Television and video Closed captioning

1 television , video

1.1 caption formatting

1.1.1 syntax
1.1.2 technical aspects


1.2 caption channels

television , video

for live programs, spoken words comprising television program s soundtrack transcribed human operator (a speech-to-text reporter) using stenotype or stenomask type of machines, phonetic output instantly translated text computer , displayed on screen. technique developed in 1970s initiative of bbc s ceefax teletext service. in collaboration bbc, university student took on research project of writing first phonetics-to-text conversion program purpose. sometimes, captions of live broadcasts, news bulletins, sports events, live entertainment shows, , other live shows, fall behind few seconds. delay because machine not know person going next, after person on show says sentence, captions appear. automatic computer speech recognition works when trained recognize single voice, , since 2003, bbc live subtitling having re-speak being broadcast. live captioning form of real-time text. meanwhile, sport events on channels espn using court reporters, using special (steno) keyboard , individually constructed dictionaries.

in cases, transcript available beforehand, , captions displayed during program after being edited. programs have mix of pre-prepared , live content, such news bulletins, combination of above techniques used.

for prerecorded programs, commercials, , home videos, audio transcribed , captions prepared, positioned, , timed in advance.

for types of ntsc programming, captions encoded line 21 of vertical blanking interval - part of tv picture sits above visible portion , unseen. atsc (digital television) programming, 3 streams encoded in video: 2 backward compatible line 21 captions, , third set of 63 additional caption streams encoded in eia-708 format.

captioning modulated , stored differently in pal , secam 625 line 25 frame countries, teletext used rather in eia-608, methods of preparation , line 21 field used similar. home betamax , vhs videotapes, shift down of line 21 field must done due greater number of vbi lines used in 625 line pal countries, though small minority of european pal vhs machines support (or any) format closed caption recording. teletext fields, teletext captions can t stored standard 625 line vhs recorder (due lack of field shifting support); available on professional s-vhs recordings due fields being recorded. recorded teletext caption fields suffer higher number of caption errors due increased number of bits , low snr, on low-bandwidth vhs. why teletext captions used stored separately on floppy disk analogue master tape. dvds have own system subtitles and/or captions digitally inserted in data stream , encoded on playback in video field lines.

for older televisions, set-top box or other decoder required. in us, since passage of television decoder circuitry act, manufacturers of television receivers sold have been required include closed captioning display capability. high-definition tv sets, receivers, , tuner cards covered, though technical specifications different (high-definition display screens, opposed high-definition tvs, may lack captioning). canada has no similar law receives same sets in cases.

during transmission, single byte errors can replaced white space can appear @ beginning of program. more byte errors during eia-608 transmission can affect screen momentarily, defaulting real-time mode such roll style, type random letters on screen, , revert normal. uncorrectable byte errors within teletext page header cause whole captions dropped. eia-608, due using 2 characters per video frame, sends these captions ahead of time storing them in second buffer awaiting command display them; teletext sends these in real-time.

the use of capitalization varies among caption providers. caption providers capitalize words while others such wgbh , non-us providers prefer use mixed-case letters.

there 2 main styles of line 21 closed captioning:

roll-up or scroll-up or paint-on or scrolling: real-time words sent in paint-on or scrolling mode appear left right, 1 line @ time; when line filled in roll-up mode, whole line scrolls make way new line, , line on top erased. lines appear @ bottom of screen, can placed on of 14 screen rows avoid covering graphics or action. method used when captioning video in real-time such live events, sequential word-by-word captioning process needed or pre-made intermediary file isn t available. method signaled on eia-608 two-byte caption command or in teletext replacing rows roll-up effect , duplicating rows paint-on effect. allows real-time caption line editing.


a still frame showing simulated closed captioning in pop-on style



pop-on or pop-up or block: caption appears on of 14 screen rows complete sentence, can followed additional captions. method used when captions come intermediary file (such scenarist or ebu stl file formats) pre-taped television , film programming, commonly produced @ captioning facilities. method of captioning can aided digital scripts or voice recognition software, , if used live events, require video delay avoid large delay in captions appearance on-screen, occurs teletext-encoded live subtitles.

caption formatting

tvnz access services , red bee media bbc , australia example:

i got machine ready.


engine starting
(speeding away)

uk ims itv , sky example:

(man) got machine ready. (engine starting)

us wgbh access services example:

man: got machine ready.
(engine starting)

us national captioning institute example:

got machine ready.

us other provider example:

i got machine ready.
[engine starting]

us in-house real-time roll-up example:

>> man: got machine ready.
[engine starting]

non-us in-house real-time roll-up example:

man: got machine ready.
(engine starting)

syntax

for real-time captioning done outside of captioning facilities, following syntax used:

>> (two prefixed greater-than signs) indicates change in single speaker.

sometimes appended speaker s name in alternate case, followed colon.


