In Australia, television captions are broadcast as teletext data encoded into the television signal.
Preprepared captions (often referred to as offline captions) are matched with the vision and sound of a program through the timecodes embedded in the master tape. When it is created, each caption has an ‘in’ timecode point, which is the signal for it to appear on screen, and an ‘out’ timecode point, which clears it from the screen.
Preprepared captions are created using specialised software which allows the captioner to watch the program (on tape or as a video file such as MPEG), pause it after each section of dialogue, create the caption text (reducing the dialogue if necessary), assign ‘in’ and ‘out’ timecode points, and colour and position the caption. This is a labour-intensive process, and it takes about 8 hours to caption a half-hour program.
Sometimes, when TV stations broadcast programs, the final few seconds of them are cut off. This can mean that the ‘out’ timecode point of a caption is not broadcast, and results in the caption ‘hanging’ onscreen until a caption in a subsequent program appears and clears it. This often happens with the final ‘credit’ caption in a program, and the last caption in commercials.
When overseas caption files are bought for use in Australia, the timecodes are altered to match the ones on the local master tape, and to accommodate different commercial breaks and other changes. The original text is generally not changed. Australian programs are usually captioned at approximately 180 words per minute. Overseas captions may be slower or faster than this. It is usually possible to tell American captions from locally produced ones because they are all in capitals and in white. British captions look much like Australian ones. Almost all overseas programs now shown on TV will have captions which originated overseas.
News captions (often referred to as online captions) are usually prepared in one of two ways. If there is sufficient time, a captioner will prepare a caption file from tapes and scripts of the various stories to be included in the news program. This file will not have any timecoded ‘in’ and ‘out’ points on it, so the captions have to be cued manually as the program goes to air.
If a program is genuinely ‘live’ and there is no time to prepare a caption file, it will need to be captioned by a stenocaptioner using a phonetic keyboard (as used by court reporters). Unlike preprepared captions, which appear on screen as blocks of text, the captions produced by a stenocaptioner appear one word at a time, and scroll up the screen. Because of the way they are created, stenocaptions are almost always verbatim.
The quality of live captions depends on the experience and skill of the stenocaptioner, the quality of their equipment, and the amount of time they have had to prepare for the program. If they know that certain names and technical terms are likely to crop up in a program, they can enter them in their stenographic machine’s ‘dictionary’ prior to captioning, so that these will come up correctly during the broadcast. High quality stenocaptions can be almost as good as preprepared captions. Unfortunately, stenocaptions can also be very poor, with words and whole sentences scrambled.
Recently, a new way of captioning live programs has been developed using speech recognition technology. With this method, a captioner repeats the dialogue of a program into a microphone, and speech recognition software converts this to captions that are automatically broadcast. The technology is still being developed, but is already being used in Australia to caption sports events.
For digital broadcasting, the analogue television signal (with its encoded teletext caption data) is passed through software which converts it to an MPEG file. All the caption data should be preserved during this process, although the way that captions appear on the screen (e.g. their positioning) can be slightly different to the analogue version.
Programs broadcast nationally, regardless of the method of captioning, are generally produced at a central point, then sent out to regional stations, where advertising specific to that region is added.