More Detail

Transcription

Transcription is the process of taking the videotape and making a verbatim ASCII file of the program so that the text is ready to put into the captioning software.

If the transcriber has difficulty understanding what is being said, he or she makes a best guess and makes a note of it--mental or other otherwise--as an item to be listened to more closely and researched later. Often a second set of ears is the best thing, but a document of troublesome or questionable words, phrases, and terms is created to be used in the proofing process. This document is called a research record.

Proper names are spelled as accurately as possible and noted on the research record. Any unfamiliar word is noted also so that it can be spelled the same way every time. That way a global change can be made when the correct spelling is found - for example, changing John to Jon, or Thompson to Thomson.

Depending on the program, interjections such as "um" and "uh" may not be included in the transcription. In general, these verbal stutters interfere with the overall written clarity of programs like documentaries or informative content shows. Captioning of sitcoms and dramas, however, will often include those parts of speech, as they convey a sense of characterization. Ums, ahs, and wells can translate to nervousness, timidity, uncertainty, even brilliance. The key is to get the intent of the show across to the closed captioning viewer.

The final step in the transcription is spellchecking. And as anyone who's written a term paper knows, spellcheckers don't know everything. But usually words the spellchecker doesn't recognize are words to include on the research record.

Beginning a Project

Captioners pull up the transcription and a digital copy of the program from the computer network. A project record and the research record are kept on a shared network drive. Included in the project record should be information such as the air date of the show, the style (pop-on or roll-up), whether it is for a client with specific captioning guidelines, and the target hours. The research record lists all troublesome or questionable words, phrases, and terms to be checked in the proofing process.

Before beginning, CaptionMax captioners check style guides for particular shows. CaptionMax has a company style guide that serves as the default. Some customers have their own style guides, which then take precedence.

Breaking Apart

Breaking apart is the first real step of editing. The captioner starts by importing the ASCII file from the disk. The job now is to break the text into two-line captions, then to break those captions into clauses--more or less.

First a caption is broken at the end of a sentence. Captions will rarely contain two sentences unless the sentences are very short or repeated. More often, a sentence will have to be broken down into two or more captions. Since we've already broken it after the period, the next step is to scan the sentence for any conjunctions ("and," "so," and "but" are examples of conjunctions).

If the conjunction is joining two independent clauses, a break is introduced before the conjunction. Next, a break can be made at dependent clauses, then at prepositions, and finally between the subject and verb. Here's some examples.

1. At end of sentence.

I walked down to the deli.
I purchased a sandwich.

2. At Conjunctions (and, but, yet, then )

I walked down to the deli,

and I purchased a sandwich,
some egg salad, and cheese.

3. At clauses (since, because, who, which, that). Clauses containing verbs first.

I walked down to the deli
because I wanted a sandwich,
some egg salad, and cheese.

OR

I walked down to the deli
Because I wanted a sandwich,
some egg salad, and cheese
with a side of french fries.

In this case, it's best to keep the 'who' clause with its subject rather than break it off.

When I finally arrived
at my neighborhood deli,
I ordered some egg salad
from the waiter,
who makes giant sandwiches.

4. At prepositions (at, in, for, on, by):

I was really in the mood
for my favorite sandwich
at my favorite deli,
O'Malley's.

5. Between subject and verb:

A well-made egg salad sandwich
served with fries
is my favorite lunch.

There are spots at which a captioner should avoid breaking. Proper names should be kept on the same line, if at all possible. Articles (a, an, the, my, his, etc.) and adjectives should be kept with their nouns. Of course, if there's no room, there's no room. We don't want an entire file of one-line captions.

Correct:

Robin, will you hand me
that power drill?

Incorrect:

Robin, will you hand me that
power drill?

Correct:

I'd like to introduce
Jefferson Davis Hogg.

Incorrect:

I'd like to introduce Jefferson
Davis Hogg.

Timestamping

Timestamping is the process of letting the computer know when the captions should appear and disappear. This is done using a timecode "address" as a reference that both the VCR and the computer understand. Timecode consists of hours, minutes, seconds, and frames. In NTSC dropframe, there are 29.97 frames in a second. In NTSC non-dropframe, there are 30 frames in a second.

The captioner views a copy of the program with a timecode clock burned, or visible, in the video. The captioner uses certain keystrokes in the software that read the current timecode address when struck and assign that time to each caption as either an "in time" or "out time." Typically, captions are made to appear when the words or sounds are first heard, with adjustments made for cuts.

Cuts

A cut is a scene change in the video. Moving to cuts is the process of slightly rearranging the times that the captions appear so that they appear exactly on cuts, disappear exactly on cuts, et cetera. This creates a cleaner overall product and one that is much more attractive to watch.

At this stage, the captioner will also check punctuation and spelling. This is also the time to add speaker identification if it is needed, blank captions, captioning credits, and anything else that may not have made it into the file earlier.

