How to Write Captions for the Deaf

How to Write Captions for the Deaf and Hard of Hearing (SDH)

By BeanyOne

Subtitling is a fairly common practice in television, video games, and movies these days. This is quite a welcome change from the old days, where someone like me with a hearing problem was limited in how he/she could enjoy these forms of entertainment. We tend to take our senses for granted and don't think much about a life without them. Unfortunately, this leads to several of the fallacies that make current styles of captioning insufficient for the deaf audience. The purpose of this guide is to rectify that situation, one person at a time.

[I] The Captioning Mindset

Most captioning is doomed before it begins as far as the deaf community is concerned. This is because people who write subtitles are generally approaching with the wrong mindset. A subtitler generally will do things that only make sense to someone who can hear, such as only subtitling dialogue, not giving indications as to who is speaking, and not indicating when someone is singing. The problem is that these cues are already apparent to someone who can hear them, but a deaf person would not pick up on them at all without some visual tip-off.

This is a fundamental issue of audience. Subtitlers are generally captioning for a hearing audience, which likely does not need it. In order to adopt the proper format for captioning for its intended purpose, you must think about the auditory cues that are important for understanding what is going on. There's almost always a lot more than you realize. Things like ambient music, indicating who is talking (especially with off-screen characters), and sounds that act as dramatic foreshadowing are all examples of things that should be captioned but almost never are.

The mindset is only the first step, albeit a very important one. There's a number of rules that can help.

[II] Captioning DO's and DON'TS

1. DO caption sound effects exhaustively.

While you do not necessarily have to caption every sound that crops up over the course of the work you are captioning, it pays to caption anything that might contribute even slightly to context. Sound effects are often used as an atmospheric effect, foreshadowing, a joke, or "punctuation" for character dialogue. Many important cues are lost when you do not caption these.

In the case of my My Little Pony: Friendship is Magic (FiM) captions, the only sound I typically do not caption is the clopping of the characters' hooves on the ground as they run, which can be assumed to occur every time they run. Captioning that particular sound every time it occurs would needlessly clutter my work. The only time I caption it is when a character is running off-screen, and the sound is the only indicator that someone is there.

2. DON'T use onomatopoeias.

This is somewhat of a fuzzy rule, but it is worth mentioning. An onomatopoeia is a word that is used to describe a sound by imitating the sound it makes. Words like kaboom, vroom, and whoosh are all onomatopoeia. The problem with using words like these in captions is because they only make sense to someone who knows what sound the word is referencing. You can explain what "vroom" sounds like until you're blue in the face to a deaf person, but it won't have nearly the same meaning as it would to someone who hears the sound and can identify with it immediately. It's similar to the sense that the concept of color has no meaning to someone who has been blind since birth.

As such, if you are describing a sound, be as descriptive as you can. Indicate the source, the action that prompted the sound, how loud or soft it is, the pitch (if it has an unusual pitch), etc. There's a huge difference between "(whoosh)" and "(air rushing past Rainbow Dash as she flies)." Remember, this is all information that is not readily apparent to your intended audience, so you owe them every scrap of help you can muster.
As an addendum, it is unreasonable to expect you're going to be able to caption every sound without using an onomatopoeia once in a while. Sometimes, hearing people do not have a better way to describe a sound than by imitating it. In such cases, include as much information as you can to describe the sound in question. The most important thing in the long run is that you're making an effort to include necessary information.

3. DO indicate who is talking.
This should be a universal rule. You should ALWAYS indicate when a new speaker is speaking, regardless of whether that character is on-screen or not. This rule is primarily intended to avoid confusion during conversations, but it is also useful in things like songs with multiple characters, after a scene transition, or after a short while with no characters speaking.

Generally, if in doubt, it's better to err on the side of too much information and include the speaker.

4. DON'T spoil unknown information before it becomes known to the hearing audience.

This rule requires some explanation. Since what is being captioned is most likely for entertainment, there will be dramatic foreshadowing and other uses of sounds and voices that might hint at something that is later revealed to the audience. For example, the first time Discord speaks in "The Return of Harmony," it is a disembodied, unidentified echo. It is not attributed to Discord until Princess Celestia specifically calls him by name.

Part of willing suspension of disbelief is that we accept that the information the scriptwriters intend to convey will be revealed at a pace they feel is appropriate. By giving information that would not be known to the hearing audience beforehand, the captions feel like they are nudging us and speaking in our ear. It's an informal sense of "hey, buddy, I got this information on the down-low, but let's keep it between you and me, hm?"

There are two general exceptions to this rule, both of which require a judgment call on the part of the captioner. The first is when a character suddenly interrupts dialogue, usually by screaming something such as "WAIT!" or "QUIET!" It is up to the captioner to decide whether to include this in the same caption as the dialogue that is being interrupted, thereby tipping off that it will occur, or use caption timing to "simulate" the effect of the interruption. In either case, the interrupted dialogue should be cut off with a dash: "What?! Pinkie, I-"

The second is when a character's name is never given in the show's canon but may have an "expanded" canon name (in the case of FiM, these could be toy names or production notes from the scriptwriters). The "correct" way to do this would be to describe the character, not give their "expanded" canon name. However, it is perfectly acceptable to give a name so long as it can be attributed to the character who is speaking. The primary goal of revealing this information is reducing conversational confusion.

