Subtitles

• 1. Words Beneath the Image

  ›

  Subtitles can reveal what a character refuses to say aloud. In Annie Hall, on-screen text exposes the inner thoughts of two characters, thoughts that flatly contradict what they are saying out loud. That use of the form for irony and humor is rare. Far more often, subtitles serve a simpler purpose. They represent the audio of a film, television show, or opera in text form, so that viewers who cannot hear clearly can follow along.

  Creating, delivering, and displaying subtitles involves multiple distinct steps, each with its own requirements. Text must be written. Each line must be timed. The finished file must be encoded in a format the viewer's equipment can read and displayed without overwhelming the image. That chain of work can be accomplished weeks in advance or within seconds of live speech. It can be done by a single enthusiast using free software or by a specialist team at a broadcasting network.

  What shapes those choices and who the system serves turns out to be a question that reaches much further than any cinema screen.

  39

• 2. Making the Words Appear

  ›

  For cinema exhibition, subtitle timing is measured in feet and frames of film; for television and video, the markers are based on electronic timecode. Professional subtitlers work with specialized software that makes each frame of video instantly accessible. That instant access is what allows each subtitle to be set to the precise moment it should appear.

  Subtitle Workshop, MovieCaptioner, and Subtitle Composer are among the free tools that allow individuals to create subtitle files outside a professional setting. Subtitles written by fans rather than official sources are called fansubs. Fansubs in soft subtitle formats can be created and modified relatively easily, which is one reason they are common among fan communities.

  Hard subtitles, also known as hardsubs, are irreversibly merged into the original video frames during production. Because they are part of the image itself, they cannot be turned off and require no special player support. Closed subtitles, stored separately, allow a viewer or distributor to switch between language options without altering the underlying video. That choice between open and closed matters most when there is no time to revise the text before it goes to air.

  39

• 3. Seconds to Spare

  ›

  Communication Access Real-Time Translation, or CART, allows stenographers to convert live speech into on-screen text within two to three seconds. Those stenographers use stenotype or Velotype keyboards to enter phonetic codes, which software then renders as readable text. Even so, real-time subtitles typically lag the audio by several seconds, reflecting the time required to transcribe, encode, and transmit before anything appears.

  English's many homophones make the work especially demanding. Pairs like "there," "their," and "they're" are phonetically identical but carry entirely different meanings. Stenographers in courts and formal inquiries have twenty-four hours to review their phonetic entries and resolve ambiguities. A real-time broadcast stenographer has no such window. Under immediate pressure to deliver accurate text, they must develop separate keying strategies for words that sound identical.

  In November 1998, the Australian Caption Centre put three captioning approaches before the National Working Party on Captioning. Each illustrated a different way of handling a live sports broadcast. The first, heavily reduced, gave only scores and essential words such as "try" or "out." The second, significantly reduced, used standard keyboard input to summarize the commentary rather than transcribe it in full. The third, comprehensive real-time stenography, captured the commentary in its entirety. The Working Party concluded that only the comprehensive method gave viewers genuine access to the broadcast.

  By 2005 in the United States, approximately half of all live captioning was being handled through speech recognition software rather than stenography. Automated systems have made the process faster and cheaper to deploy, but they regularly fail to distinguish homophones. The difficulty compounds when audio quality is poor, when multiple speakers overlap, or when the content involves specialist vocabulary or proper names. Those limitations have raised questions about whether automated captions reliably serve the people who depend on them most.

  54

• 4. Closed Captions and the Law

  ›

  Closed captioning is the American term for subtitles designed specifically for people who are deaf or hard-of-hearing. Unlike translation subtitles, closed captions are a transcription of the original language. They also include written descriptions of non-dialogue audio, placed in parentheses: glass shattering, a door creaking, wind howling, or a cat yowling. Those descriptions are what distinguish captions from subtitles in the strict sense.

  The DVD industry introduced a separate term: Subtitles for the Deaf or Hard-of-Hearing, or SDH. The practical difference for the viewer is largely visual. Closed captions appear as white text on a solid black band that blocks a significant portion of the image. SDH subtitles use the same proportional font as translation subtitles, displayed with an outline or shadow for legibility. Closed captions also typically specify where on screen text appears. This helps viewers track who is speaking when a character is off-screen or when conversations overlap. Some SDH releases, including those from Universal Pictures and 20th Century Fox, include position markers, but this is not common across the industry.

  Song lyrics present a specific problem. Studios have often omitted them from subtitle tracks, arguing that captioning lyrics requires additional copyright permissions. In October 2015, major studios and Netflix were sued over this practice. Plaintiffs cited false advertising and violations of California's Unruh Civil Rights Act. Judge Stephen Victor Wilson dismissed the suit in September 2016. He ruled that the civil rights allegations lacked evidence of intentional discrimination. He wrote that advertising claims "fall far short of demonstrating that reasonable consumers would actually be deceived as to the amount of subtitled content provided."