Load Times

The first caption of any file must be blank, center bottom position on the screen, and 10 seconds before the second caption. This caption is necessary to 'jumpstart' the encoder. The encoder may miss or scramble the first caption, so we load a blank in there; this ensures that the second caption will be shown normally.

The second caption (the first caption with text) should be between 1.5-2 seconds after the start of the video signal. We explained above that a blank caption was necessary to jumpstart the encoder. Here we need to make sure that we've got enough time to load the caption. Look at it this way. Each caption takes a certain amount of time to load into the system before it can be displayed. Thus the information for that caption is being read from the tape BEFORE the caption is displayed, often 1 to 2 seconds before the caption comes on, so we allow the caption time to load into the decoder before we need it to appear.

The easiest way to guarantee an adequate load time for the first caption would be for the caption to load 1 second before the start of the video signal, allowing the caption to appear at the start of the signal. But this would require television broadcasters to show video with nothing on it, something they don't like at all, so rather than having the caption appear exactly at the start of the video, we time the appearance of the caption so that the caption loads immediately after the start of the video signal. This is something we deal with every time there is a break in programming, as for a commercial.

Blank captions

The same conditions that apply to the start of the video signal apply to blank captions.

Blank captions should be inserted to queue up the next caption after a 15-second break between captions. The first caption (see "Load Times") is an example of a blank caption. Without a blank caption resetting the loading queue after a 15-second lapse, the next caption will not load.

Blank captions are put 10 seconds before the next ontime both to allow loading and to make certain they do not interfere with the loading of the following caption by being too close.

Offtimes

"Offtime" is a signal that clears the previous caption from the screen. The captioner executes an offtime signal whenever starting a new scene or when restarting the show after a commercial break. This prevents "hangover" captions from previous scenes as well as commercials.

Offtimes require a 1- to 2-second time lag after the signal to allow the system to load the following caption. If the space between an offtime and the next ontime is not sufficient, the caption following the offtime will not display. Generally, we allow 3 seconds between offtimes and ontimes.

Alignment of Captions

Captions should be placed center bottom unless:

• There are two or more people speaking on camera and need to be differentiated.
• There is a graphic, name, map, or other object integral to the video that the caption would cover.
• It's part of a series of top- and bottom-aligned captions that produces a "jumping" or "bouncing" effect. In this case, move all captions to top. Additionally, if the caption would cover a speaker's eyes, we modify its placement.

Italics/Quotation marks

If a person is quoting a written source or if they are repeating dialogue, then that statement needs to be in quotation marks. Quotation marks are placed at the beginning of a quotation, and if the quote goes over several captions, quotation marks go in the first space of all but the first and last caption.

Example:

Sean Connery's character
said,

"If he sends one of yours
to the hospital,

"you send one of his
to the morgue.

"That's the Chicago way,

and that's the way
you'll get Capone."

Speaker Identification

The captioner may wish to add a description of the tone of voice or other relevant sounds. These descriptions go in brackets over the caption. We do not add descriptors if they do not change the meaning of the action or if they are very obvious.

For example, if we see Jane crying, we do not need to caption [crying]. But if she is speaking from offscreen, the descriptor can greatly alter the meaning of the sentence.

Example:

(Jane)
Sure, I'll see you there.

[Jane, sarcastically]
Sure, I'll see you there.

When identifying speakers whose names the captioner does not know, he/she uses lower case: (police officer)(woman)(male announcer). A speaker with a known name is identified by last name unless only the first name is available.

If the audience is not supposed to know the identity of the speaker, we ID the person only as "man" or "woman" until the audience finds out the person's identity.

Music

The title (if known) and the singer/instrument should be in brackets above the beginning of the lyrics. The lyrics themselves should be captioned verbatim. Instrumental pieces should be identified by name whenever possible.

A music note should appear at the beginning and end of each caption of lyrics.

To indicate a long sequence of music without lyrics, two musical notes may be put on screen after the description for up to 8 seconds.

Spelling

The captioner must write down and look up every word that he/she is not absolutely sure how to spell. Often there are words that are very unusual and used almost exclusively when discussing a given topic, so they must be looked up even if the captioner feels relatively certain of the spelling.

Foreign words, except for those like "amigo" and "et cetera" that are technically not English but are extremely common, should be written in italics.

Mispronounced words

A mispronounced word is spelled correctly unless the mispronunciation is used for comedic effect. For example, a character said, "deconamitated" when they actually meant "decontaminated." In this particular instance, it added to the comedic effect, so it was left in.

On the other hand, an expert on the Civil War referred to the development of the lifle (rifle). This was the result of a speech impediment and not a point of emphasis in the program, so it was changed to reflect his intent.

Numbers and Dates

As a general rule, if a number is ten or less, it is written out. Numbers over ten are written as numerals. But since there are always exceptions to the rules and the rules themselves can get confusing, here are some examples: