SDH vs CC vs Subtitles: What Your Company Videos Actually Need
SDH vs CC comes down to audience and delivery. Subtitles translate dialogue for viewers who can hear but do not speak the language. Closed captions (CC) and SDH, subtitles for the deaf and hard of hearing, also show who is speaking and the sounds that carry meaning. Most company videos need one accessible track per language. DEMA Solutions produces subtitles and SDH in many languages.
Who this guide is for
This guide is for the people who commission video rather than watch it: marketing and communication managers publishing product launches and brand films, HR and learning teams building onboarding and compliance training, and product managers recording demos and tutorials. At some point each of them is asked: subtitles, captions or SDH? SRT or VTT? Burned in or as a separate file? In which languages? Most explanations are written for viewers choosing a TV setting. This one is written for the person who has to order the right file and keep it consistent in every language.
Subtitles, closed captions, open captions and SDH: the definitions
The terms overlap and are used differently on each side of the Atlantic, which is why briefs so often go wrong. These are the working definitions most professionals use.
Subtitles are on-screen text that renders the spoken dialogue, usually in another language. They assume the viewer can hear the soundtrack: the music, the doorbell, the change in someone’s voice. Their job is to bridge a language gap, so they leave out anything that is not speech.
Closed captions (CC) render speech and the relevant non-speech audio, and they are “closed” because the viewer can switch them on or off. The term comes from North American broadcasting, where captions were encoded in the television signal and traditionally displayed as white text on a black box that could be placed anywhere on the screen. In the accessibility guidelines used across the web, captions are defined as a synchronised text alternative for both speech and non-speech audio needed to understand the content.
Open captions are always visible because they are burned into the picture. Nobody can switch them off, translate them or hide them. They are common on social media clips that play without sound.
SDH, subtitles for the d/Deaf and hard of hearing, carry the same information as captions, dialogue plus speaker identification, sound effects and relevant music, but follow subtitle conventions in their presentation. The format grew with DVDs and streaming, where a single subtitle file had to serve both accessibility and translation. In Europe, “subtitles for the deaf and hard of hearing” is the usual term, and in the United Kingdom “subtitles” often covers both.
So what are SDH subtitles in practice? An accessibility track for viewers who cannot rely on the audio, delivered in the same format as ordinary subtitles.
SDH vs CC: what is actually different
For a company publishing video online, SDH and closed captions differ less in purpose than in tradition and format. Both give someone who cannot hear the soundtrack the same information as someone who can.
Origin and format. Closed captions were born in broadcast standards and can be positioned on screen, near the speaker or away from on-screen text. SDH is delivered as a subtitle file and usually looks like normal subtitles, centred at the bottom, in the player’s subtitle style.
Language. Closed captions are typically in the same language as the audio, because their purpose is access, not translation. SDH can be produced in the source language or in a target language, which is why it is the natural choice for multilingual corporate video: a German SDH track for an Italian training film is both a translation and an accessibility track.
Conventions. Captioning traditions describe sounds in brackets and label speakers by name; some European SDH guidelines use colours or positions instead. Either works if applied consistently.
When asked to choose between SDH and CC, ask back: will the file play in a broadcast or a web player, and must it be translated? For web video in several languages, an SDH-style track per language, correctly labelled in the player, is usually the answer.
What SDH adds: speaker identification, sound cues and music
The difference between an ordinary translated subtitle file and an accessible track is the information a hearing viewer gets from sound without noticing it.
Speaker identification matters whenever the speaker is off screen or several people talk. A deaf viewer cannot tell whether an instruction comes from the trainer or the safety officer in the next shot unless the subtitle says so.
Sound effects matter when they change meaning. In a product video, the click that confirms a part is locked, the alarm that signals a fault or the beep at the end of a cycle are information, not decoration. An SDH track describes them briefly, for example as an alarm sounding or a click, at the moment they occur.
Music and tone matter when they set context, such as a brand film that turns from sombre to upbeat or a remark that is clearly ironic. Good SDH describes these only when relevant, because describing every chord makes the text harder to read.
All of this requires judgement about which sounds matter and how to describe them in a few words, repeated in every language. This is why DEMA treats SDH subtitles for d/Deaf and hard-of-hearing viewers as a separate deliverable with its own conventions, agreed with the client before work starts.
Subtitles vs closed captions: which does your audience need?
The subtitles vs closed captions question is easiest to answer by asking who will watch the video and how.
If your audience can hear but speaks another language, for example distributors in Spain watching an Italian product launch, translated subtitles bridge the language gap. If the video is internal and in the audience’s language, source-language captions make it accessible to colleagues who are deaf or hard of hearing, and to everyone watching on a train. If the video is public, or goes to employees and customers in several countries, the safest choice is an accessible track in every language, because you cannot know who in each audience cannot hear.
Captions vs subtitles is therefore rarely either-or. Many companies settle on SDH in the source language, SDH or subtitles in each target language and burned-in captions for social clips, decided once as a policy rather than video by video.
Standards and guidelines: WCAG, EN 301 549 and national practice
There is no single European law that prescribes how a caption must look, but there are clear reference points.
WCAG. The Web Content Accessibility Guidelines, in their current versions 2.1 and 2.2, contain the criteria most organisations test against. Success criterion 1.2.2 Captions (Prerecorded), level A, requires captions for all prerecorded audio content in synchronised media, unless the video is itself a clearly labelled alternative to text. Criterion 1.2.4 Captions (Live), level AA, extends this to live content such as webinars and broadcasts. The audio description criteria, 1.2.3 and 1.2.5, address the opposite need: making visual information available to people who cannot see it. Because WCAG defines captions as including non-speech audio, a dialogue-only subtitle file does not meet the wcag captions requirement on its own.
EN 301 549. The European standard for accessibility requirements of ICT products and services incorporates WCAG 2.1 level AA for web content and adds requirements for technology that plays video, such as displaying captions, keeping them synchronised, preserving them when content is converted and giving users controls to switch them on. It is the reference used under the Web Accessibility Directive for public sector bodies and the benchmark most companies use to check compliance in practice.
National practice. Several countries have their own SDH conventions. In Spain, standard UNE 153010 sets requirements and recommendations for subtitling for deaf people and people with hearing disabilities, covering presentation, timing, speaker identification and sound information. In German-speaking countries, the public broadcasters of Germany, Austria and Switzerland have agreed shared principles for subtitles for hearing-impaired viewers, with house style guides on top. Other markets rely on broadcaster or platform guidelines. For a company video, the practical rule is to agree which convention each language follows and to apply it consistently.
Formats: SRT, WebVTT and burned-in captions
The format decides where your subtitles can be played and how much of the SDH information survives.
SRT is the simplest and most widely accepted subtitle file: numbered cues, start and end times, and text. Almost every player and video platform accepts it. Its weakness is presentation: positioning and styling are not reliably supported, so speaker labels and sound cues have to be written into the text.
WebVTT is the W3C format for web video and the one HTML5 players use natively. It supports positioning, basic styling and voice tags that mark who is speaking. In an HTML page, a WebVTT file is attached with a track element whose kind can be “subtitles” or “captions”, which tells assistive technology what the track contains. Labelling an SDH file as captions, with the right language code, is a small step that makes a big difference for users.
Broadcast and streaming workflows use their own timed text formats, rarely needed for corporate video but worth asking for when a film goes to a broadcaster or festival.
Burned-in (open) captions are rendered into the image. They work everywhere, which suits social clips that autoplay without sound, but viewers cannot switch them off or resize them, assistive technology cannot read them, and every language needs its own video file. For training and product libraries, sidecar files are the better choice.
Reading speed, line length and segmentation
A subtitle that is correct but appears too briefly fails the viewer just as surely as a wrong one. Professional guidelines therefore limit how much text appears and for how long.
Most guidelines allow two lines per subtitle and roughly 37 to 42 characters per line, depending on the language and the screen. Reading speed is usually expressed in characters per second, and adult guidelines typically fall somewhere between 12 and 17 characters per second, with lower speeds recommended for children and for some deaf audiences. Each subtitle should stay on screen long enough to be read, not cross a shot change where avoidable, and break lines at natural points in the sentence rather than in the middle of a phrase.
As a result, subtitles rarely reproduce every word. Speakers talk faster than anyone can read, so the subtitler condenses, keeping meaning, terminology and tone while removing filler. In SDH the pressure is higher, because sound descriptions and speaker labels compete for the same space, and languages that run longer than the source need even more careful condensing. That is why a verbatim transcript is not a subtitle file.
Multilingual subtitles and SDH in several languages
Once a video library goes international, every new language is a new track, and every edit of the video affects every track.
Start from a clean source. Accurate source-language captions are the best basis for translation; an uncorrected automatic transcript carries every error into every language. The source timing is often used as a template so that all languages share the same segmentation, with small adjustments where needed.
Keep terminology consistent. Product names, interface labels and technical terms should match your manuals, your website and the software itself. If the subtitle calls a button by one name and the screen shows another, viewers notice. The same terminology control that keeps technical content consistent across suppliers applies to subtitles, and product demos that film a user interface benefit from the same glossary used for UI translation for HMI and machine software.
Adapt the sound descriptions. SDH conventions differ between countries, so sound cues and speaker labels have to be adapted to the target market’s conventions, not just translated word for word.
AI auto-captions vs professional captioning
Automatic speech recognition has improved enormously, and auto-captions are now built into most video platforms and editing suites. For a first draft, a rough internal recording or a quick search index, they are useful. For published company video, they have predictable weaknesses.
Recognition errors cluster where they hurt most: product names, technical terms, numbers, accents and overlapping speech. Automatic captions rarely identify speakers or describe sounds, and often break text at pauses rather than meaning. Machine-translated subtitles built on them inherit every error. General-purpose assistants such as ChatGPT can help draft or condense text, but they do not watch the video, check the timing or know your terminology.
The approach that works for most companies is a hybrid: automatic transcription as a starting point, then a qualified subtitler who corrects the text, adds speaker identification and sound information, condenses for reading speed and checks the result in the video. Accessibility guidance generally treats uncorrected automatic captions as insufficient, because a caption that is only mostly accurate is not an equivalent. Where machine output stops being good enough is a broader question, discussed in our guide to post-editing machine translation for technical content. How that hybrid route is priced, and where automatic subtitles stop paying off, is broken down in our guide to subtitling rates.
The European Accessibility Act, briefly
The European Accessibility Act, Directive (EU) 2019/882, has applied through national laws since 28 June 2025 to specific consumer products and services, including e-commerce, consumer banking, e-books, passenger transport and services giving access to audiovisual media. For services in scope, websites and apps must be perceivable, which in practice means captions for prerecorded video published after that date.
Many companies producing training or B2B video are not directly in scope, but their customers or consumer channels may be. Our guide to European Accessibility Act requirements explains who is covered. The manuals and reports published next to your videos raise the same questions, and a practical guide on how to make a PDF accessible covers that side.
Training, product and marketing videos: what each one needs
Different video types put the weight in different places, and a single policy should reflect that.
Training and e-learning. Employees must understand the content, so SDH in the source language, plus SDH or subtitles in each language your workforce uses, is the sensible default. Accuracy of instructions matters more than style, and safety training should never rely on uncorrected automatic captions.
Product videos and tutorials. These sit on product pages, in support portals and sometimes inside the software. Their captions must match the names on the product and in the manual, and sound cues such as alarms or confirmation signals are often essential information. Sidecar WebVTT files keep the video reusable across languages, and they fit the software localization process when tutorials are released with each product update.