>>> (three prefixed greater-than signs) indicates change in news story or multiple speakers.

styles of syntax used various captioning producers:

capitals indicate main on-screen dialogue , name of speaker.

legacy eia-608 home caption decoder fonts had no descenders on lowercase letters.
outside north america, capitals background coloration indicate song title or sound effect description.
outside north america, capitals black or no background coloration indicates when word stressed or emphasized.


descenders indicate background sound description , off-screen dialogue.

most modern caption producers, such wgbh-tv, use mixed case both on-screen , off-screen dialogue.


- (a prefixed dash) indicates change in single speaker (used captionmax).
words in italics indicate when word stressed or emphasized , when real world names quoted.

italics , bold type supported eia-608.
some north american providers use narrated dialogue.


text coloration indicates captioning credits , sponsorship.

occasionally, karaoke effect music videos on mtv or vh-1.
in ceefax/teletext countries, indicates change in single speaker in place of >> .
some teletext countries use coloration indicate when word stressed or emphasized.
coloration limited white, green, blue, cyan, red, yellow , magenta.
uk order of use text white, green, cyan, yellow; , backgrounds black, red, blue, magenta, white.
us order of use text white, yellow, cyan, green; , backgrounds black, blue, red, magenta, white.


square brackets or parentheses indicate song title or sound effect description.
parentheses indicate speaker s vocal pitch e.g., (man), (woman), (boy) or (girl).

outside north america, parentheses indicate silent on-screen action.


a pair of eighth notes used bracket line of lyrics indicate singing.

a pair of eighth notes on line of no text used during section of instrumental music.
outside north america, single number sign used on line of lyrics indicate singing.
an additional musical notation character appended end of last line of lyrics indicate song s end.
as symbol unsupported ceefax/teletext, number sign - resembles musical sharp - substituted.



technical aspects

there many shortcomings in original line 21 specification typographic standpoint, since, example, lacked many of characters required captioning in languages other english. since time, core line 21 character set has been expanded include quite few more characters, handling requirements languages common in north , south america such french, spanish, , portuguese, though extended characters not required in decoders , unreliable in everyday use. problem has been eliminated market specific full set of western european characters , private adopted norpak extension south korean , japanese markets. full eia-708 standard digital television has worldwide character set support, there has been little use of due ebu teletext dominating dvb countries, has own extended character sets.

captions edited make them easier read , reduce amount of text displayed onscreen. editing can minor, few occasional unimportant missed lines, severe, virtually every line spoken actors condensed. measure used guide editing words per minute, commonly varying 180 300, depending on type of program. offensive words captioned, if program censored tv broadcast, broadcaster might not have arranged captioning edited or censored also. tv guardian , television set-top box, available parents wish censor offensive language of programs-the video signal fed box , if detects offensive word in captioning, audio signal bleeped or muted period of time.

caption channels

a bug touting cc1 , cc3 captions (on telemundo)

the line 21 data stream can consist of data several data channels multiplexed together. odd field 1 can have 4 data channels: 2 separate synchronized captions (cc1, cc2) caption-related text, such website urls (t1, t2). field 2 can have 5 additional data channels: 2 separate synchronized captions (cc3, cc4) caption related text (t3, t4), , extended data services (xds) now/next epg details. xds data structure defined in cea-608.

as cc1 , cc2 share bandwidth, if there lot of data in cc1, there little room cc2 data , used primary audio captions. similarly, cc3 , cc4 share second field of line 21. since caption decoders supported single field decoding of cc1 , cc2, captions sap in second language placed in cc2. led bandwidth problems, however, , current u.s. federal communications commission (fcc) recommendation bilingual programming should have second caption language in cc3. many spanish television networks such univision , telemundo, example, provides english subtitles many of spanish programs in cc3. canadian broadcasters use cc3 french translated saps, similar practice in south korea , japan.

ceefax , teletext can have larger number of captions other languages due use of multiple vbi lines. however, european countries used second subtitle page second language audio tracks either nicam dual mono or zweikanalton used.