Captioning: The Essentials
What are closed captions, and why do I need to worry about them?
How can I tell if the captions are okay?
What can go wrong during videotape playback?
What problems can arise in the transmission chain?
What about duplicating or editing captioned material?
What can caption technology do, and what are its limitations?
If you are a broadcaster, a cable operator, a post-production house, a duplication facility, or a producer, you may have questions about closed captioning. With the implementation of the Americans with Disabilities Act and the Television Decoder Circuitry Act, producing and handling captions correctly is more critical than ever. The Media Access Group at WGBH, with more than 30 years of experience handling captions, has for years helped many television and video professionals address these concerns. We hope you will find the following information helpful as you encounter questions about captioning.
Production, Post-Production, and Duplication of Closed-Captioned Videotapes
To create captions for a pre-produced program, highly trained caption writers transcribe the audio portion of the program using special computer software. Captions are carefully timed using the master videotape's own SMPTE time code. Quality captions are also carefully placed to indicate who is speaking (while not obscuring visual elements), and they convey relevant information such as narration, offscreen speakers, and sound effects.
The captioning process does not alter the master videotape in any way. Caption data are inserted into the vertical blanking interval (VBI) though a process called encoding, during which a captioned copy of the master is created. During this process, the video from the master is routed through an encoder. Also routed through the encoder is a data signal from a PC or disk reader that contains the caption information (this comes from the data file created prior to the encoding session). As the master is played back, the encoding system monitors the time-code track on the tape and sends caption data to the encoder at appropriate times. Only line 21 of the VBI is affected during the encoding process; the rest of the video signal passes through unchanged.
Usually, line-21 caption data are encoded only on field 1. In mid-1993, however, the FCC issued a rule allowing data on line 21, field 2, as well. This enables a second stream of captions to be encoded.
The newly encoded video, with caption data now on line 21, emerges from the encoder and is routed to a record machine. This new tape of the program is sometimes referred to as the closed-captioned submaster, or CC master.
Encoding is always a dubbing process. Caption data can never be placed on line 21 of the original master. (Theoretically, caption data could be added to a digital master using the VTR's write-back feature, but this is not recommended.) The closed-captioned master is always a dub or clone of the original tape. Therefore, it is important to make sure that the closed-captioned submaster is clearly labeled as such and is always used for air or duplication.
Closed captions may be seen only with the use of a decoding device, now built in to most televisions as a result of the Television Decoder Circuitry Act, or with a set-top device. A program can also be open captioned (where captions are always visible as part of the video) if desired. If an open-captioned master is needed, one is easily produced during the encoding session: After the video (with caption data on line 21) emerges from the encoder, it is simply routed through a decoder on its way to the record machine. The resulting dub will be open captioned. Whenever it is played back or duplicated, the captions will always be visible, so a decoder is not necessary.
When using a closed-captioned videotape as a source for post-production or duplication, please observe these simple rules:
Broadcasting Line-21 Captioning
- Any duplication of captioned video should be handled through equipment that is adjusted so that it does not blank line 21. Equipment or processes that may blank line 21 include time base correctors, processing amplifiers, frame stores, digital video effects, and switchers.
- Just as line-21 data should be allowed to pass, they should also not be shifted to another line. If your video somehow gets bumped up or down a line, for example, the captioning data will be transmitted on a line other than line 21. Caption-capable television sets and set-top closed-caption decoders look only at line 21, so caption data on the wrong line are effectively lost.
- The video fields should not be flipped. Caption decoders expect the main caption data to be encoded on line 21, field 1. If the fields are flipped, putting the data on field 2, most decoders will fail to display the captions. Even if the video contains two independent caption services on the two fields of line 21, field flipping will render both services unusable. This is because the coding conventions on the two fields are slightly different.
- Caption data are embedded in the video, so certain audio edits such as sweetening and remixing will not affect the captions. For example, you may make changes to incidental audio, like background music, without affecting the captions. You may also replace narration or dialogue with a different voice, as long as the text and the timing remain identical to the original material. If you are changing the words or the timing of the delivery, new captions will be required.
- Editing closed-captioned tapes almost always requires reformatting and re-encoding of the captions. You should be aware that caption data are transmitted just before a caption is to appear; that is, you will see data pulses on line 21 before a person speaks. Even if you edit at audio transitions using simple A/V cuts, you will probably end up damaging some of the captions. It is best to edit the uncaptioned master and have the material reformatted by the original caption provider. Reformatting is a simple and inexpensive process.
- As long as line 21 is not blanked (and as long as head switching does not occur on that line), you can take captioned video down as many generations as the video itself can withstand. It's quite possible, for instance, to prepare a generic advertising spot, leaving off any locality-specific tags, and have your captioning done at that point. Then you can make multiple dupe masters, adding a tag to each.
- Time compression usually destroys the captions, since it involves deleting frames of video. If time compression is necessary, plan on having the captioning reformatted and re-encoded afterwards.
- Signal compression, such as MPEG, involves recovering the caption data from line 21 before compression, embedding the data as user bits in the MPEG data stream, and re-encoding the data on line 21 after decompression. Make sure your equipment vendors are aware of this requirement.
No special equipment is needed to pass captioning along to the viewer, and any broadcaster or cable operator can carry closed captioning. In fact, if everything is correctly adjusted, you may be broadcasting captioning data without even knowing it. As with audio and video, though, it is always possible for the captioning to be lost or distorted somewhere in the broadcast chain. Just as you monitor your audio and video signals, we recommend that you monitor line-21 caption data by obtaining a closed-caption decoder. Make sure that your staff knows how to use it, and be sure staff members are aware of what captioning should look like. "Misspelled" words are usually the result of data errors and are a sign of technical problems.
Broadcasters, cable operators, equipment manufacturers, and anyone else involved in delivering a television signal should exercise the same care in ensuring delivery of captions as they do with audio and video.
- A program's captions are part of the video signal and can be easily duplicated and transmitted.
- Use a closed-caption decoder to monitor your signal and familiarize your staff with what correct captions should look like.
- If you are duplicating a tape, monitor the duplication process.
- Spot-check all recordings to make certain that your equipment is not stripping the caption data or placing them on the wrong line or field. Most problems can be traced to processing amplifiers, time base correctors, frame stores, digital video effects equipment, and switchers.
- Captioned submasters should be clearly labeled as such.
- When reediting material that has been closed captioned, edit from the uncaptioned master, and then have the original caption provider reformat and re-encode the caption data.
- Audio edits such as changes to incidental audio, background music, or replacing narration or dialogue with a different voice (as long at the text and timing remain the same) can be made on a captioned master without affecting the captions.
For more information about captioning, Contact us