Marketing and brand films. Here tone, rhythm and brand voice matter as much as accuracy, and much of the audience watches without sound. A common set-up is burned-in captions for social cut-downs plus switchable tracks on the website and video platforms. Taglines and claims may need adaptation rather than translation, which is the territory of marketing translation and transcreation.
Webinar recordings. If you publish a webinar afterwards, plan corrected captions for the on-demand version, because live output contains errors that are tolerable in the moment but not in a permanent asset.
How to brief a supplier
Most problems with subtitles start with a brief that says only “subtitles in four languages”. A short, complete brief saves several rounds of corrections.
- Audience and purpose: translation only, accessibility, or both, and for which languages.
- Deliverable: SDH, captions or translated subtitles, with the conventions to follow for speaker labels and sound cues in each language.
- Format and platform: SRT, WebVTT or a broadcast format, burned-in versions if needed, and where each file will be published.
- Source material: final video (not a rough cut), any script or transcript, on-screen text that must be translated, and your glossary of product names and terms.
- Review: who on your side approves terminology, and whether you want the subtitles checked in the video before delivery.
Ask the supplier, too, who reviews the files in context and how a re-edit of the video will be handled. The broader questions in our guide on checking a supplier’s quality before you hire them apply here as well.
How DEMA Solutions creates subtitles and SDH
Our subtitling and subtitle translation service is built on three core steps: transcription of the source audio, translation or adaptation by a qualified linguist, and segmentation and synchronisation to speech, scene changes and reading rhythm. Around that core, projects can include review in context with the video, closed captions, SDH, and delivery either as separate subtitle files or encoded into the video.
For SDH, we transcribe the complete audio and agree the conventions for speaker labels, sound descriptions, music and formatting with you before work starts, then adapt both dialogue and sound information for each language. The SDH track is checked in context, and where appropriate we work with associations supporting d/Deaf and hard-of-hearing viewers to test accessibility and readability. Translation follows our ISO 17100 certified process, delivered by our teams in Italy, Serbia and Spain, and your glossary keeps product names and terms consistent with your other content.
SDH can also sit within a wider accessibility programme, alongside accessible PDFs, plain language and Easy-to-Read versions, which our accessibility and inclusive language services bring together.
Frequently asked questions
Both make speech and relevant sounds available to viewers who cannot hear the soundtrack. Closed captions come from broadcast tradition, are usually in the source language and can be positioned on screen. SDH is delivered as a subtitle file, looks like ordinary subtitles and can be produced in the source language or in a translation.
Subtitles for the deaf and hard of hearing. Besides the dialogue, they identify speakers when this is not obvious and describe sound effects and music that are relevant to understanding the video.
Not usually. Translated subtitles render dialogue only, while WCAG defines captions as covering both speech and the relevant non-speech audio. For accessibility, you need captions or SDH.
Criterion 1.2.2 Captions (Prerecorded) at level A and criterion 1.2.4 Captions (Live) at level AA. Both appear in WCAG 2.1 and 2.2 and, through EN 301 549, in the European accessibility benchmark.
SRT is the most widely accepted and works on almost every platform. WebVTT is the web standard, supports positioning and speaker tags, and can be labelled as captions in HTML players. Check which formats your platform accepts; many companies keep both.
As a draft, yes. For published video, automatic captions need correction by a person, because they often get names, terms and numbers wrong and rarely identify speakers or describe sounds.
Only for short clips on platforms that play without sound. Burned-in captions cannot be switched off, resized or read by assistive technology, and each language needs its own video file.
Make your videos accessible in every language
If you are planning training, product or marketing video for several markets, tell us which videos, languages and platforms you need. We will recommend the right mix of SDH, captions and subtitles and plan the work around your publishing schedule.
How can we help?
Tell us what you need and we’ll get back to you with the solutions.