5. DO caption songs and song lyrics.

Musical presentations are an increasingly popular trope in modern entertainment, and with good reason. Songs are very memorable, tap into our emotions, and auditorily pleasing when well-written. However, songs are very difficult to adequately convey to the audience you are trying to reach, for several reasons.

It's not enough to simply caption the lyrics. There is a lot of "soundplay" in singing. A singer might do things such as hold a note, trill a note, change pitch, change tempo, pronounce a word differently, and even vacillate between emotional states. This is a lot of information to convey in a tiny box at the bottom of the screen. However, you are obligated to include as much as you think will be necessary to understand the proceedings.

Sometimes characters in these musical proceedings will do things such as complete each others' sentences or sing in harmony. This is actually somewhat of a procedural headache for a captioner, but it can be solved with a bit of ingenuity. In the case of the former, include all of the complete sentence in one caption, arranged in order according to who is speaking. Punctuate the "cutoff" (both the end of the first line and the beginning of the next) with dashes. In the case of the latter, usually the harmony is between a chorus line and the main singer. Keep the main singer the same format as always, but enclose the chorus line in [] brackets. This indicates that the chorus line is "singing in the background," as it were.

Songs are not just lyrics. The music is a large part of the experience. While deaf viewers may not necessarily pick up on all the subtle cues of the music, it is still necessary to include terms that will broaden understanding of what the music is doing. Always start off by attempting to quantify what kind of music is playing. If it has a clear, well-defined genre, indicate that and the tempo. If it's intended to convey emotion, state what that emotion is. As the song gets underway, be sure to indicate changes in pitch, tempo, or emotional inflection. Terms like "crescendo" or "swell" have a strong place, as most musicals use what is called a claptrap ending: they build up to a huge finale, then everyone poses for applause. Modern entertainment is no different, so make sure you indicate when this is happening.

6. DON'T split sentences needlessly.

One common problem I notice when people caption is that they are so afraid of putting too much information in a single caption that they risk losing the viewer or running outside the bounds of the screen. This is a valid concern, but the solution often proposed is not exactly useful. Captioners will typically solve this problem by splitting their captions and compensating for them with timing. The problem this presents is that often the sentences are split in a way that leaves only one or two words in the following captions or breaks up a subordinate clause/phrase.

Consider the following example:

Peter Piper picked a peck of pickled peppers.
Peter Piper picked a peck of
pickled peppers.
Peter Piper picked a
peck of pickled peppers.

All three of these are instances of the same sentence. However, the second sentence splits up the prepositional phrase "of pickled peppers," which can mess with contextual reading of the sentence. Similarly, the second sentence splits the article "a" from the word it modifies, "peck."

Splitting a sentence is always a judgment call, but if a sentence must be split into more manageable fragments, make sure the fragments individually flow and make sense. If necessary, test your own captions to make sure that you are properly conveying the information in a way that is easy to follow.

7. DO use parenthetical indicators.

There are two important kinds of parenthetical indicators in captioning: stage directions and transcription notes.

Stage directions are any description of non-dialogue action. In the case of captions, this usually falls to descriptions of sounds and music. These should be separated by () parentheses from the dialogue. It is important to note that stage directions are actions, not complete sentences. They do not have sentence-ending punctuation, unnecessary capitalization, or even a full independent clause fragment. They are merely intended to indicate that something is going on besides the dialogue currently being spoken.

Transcription notes are anything the captioner inserts into the captions to clarify something. The most common use of this is to explain a joke that would go over the head of someone who can't hear it. Puns are a common source of frustration for deaf viewers, as they are based on words that sound similar. In such case, after the joke is made in the caption, a parenthetical indication in [] brackets alluding to the word or phrase it sounds like is sufficient. Example: "Cranky Doodle Donkey [Yankee Doodle Dandy]." Sometimes the joke is a sound, music, or a character imitating another well-recognized pop culture character. In this case, put stage directions inside [] brackets to indicate the joke. Example: "Iron Will: [a la Mr. T] I pity the foal [fool]."

Sometimes you'll have to make a lot of parenthetical indications, and they might get confusing. Try to keep them as ordered as possible. If you need a third parenthetical indicator, {} curly braces can work in a pinch, but this is inadvisable as {} curly braces are used most often in computer programming. Try to limit usage of them if you can, but above all, be consistent. If you change style midway through your captions, you risk frustrating and confusing your viewers.

8. DON'T let personal bias sneak into your captions.

As a captioner, you must remain faithful to the source material. Characters you don't like must receive equal treatment, even if you'd rather not listen to or caption what they're saying or doing. In addition, you cannot let your likes and dislikes sneak into your language regarding stage directions. Try to remain as objective as possible. "Sinister" is a better description for "evil" than "murderous," for example. Sinister tones are a trope in speech to portray "evil" characters, but they are not necessarily murderous, nor are they even necessarily evil. Often a character will act one way with motivations that seem almost completely opposite what they're doing.

Ultimately, you are informing people what is going on, not making their judgments for them.

[III] Further Reading

This is the bulk of what you need to know in how to write captions for the deaf, but I would highly recommend turning on captions in a Valve game such as Left 4 Dead or Half-Life 2. Full captions, not just dialogue. This is an EXCELLENT indication of what captions should be like in video games, but even done right, there's a lot of hard-to-catch bugs and procedural issues that make even well-done and properly timed captions screw up.

Above all, though, practice writing captions for stuff. Kids' shows are actually great practice for this. Try to get feedback from deaf friends, if you have them. It can be very hard to identify what is and isn't important information when your hearing is good enough to understand something without assistance.