  Some US DVDs now carry three separate English subtitle tracks. One is SDH. A second provides plain translation text for viewers whose first language is not English. A third carries closed caption data decoded by the viewer's own equipment. That multiplying of tracks reflects how many different purposes a subtitle channel had come to serve. Why some countries defaulted to subtitles while others chose dubbing traces back to a pivotal moment in cinema history.

  93121311

• 5. The Choice to Dub

  ›

  With the arrival of sound film in the late 1920s, film importers in different countries made choices that would define audience experience for generations. Germany, Italy, France, Switzerland, Luxembourg, Austria, and several other countries chose dubbing, replacing the original actors' voices with recordings in the local language. The rest of Europe generally chose to display dialogue as translated subtitles instead.

  Financial considerations drove much of the split. Subtitling is both cheaper and faster to produce than dubbing. During the 1930s, however, a more troubling rationale emerged in Germany, Italy, and Spain. Dubbing allowed those governments to exercise a form of censorship over imported films. By replacing original dialogue entirely, it became possible to create a version that conveyed a message entirely different from the one the filmmakers intended.

  Lectoring, a third approach, became common in Russia, Poland, and a few other Eastern European countries for television broadcasts. A narrator speaks aloud what the characters are saying while the original audio remains audible underneath. Cinemas in those same countries typically show films either dubbed or subtitled rather than lectored.

  Sometimes the deliberate absence of subtitles is itself a storytelling choice. In Not Without My Daughter, the Persian dialogue spoken by Iranian characters is left untranslated. The film follows Betty Mahmoody, who does not speak Persian. By withholding subtitles, the filmmakers place the audience inside her disorientation. Countries that settled on subtitles as their default practice eventually developed distinct approaches to what those subtitles could convey. In parts of East Asia, those approaches looked very different from anything that had emerged in Europe.

  3

• 6. East Asia's Different Path

  ›

  Run Run Shaw of Shaw Brothers Studio pioneered Standard Chinese film captioning in Hong Kong in the 1960s, according to HK Magazine. Shaw had already chosen to record his productions in Mandarin, reasoning that it would reach the broadest possible Chinese-speaking audience. Mandarin alone was not sufficient for non-Mandarin viewers, and dubbing into regional varieties was judged too costly. The solution was to include Standard Chinese subtitles in every Shaw Brothers release.

  Shaw added English subtitles as well, as the films were made in British-ruled Hong Kong. These served the English-speaking population there and made the films exportable outside Asia. The approach addressed a fundamental fact about the Chinese-speaking world. Written Standard Chinese is broadly uniform, but spoken regional dialects can be mutually incomprehensible. Subtitles bridged that gap at a fraction of the cost of dialect dubbing.

  Japanese variety and reality television developed an entirely different relationship with on-screen text. In those programs, subtitles function as decoration rather than as an aid to comprehension. Words appear one by one as they are spoken. Fonts, colors, and sizes are chosen to capture the mood or energy of the moment. Japanese has a rich vocabulary of onomatopoeia, and those words appear in captions as well. The practice spread from Japan to South Korea and Taiwan. ATV in Hong Kong adopted the same decorative style in 2009, while Want Want Holdings of Taiwan owned the network. Want Want Holdings also owns CTV and CTI. That spread of stylistic captioning pointed to a larger shift. On-screen text had become something audiences expected and sought out, even when they had no difficulty following the audio.

  23

• 7. The Hearing Majority

  ›

  A 2021 survey in the United Kingdom found that 80% of viewers between 18 and 25 regularly used subtitles. Among those between 56 and 75, fewer than a quarter said the same. The phenomenon has a borrowed name from urban design: the curb cut effect. A feature built for one group ends up serving a far wider population than its creators anticipated.

  Multiple factors have driven the shift. Advances in microphone technology have allowed actors to speak more softly and to mumble in ways that earlier recording equipment would not have permitted. Studios have grown less willing to spend money re-recording dialogue that does not come through clearly. Theatrical surround sound loses clarity when compressed to the stereo or mono output of a home television, phone, or computer. Some filmmakers have also pursued a wide dynamic range, contrasting quiet dialogue with sudden loud effects.

  Christopher Nolan addressed the complaints directly, recalling the reaction to his films: "We got a lot of complaints. I actually got calls from other filmmakers who would say, I just saw your film and the dialogue is inaudible. Some people thought maybe the music's too loud, but the truth was it was kind of the whole enchilada of how we had chosen to mix it."

  Same-language subtitling serves a different purpose: literacy education. Studies by the University of Nottingham found that subtitles can improve reading comprehension in school-aged children. The What Works Clearinghouse of the United States Department of Education reached similar conclusions. National broadcasters in China and in India, including Doordarshan, use same-language subtitling for this purpose. The most effective implementations change the text color in syllabic synchronization with the spoken audio. That allows readers to track pronunciation one syllable at a time. Disability-rights groups have called for automatic captions to be reviewed by a human before publication. They point especially to educational settings, where inaccurate captions can directly affect students' grades. In the United States, the Federal Communications Commission recommends placing translation subtitles in CC3. CC1 and CC2 share bandwidth with same-language captions, and mixing them would compromise both tracks.

  1415181920