SDH subtitles go beyond words.
Here's what you need to know.
Regular subtitles show dialogue. SDH subtitles show everything: who is speaking, what sounds are happening, and what music is playing.
If your content needs to be accessible to deaf and hard-of-hearing audiences, SDH is what you should be producing.
Gary Sztajnman
Founder, Hello8
TL;DR
- SDH stands for Subtitles for the Deaf and Hard of Hearing. It includes sound effects, speaker labels, and music descriptions alongside dialogue.
- SDH is not the same as closed captions. SDH is embedded in the video stream and follows subtitle positioning rules. Closed captions use a separate data channel (CEA-608/708).
- ADA Title II (deadline: April 24, 2026), the European Accessibility Act (June 28, 2025), and RGAA all have requirements that can mandate SDH-level detail.
- Good SDH requires more than bracketed sound effects. Speaker identification, music tone descriptions, and non-speech audio all need to be represented.
- SDH takes more time to produce than standard subtitles. Budget for the extra annotation work or use a platform that supports SDH workflows.
What are SDH subtitles?
SDH stands for Subtitles for the Deaf and Hard of Hearing. Unlike regular subtitles that only transcribe spoken dialogue, SDH includes additional information that hearing viewers take for granted: sound effects, speaker identification, music descriptions, and other non-speech audio.
Think of it this way: if a character whispers, regular subtitles show the words. SDH shows the words and tells you they are being whispered. If a phone rings off-screen, regular subtitles show nothing. SDH shows [phone ringing]. If two characters are talking but you can't see their faces, SDH tells you who is speaking.
SDH originated in the DVD era as an alternative to closed captions for formats that didn't support the CEA-608 caption data channel. Today, SDH is the standard approach for accessibility on streaming platforms, Blu-ray, and most digital video distribution.
SDH vs closed captions vs regular subtitles
These three terms get used interchangeably, but they are different things. Here is how they compare.
| Aspect | SDH | Closed Captions | Regular Subtitles |
|---|---|---|---|
| Target audience | Deaf and hard-of-hearing viewers | Deaf and hard-of-hearing viewers | Hearing viewers watching in another language |
| Sound effects | Yes, described in brackets | Yes, described in brackets | No |
| Speaker identification | Yes | Yes (often with positioning) | No |
| Music descriptions | Yes | Yes | No |
| Can be toggled on/off | Yes (subtitle track) | Yes (separate data channel) | Yes (subtitle track) |
| Format / delivery | SRT, VTT, ASS, TTML subtitle files | CEA-608/708 data stream | SRT, VTT, ASS, TTML subtitle files |
The practical difference: closed captions use a dedicated data channel (CEA-608 or CEA-708) and are mostly a North American broadcast standard. SDH delivers the same information but as a subtitle track, which makes it compatible with streaming platforms, Blu-ray, and international workflows. If you are producing content for digital distribution, SDH is almost always what you need.
When is SDH legally required?
Several accessibility regulations now require caption-level detail on video content. While not all of them use the term "SDH" specifically, their requirements effectively mandate the same thing: subtitles that include non-speech audio information.
ADA Title II (United States)
Requires state and local government web content to meet WCAG 2.1 AA by April 24, 2026. WCAG 1.2.1 and 1.2.2 require captions that include speaker identification and non-speech sounds. Private sector lawsuits under ADA Title III have also increased since 2020.
European Accessibility Act (EU)
Applies to digital products and services sold in the EU from June 28, 2025. Requires audio-visual content to provide alternatives for deaf and hard-of-hearing users. Member states set their own fines, which can reach millions of euros.
RGAA (France)
France's accessibility standard (Référentiel Général d'Amélioration de l'Accessibilité) mandates WCAG-compliant captions for public sector and large private organizations. Non-compliance can result in fines up to 50,000 euros per audit cycle.
WCAG 2.1 AA (International)
Success Criteria 1.2.1 (pre-recorded audio), 1.2.2 (pre-recorded captions), and 1.2.4 (live captions) all require that captions include relevant non-speech information. WCAG is referenced by ADA, EAA, RGAA, and most national accessibility frameworks.
Missing SDH can have real consequences
ADA lawsuits targeting video accessibility have risen by over 300% since 2018. The European Accessibility Act carries fines that vary by member state but can reach into the millions. Even if your organization is not directly regulated, SDH is increasingly expected by platforms, broadcasters, and procurement requirements.
What makes good SDH?
SDH is more work than standard subtitles. Here are the five elements you need to get right.
Sound effects
Describe meaningful sounds that affect the story or context. Don't annotate every background noise. Focus on sounds that a hearing viewer would notice and that change how you understand the scene.
Example: [door slams] [gunshot] [glass breaking]
Speaker identification
Label who is speaking when it is not visually obvious. This is especially important in scenes with multiple speakers, off-screen dialogue, or voice-overs. Use the character's name, not generic labels like "Man 1".
Example: SARAH: I'll be there in ten minutes.
Music cues
Describe the type and mood of music, not just that music is playing. "Upbeat jazz" tells the viewer more than "music playing". If lyrics are relevant to the story, transcribe them.
Example: [slow piano music] [upbeat rock music] [singing in French]
Non-speech audio
Include emotional vocal cues that dialogue alone doesn't convey: laughter, crying, sighing, whispering. These change the emotional context of a scene and should not be left out.
Example: [laughs] [sobbing] [whispering] [clears throat]
Consistent formatting
Use the same bracket style and terminology throughout the project. If you write [phone ringing] in scene one, don't switch to [phone rings] in scene five. Create a style guide before you start and stick to it.
