Joe Clark: Media access

You are here: joeclark.orgCaptioning and media access
CAB Forget-Me-Nots

Updated 2002.08.13

Captioning manual forget-me-nots

Here is a quickie list of topics that any credible captioning manual must cover. The Committee cannot reasonably release its style guide without coherently addressing these issues in all necessary detail.

Correcting present-day errors

The CAB manual must be edited to specifically outlaw the worst habits of current and past Canadian captioning.

  1. >> in popup captions
  2. < > as delimiters
  3. [ Gerund of noun ] as NSI format
  4. [ ||| ]
  5. All-uppercase captions
  6. Blink rate
  7. Common misspellings
  8. Defaulting to line 14
  9. Extended quotations
  10. French-language interference
  11. [inaudible], ... as placeholders for guessing
  12. Hyphenation
  13. Impossible-to-parse, overpunctuated speaker IDs
  14. Parallelogram-shaped captions
  15. Random italicization
  16. Spaces inside brackets and parentheses
  17. Speaking in unison
  18. Standard punctuation of quotations
  19. Turn NSI into verb phrase, not two-part caption

Speed & editing

  1. Cursing, bleeps, deletions, and dropouts (misuse of ellipsis and dashes)
  2. Philosophy of edited captions: Patronizing, condescension, and equality
  3. Post-postproduction sweetening
  4. Research evidence
  5. Techniques to accommodate faster text


  1. Mixed case only
  2. Upper case IDs, with lower-case prepositions and NSI
  3. Canadian spellings, save for special cases
    1. Use Oxford and/or Gage Canadian Dictionaries, but develop a collection of national dictionaries
    2. Adamantly not British spelling (organise, programme)
    3. Limited range of exceptions to standard English orthography
  4. Appositives
  5. Italicization rules
  6. No necessity to follow unsupportable “official” orthographies (kd lang, KISS), but match onscreen orthography where practicable
  7. Quotation-mark rules
  8. Rendition of numbers, including approximates
  9. URLs are case-insensitive
  10. Use of full character set, including £
  11. Words spelled out verbally

Speaker identification

  4. Combine NSI with ID: NICOLE, whispering:
  5. Do not alter single-letter names: G:K:J:Q:M:
  6. Do not edit dialogue to make more room for a speaker ID (go to three lines of dialogue text and a fourth line devoted to speaker ID if necessary, though it almost never is)
  7. ID source of sound effects
  8. Identify age when dramatically significant: NICOLE:YOUNG NICOLE:ADULT NICOLE:
  9. Identify gender whenever possible for narrators, announcers, unnamed voices, aliens, animals, and machine-generated voices
  10. Upper case, followed by colon, on its own line unless impossible
  11. Use of leading dash to denote speaker change, including myriad ways to get it wrong
  12. When speakers ID themselves, do not add another one

Non-speech information (NSI)

  1. Use brackets without spaces, with (likely) initial lower case
  2. [laughing] vs. Ha ha ha vs Ha. Ha. Ha
  3. [no voice][no audio][mouthing]
  4. [woman calling moves]
  5. EKGs and flatlining
  6. Progressive vs. indicative aspect

Scrollup captions

  1. >>> optional, possibly unnecessary
  2. Captioning sound effects
  3. Continuous vs. excerpted real-time captioning (as in hockey games)
  4. Do not centre lines (permissible in closing caption credits)
  5. In real-time captioning, convey emphasis through use of quotation marks (STUDIO 54 WAS "THE" MANHATTAN DISCOTHEQUE)
  6. Indication of interpreters
  7. Music captions identical to popup music captions
  8. Must ID gender of unnamed speakers
  9. Names of correspondents
  10. Overuse of clear-screen command
  11. Overwriting existing captions with real-time captions
  12. Prohibition of use for fictional narrative programming
  13. Return after every sentence, and exceptions
  14. Use of tab stops
  15. Use of real italics

