker
Zoom does this today in real time. Not very well for scientific talks ("Is Jay-Sigh a rapper?"), unfortunately.
 
It depends. Machines are used in some cases, but TV programs with a decent budget still use humans to transcribe, computers still make to many mistakes and that could cause real problems if what is being transcribed is say a news program or a political speech.
As far as I understand the keyboards used are versions of stenotype keyboards.

Note also that this is not as "exotic" as it might seem, there are lots of cases where transcriptions have to happen live; the most obvious case being in a courtroom, where what is said obviously have to be recorded accurately. There are also services that do live transcriptions of e.g. lectures for people with hearing impairments; a colleague of mine uses one of these services when listening to talks at conferences, the built in transcribe feature in e.g. Teams doesn't work quite well enough.

I don't see why the stenographer would be to "highly trained" just because it is live TV? Sure, mistakes might have more impact on TV, but even if it is just a "local" transcription it still needs to be correct.
I friend of mine used to work for a company that transcribes (live) conference calls between business that are e.g. negotiating contracts; needless to say mistakes can be costly.
 
In my discussions elsewhere, I've noticed a lot of disagreement regarding AI. A question that comes up is, "Is AI hype?" Unfortunately, when this question is asked, the one asking, as far as I can tell, may mean one of three things which can lead to lots of confusion. I'll list them out now for clarity. 1. Can AI do everything a human can do and how close are we to that? 2. Are corporations and governments using the promise of AI to gain more power for themselves? 3. Are AI and transhumans...
Thread 'ChatGPT Examples, Good and Bad'
I've been experimenting with ChatGPT. Some results are good, some very very bad. I think examples can help expose the properties of this AI. Maybe you can post some of your favorite examples and tell us what they reveal about the properties of this AI. (I had problems with copy/paste of text and formatting, so I'm posting my examples as screen shots. That is a promising start. :smile: But then I provided values V=1, R1=1, R2=2, R3=3 and asked for the value of I. At first, it said...
Back
Top