SDH, short for subtitles for the deaf and hard of hearing, are subtitles that render all meaningful audio as text: dialogue plus speaker identification, sound effects, music cues and tone of voice. In video accessibility, SDH combine the visual format of ordinary subtitles with the informational completeness of closed captions, so a viewer who cannot hear the soundtrack can still follow everything the audio contributes to the story.
SDH exist largely for technical and distribution reasons. Traditional broadcast closed captions are encoded in formats that many disc players, digital connections and streaming platforms do not carry, so SDH deliver caption-level information through the subtitle track system that those platforms already support.
What SDH contain beyond dialogue
The defining feature of SDH is that they assume the viewer hears nothing at all. Standard subtitles assume the viewer hears the audio but does not understand the language; SDH must therefore describe every sound that matters for comprehension, not just translate or transcribe speech.
| Element | Purpose | Typical rendering |
|---|---|---|
| Speaker identification | Shows who speaks when unclear | Name label before the line |
| Sound effects | Conveys plot-relevant noises | Bracketed description such as door slams |
| Music cues | Indicates mood or lyrics | Music note symbol or bracketed description |
| Manner of speaking | Captures tone such as whispering | Bracketed descriptor before the line |
| Off-screen speech | Signals voices not visible | Label or italics, depending on style |
| Silence or inaudible audio | Explains meaningful absence of sound | Bracketed note such as no audio |
Good SDH describe sounds that carry meaning, not every noise. A ringing phone that a character ignores may matter for the plot; traffic noise in a city scene usually does not. Choosing what to describe is an editorial judgment that separates professional work from automated output.
SDH compared with subtitles and closed captions
SDH sit between standard subtitles and closed captions: they look like subtitles but contain the information of captions. The differences matter when ordering deliverables or specifying accessibility requirements.
| Feature | Standard subtitles | SDH | Closed captions |
|---|---|---|---|
| Assumed audience | Hearing viewers | Deaf and hard of hearing | Deaf and hard of hearing |
| Non-speech information | No | Yes | Yes |
| Visual style | Subtitle font, usually bottom | Subtitle font, usually bottom | Caption style, often boxed text |
| Delivery format | Subtitle track or file | Subtitle track or file | Embedded caption data or caption file |
| Positioning | Mostly fixed | Can move to avoid on-screen text | Positioned near speakers in many formats |
| Can be turned off | Yes | Yes | Yes, unlike open captions |
Forced subtitles, which translate only foreign-language passages or on-screen text, are a different deliverable. A complete SDH track must include those forced elements as well, because the viewer relying on SDH will not have a separate forced track active.
Terminology also varies by market. In some regions the word subtitles is used for all on-screen text, including captions, while in North America captions and subtitles are kept distinct. Specifying SDH explicitly in delivery requirements avoids receiving a translation-only file.
Formats and technical parameters
SDH are delivered in ordinary subtitle formats, such as SubRip and WebVTT files, timed text formats based on TTML, or bitmap subtitle streams on optical discs. Each format supports a different level of styling and positioning.
| Parameter | Common range | What it affects |
|---|---|---|
| Characters per line | About 32 to 42 | Readability and screen coverage |
| Lines per subtitle | 1 to 2 | Amount of image obscured |
| Reading speed | About 15 to 20 characters per second | Time available to read each line |
| Minimum duration | About 0.8 to 1 second | Whether short lines can be read |
| Maximum duration | About 6 to 7 seconds | Avoids lingering, stale text |
| Gap between events | About 2 frames or more | Lets the eye register a change |
Exact values come from each distributor’s style guide and often differ for adult and children’s content, with slower reading speeds for younger audiences. SDH may need slightly more condensation than standard subtitles, because sound descriptions compete with dialogue for the same reading time.
Writing and quality practices
Professional SDH follow verbatim or near-verbatim transcription of dialogue, condensed only when reading speed limits demand it. Paraphrasing that changes meaning or removes a speaker’s dialect and personality reduces accessibility rather than improving it.
- Synchronize precisely with speech, so lip movements and text match and jokes do not appear before the punch line.
- Identify speakers only when needed, for example during off-screen dialogue or rapid exchanges.
- Keep sound descriptions concise and objective, describing the sound rather than interpreting the scene.
- Reposition text when it would cover burned-in titles, lower thirds or important visual details.
- Review against the final picture, because late edits often shift timing and break synchronization.
Automatic speech recognition can produce a first draft, but unedited machine output rarely meets SDH quality: it misses sound effects, misattributes speakers and struggles with names, overlapping speech and music.
Common mistakes
The most frequent error is delivering a standard subtitle file labeled as SDH. Without sound descriptions and speaker labels, the file fails the audience it claims to serve.
- Describing irrelevant sounds, which clutters the screen and slows reading.
- Omitting lyrics that carry plot or emotional meaning.
- Inconsistent bracket and label style within one program.
- Missing forced narrative text, such as translated signs or foreign dialogue.
A final quality check with the sound turned off is the most reliable test. A reviewer who watches the program muted quickly notices missing cues, unclear speakers and lines that disappear too fast. Many distributors also run automated checks for timing and format before acceptance, but these cannot judge whether the right sounds were described.
The short version
SDH are subtitles written for people who cannot hear the sound. They show what characters say, plus who is speaking and which sounds matter. A line such as a door slamming or soft music tells the viewer what a hearing person would notice.
Regular subtitles help people who hear the film but do not know the language. SDH help people who know the language but cannot hear. That small difference changes what has to appear on the screen.
Where it makes a difference
A documentary producer prepares a feature for release on a streaming platform and on Blu-ray. The platform requires a closed caption file, while the disc authoring house needs subtitle streams, because broadcast-style caption data is not carried through the disc player’s HDMI output.
The subtitle team creates an English SDH track from the verified transcript. They add speaker labels for off-screen interviewees, bracketed descriptions such as a helicopter approaching, and music cues for the score. They keep reading speed around 17 characters per second and move text upward whenever interview name titles appear at the bottom of the frame.
The same SDH file, converted to the required formats, serves both releases. Deaf and hard of hearing viewers get the full story on every platform, including the tension created by off-screen sound.