Example: Always use present participle: [door closing] not [door closed]
How to create SDH subtitles
SDH production follows the same workflow as standard subtitles, with additional annotation steps.
Start with accurate transcription
Get a complete, timed transcription of all spoken dialogue. Accuracy matters here because SDH viewers rely entirely on the text. With Hello8, you can generate AI transcription in 90+ languages and use the AI Reviewer to catch grammar, conjugation, and punctuation issues before moving to the SDH annotation step.
Add non-speech annotations
Watch the content and add sound effects, speaker labels, music cues, and emotional vocal indicators. This step requires watching the video, not just reading a transcript. A person who has not seen the video cannot produce good SDH.
Check timing and reading speed
SDH subtitles often have more text per cue than regular subtitles because of the added annotations. Make sure your reading speed (characters per second) stays within acceptable limits. Hello8's automated tools check CPS, line length, and overlaps automatically.
Review and deliver
Have someone review the SDH track against the video. Check that speaker labels are consistent, sound effects match what is on screen, and music descriptions are accurate. Export in the format your platform requires (SRT, VTT, TTML, or ASS).
SDH production takes longer than standard subtitles
Talk to our teamLearn about our accessibility approachCommon SDH mistakes
These are the errors we see most often when reviewing SDH subtitle files.
Over-annotating background sounds
Not every sound needs a label. Annotating ambient noise that doesn't affect the story (general traffic, background chatter in a busy scene) makes subtitles harder to read without adding value. Focus on sounds that change how you understand the scene.
Using generic speaker labels
"Man 1" and "Woman 2" are not helpful. Use character names when known. If a character hasn't been named yet, use a descriptive label like "RECEPTIONIST" or "NEWS ANCHOR" rather than a generic placeholder.
Inconsistent formatting
Switching between [DOOR SLAM], [door slam], and (door slam) within the same project makes the subtitles look unprofessional and can confuse viewers who rely on formatting patterns to distinguish sound effects from dialogue.
Ignoring music descriptions
Writing [music] or [music playing] tells the viewer almost nothing. Describe the mood and genre: [tense orchestral music], [upbeat pop song], [soft acoustic guitar]. If lyrics are plot-relevant, include them.
Pre-delivery SDH checklist
Run through this list before you deliver any SDH subtitle file.
All non-obvious speakers are labeled
Every line of dialogue where the speaker is not visually clear should have a speaker label. Check off-screen dialogue, voice-overs, and scenes with multiple characters.
Sound effects are relevant and consistent
Only meaningful sounds are annotated. The same bracket style and terminology are used throughout. No background noise clutter.
Music is described by mood and genre
Every music cue has a description beyond [music playing]. Lyrics are transcribed when relevant to the plot.
Reading speed is within limits
CPS (characters per second) stays between 15 and 20. Lines are not too long. No subtitle stays on screen for less than 1 second.
Formatting is consistent
Bracket style, capitalization, and terminology follow the same rules from first subtitle to last. A style guide was used during production.
Frequently asked questions
What is the difference between SDH and closed captions?
SDH and closed captions contain the same type of information (dialogue, sound effects, speaker identification, music cues), but they are delivered differently. Closed captions use a dedicated data channel (CEA-608 or CEA-708) and are primarily a North American broadcast standard. SDH is delivered as a subtitle track in standard formats like SRT or VTT. For streaming and digital distribution, SDH is the standard approach.
Are SDH subtitles required by law?
It depends on your jurisdiction and sector. In the US, ADA Title II requires government web content to meet WCAG 2.1 AA by April 24, 2026, which includes caption requirements. The European Accessibility Act applies from June 28, 2025. France's RGAA has its own requirements. While these laws don't always say "SDH" specifically, their requirements for captions that include non-speech audio effectively mandate SDH-level detail.
What file formats support SDH?
All standard subtitle formats support SDH content: SRT, VTT, ASS/SSA, and TTML. SDH is not a format itself. It is a content standard. You add SDH annotations (sound effects, speaker labels, music cues) to whichever subtitle format your platform requires. Hello8 exports all of these formats.
Do I need SDH for social media videos?
Social media platforms do not currently require SDH by law. However, adding sound effect descriptions and speaker labels makes your content accessible to a wider audience. Over 430 million people worldwide have disabling hearing loss. If you already produce subtitles for social media, adding SDH-level detail is a relatively small extra step that increases your reach.
How are SDH subtitles different from regular subtitles?
Regular subtitles transcribe spoken dialogue for viewers who can hear the audio but don't understand the language. SDH subtitles transcribe everything a hearing viewer would perceive: dialogue, sound effects, speaker identification, music descriptions, and emotional vocal cues. SDH is designed for viewers who cannot hear the audio at all.
Can AI generate SDH subtitles?
AI can generate accurate transcriptions of dialogue, which is the foundation of SDH. Hello8's AI transcription works in 90+ languages and the AI Reviewer catches grammar and translation issues. However, the SDH annotation step (adding sound effects, speaker labels, music descriptions) still requires a human watching the video. AI can't reliably identify which sounds are meaningful to the story. The best workflow is AI transcription followed by human SDH annotation.
ADA Title II deadline: April 24, 2026
Need SDH subtitles for your content?
Send us a video and we'll show you what professional SDH subtitles look like. Accurate transcription, proper sound annotations, and compliant formatting.
Send us one video and we'll produce an SDH sample