Scenarist Closed Caption
American TV captions as they travel inside the picture: the two bytes each frame carries, in hex, after a timecode. Named after Scenarist, the DVD authoring software, and still what broadcasters, streaming services and editors take for CEA-608 captions.
At a glance
| Extension | .scc |
|---|---|
| Media type | None registered |
| Specification | The bytes: CTA-608 (formerly EIA-608, then CEA-608). The file around them: Scenarist's, never a standard of its own. |
| Time precision | A frame, at 29.97 a second |
| Text encoding | Hex digits. The characters are CEA-608's own: Latin letters, the accents of Western European languages, a few symbols. |
| Styling | Italics, underline and seven colours. No bold. |
| Positioning | A grid of 15 rows of 32 columns |
| Speakers | None in the format. Captioners write >> at each change of speaker. |
| In containers | The captions, not the file: inside the video for broadcast and HLS, a c608 track in QuickTime and MP4 |
Anatomy of a file
A header, then lines of a timecode and the words sent from that frame on, one word a frame. A blank line between each.
Scenarist_SCC V1.0 00:00:22;07 9420 9420 94ae … 00:00:22;29 942f 942f 00:00:24;14 942c 942c
- 1 · Header
Always
Scenarist_SCC V1.0. - 2 · Timecode
The frame the line's first word goes out on. Each word after it takes the next frame.
- 3 · Words
Two bytes each, in hex: a command, or two characters.
A caption, word by word
Captions are not text with times, but the commands a TV's caption decoder follows. These load You're a jerk, in yellow, show it, and clear it.
| Word | What the decoder does |
|---|---|
9420 |
RCL Start loading a caption, out of sight |
94ae |
ENM Clear what is being loaded |
94f2 |
PAC Go to row 15, column 4 |
97a1 |
Tab offset One column further |
912a |
Mid-row code Yellow from here. It takes a column, shown as a space. |
d9ef 75a7 |
Characters Yo, then u': two a word |
942f |
EOC Show what was loaded, at once |
942c |
EDM Clear the screen |
Commands go twice. 9420 9420 is one command: a decoder takes the second as a spare, sent in case the first is lost. The first digit is 9 rather than 1 because the top bit of each byte is a parity bit.
The timecode
SCC counts frames at 29.97 a second, the rate of American TV. Drop-frame timecode, with ; before the frames, skips the frame numbers 00 and 01 at each minute but every tenth, so the clock keeps time. Non-drop-frame, with :, numbers every frame and falls 3.6 seconds behind each hour.
Three ways to caption
| Style | What viewers see | In Subtitld |
|---|---|---|
| Pop-on | A caption appears whole and stays until cleared or replaced. The usual way for recorded programmes. | A subtitle for each |
| Roll-up | Lines appear as they are typed and scroll up, two to four at a time. Live news and sport. | A subtitle for each line, until the next begins |
| Paint-on | Text appears on screen as it is written. | A subtitle for what it paints |
Characters
A row holds 32 characters, and the screen 15 rows. Two characters go in a frame, about sixty a second, so a long caption takes a moment to load.
The character set is not ASCII. Ten of ASCII's codes mean something else (* is á, { is ç, ~ is ñ), a second set adds symbols such as ♪, ½ and ®, and two more add the accents of Spanish, French, Portuguese, German and Danish.
An accented capital comes in two steps. É is sent as a plain E, then a code that replaces it, so decoders without the extended characters still show a letter. A writer that skips the plain letter loses the character before it: Ação comes out Aão. Subtitld sends both.
Styles and places
| Code | Effect | In Subtitld |
|---|---|---|
| PAC | A row, and a column (by fours) or a colour, or italics, to start it | Kept, as one of the nine places, and the style |
| Tab offset | One to three columns further | Kept, in the place |
| Mid-row code | Italics, underline or a colour from here on, shown as a space | Kept |
| Colours | White, green, blue, cyan, red, yellow, magenta | Kept; others become the nearest of the seven |
| Bold, fonts, sizes | — | No place in the format |
SCC in Subtitld
- Every subtitle's text, in rows of 32 columns
- Italics, underline and seven colours
- Where it sits: top, middle or bottom; left, centre or right
- Speakers, as
>>at each change - Start and end, to the frame
- Bold, fonts and sizes
- Characters outside CEA-608's set, which become their plain letter or "?"
- Translations beside the original
- Dubs, speakers' names and faces
Save a .usfx project to keep what is Subtitld's.
Opening an SCC file
Subtitld decodes the file as a caption decoder does, on its screen of 15 rows by 32 columns:
- pop-on captions are subtitles from the frame they show to the frame they are cleared or replaced;
- roll-up captions are a subtitle a line, and paint-on ones a subtitle for what they paint;
- italics, underline and colours become the subtitle's formatting, and its rows and columns its place on the screen;
- the space a mid-row code takes is left out before punctuation;
- drop-frame and non-drop-frame timecodes both read right;
- the first caption channel is read, or the second when the first is empty.
Exporting SCC
Export, then Subtitles and SCC, offers:
- TextThe original, or a translation.
- Speaker namesLeft out, before the text (Thom: …), or as
>>at each change of speaker. - FormattingKept, or removed.
- Shift timesEvery subtitle earlier or later by the same amount.
- TimecodeDrop-frame, as broadcast uses, or non-drop-frame.
Subtitld writes pop-on captions in the first channel. Each loads in the frames before its start and shows on its own frame, its show and clear commands on lines of their own, so that decoders that read a line at a time show it on time too. Rows longer than 32 columns are wrapped, and each row is centred, or put left or right as the subtitle asks. A caption that comes too soon after the last to load in time shows as soon as it has loaded.
Trivia
The first closed-captioned network broadcasts aired on 16 March 1980, on ABC, NBC and PBS, in line 21 of the picture. Viewers needed a decoder box to see them.
From 1993, the Television Decoder Circuitry Act of 1990 required a caption decoder in every television of 13 inches or more sold in the United States.
Characters a second: two a frame, at 29.97 frames a second. A full four-row caption takes over two seconds to send.
Related formats
Open it in Subtitld with the video and retime it on the waveform.