Exception dictionaries

  1. Stop cold and check anything marginally unfamiliar; “Gee, that was weird” and “I wonder what that meant” will not cut it in captioning
  2. alright
  3. any[ ]day, every[ ]day, any[ ]more, any[ ]time, a[ ]while
  4. O
  5. oh-so
  6. So[,] ¶ And[,] ¶ But[,] ¶ Then[,] ¶ Now[,]
  7. Adjoining verbs, e.g., those who can do
  8. Approximate vs. exact figures and ranges
  9. Articulated pronunciation (“quote-unquote,” “slash,” “period”)
  10. Comma before addressees
  11. Contractions
  12. Currencies: $5 vs. $5.00, $16.50 vs. 16.50, “eighteen pounds 15” as £18.15
  13. Long noun phrases, e.g., what it is is
  14. Nonverbal utterances: uh, ah, ahh, ahhh, eh, ehh, ehhh, mm, um [not erm], umm, ummm, er (not uh), uh-huh, uh-uh[-uh], mm-mm, mm-hmm, duh, d'oh, oi vs. oy
  15. Number and fraction ranges
  16. Sacred texts: Caption verbatim
  17. Slang and idiom dictionaries
  18. Use current orthography unless source uses outdated original (e.g., Zedong vs. Tse-tung)
  19. [Be]cause


The following list could be subdivided to some degree, a task that remains uncompleted.

  1. 608-to-708 upconversion
  2. Must displace captions for onscreen text and Chyrons
  3. [asks question in Japanese][responds in Japanese][Japanese dialogue continues]
  4. [singsong] delivery
  5. Abandonment of TeleCaption and TC II features
  6. Actual utterance of “(1),” “(c),” and other section markers
  7. Adding blank spaces to fill in blocks
  8. Alternate caption position to cover and uncover fixed full-screen text
  9. Beginning and ending caption credits
  10. Caption chunking
  11. Caption credits and simultaneous dialogue
  12. Caption sponsorships
  13. Captioning over opening and closing titles
  14. Captioning subtitled and/or dubbed productions, including dubbed reading of onscreen text
  15. Clearing screen during long pauses
  16. Colour
  17. Completion of onscreen text, and replication of slight variation between onscreen text and verbal delivery
  18. Conversion to DVD subpictures
  19. Covering mouths
  20. Disclaimers and bumpers, including audio-description bumpers
  21. Doggerel
  22. Error-checking
  23. Fake overdubbing
  24. Foreign languages
  25. Historical re-enactments
  26. Identifying music pieces and genres
  27. Interpreted dialogue
  28. Lines varying by one or two characters in length
  29. Magnetic theory of placement
  30. Manner of speech
  31. Mixed sound effects and singing, speech and singing
  32. Mockery
  33. Mode choice: Popup vs. scrollup vs. paint-on, and mixtures
  34. Montages of voices
  35. Multiple simultaneous caption blocks
  36. Must list feedback contact address
  37. No end-of-caption punctuation in music
  38. No use of . as clear-screen manqué
  39. No use of || in music
  40. Non-numeric time designations (“quarter of seven”)
  41. Presence and nature of theme music
  42. Preserve dramatic exposition: [thud] not [body thuds]
  43. Puns
  44. Quotations within quotations
  45. Re-encode to fix reported or discovered errors
  46. Recaps from previous episodes, previews of future episodes; flashbacks, flashforwards
  47. Reformatting and recaptioning; strip old captions when recaptioning (includes small snippets in larger programs)
  48. Remove extraneous punctuation
  49. Rule-governed editing for children
  50. Sampling
  51. Segue between thought or narration and diegetic dialogue; thinking vs. inner voices vs. narration
  52. Speech rendered through telephones, answering machines, radios, communicators
  53. Splitscreens
  54. Start at frame 7
  55. Technical compatibility and overcoming limitations
  56. True right justification
  57. TTY numbers
  58. Understanding breaks around shot and scene changes
  59. URLs
  60. Use of CC logotypes
  61. Use of multiple caption blocks
  62. Use of paint-on style
  63. Use of slash vs. comma in long NSI chains
  64. Ventriloquism and mouthing along