EAA Video Accessibility: What Captions, Audio Description, and Accessible Players Actually Require
If your organisation publishes video to EU audiences - whether on a streaming platform, a bank's help centre, an e-commerce product page, or a mobile app - the European Accessibility Act is not a future concern. The EAA has been enforceable since 28 June 2025 across all 27 EU member states. Enforcement is ramping through 2026 as national market-surveillance authorities get up to speed, and the obligations are specific.
This guide cuts through the noise. It explains which video is in scope, what the concrete technical requirements are for captions, audio description, and accessible media players, and how live content differs from pre-recorded. It does not duplicate our existing posts on e-commerce, mobile apps, or e-books - but it does show how video obligations sit inside those broader compliance pictures.
Is your video even in scope?
The EAA covers two distinct populations of video publishers, and the obligations differ between them.
1. Audiovisual media services (AVMS)
Streaming services and on-demand providers - platforms whose primary purpose is to inform, entertain, or educate the general public through video - fall under the audiovisual media services category. This includes VOD platforms, broadcaster catch-up services, and the apps and websites through which users access them. For these organisations, the EAA works alongside the Audiovisual Media Services Directive (AVMSD), which requires media service providers to make their services continuously and progressively more accessible. The EAA adds a harder floor: the interface through which users find, choose, and watch content must itself meet EN 301 549 / WCAG 2.1 AA, and access services (subtitles, audio description, sign language) must be transmittable and user-controllable.
Article 32 of the EAA allows a five-year transitional period for audiovisual media services to bring existing back-catalogue content into conformance - meaning legacy content must be remediated by 2030. New content published after 28 June 2025 is subject to the full requirements immediately.
2. Video embedded in other in-scope products and services
This is the category that catches many organisations off guard. If your website, mobile app, e-commerce store, or banking service is in scope under the EAA - and for most consumer-facing digital services it is - then video embedded in those products is also in scope. That includes product demonstration videos, explainer clips, onboarding tutorials, and marketing content hosted on your own player.
Pre-recorded time-based media published before 28 June 2025 on websites and mobile apps is exempt from accessibility requirements under Article 2 of the EAA - but this exemption does not apply to audiovisual media services. For content published after that date, the full requirements apply immediately.
Third-party content you do not control - such as a user-uploaded video embedded from a platform you neither fund nor manage - may be exempt. If you license or control the content, you are not exempt.
Captions: what the standard actually requires
Captions are the most visible video accessibility requirement, and also the most frequently misunderstood. The EAA's technical baseline is EN 301 549, whose Clause 9 incorporates WCAG 2.1 Level AA in full. Under WCAG 2.1 AA, two success criteria govern captions directly:
- SC 1.2.2 Captions (Prerecorded) - Level A: Captions must be provided for all pre-recorded audio content in synchronised media.
- SC 1.2.4 Captions (Live) - Level AA: Captions must be provided for all live audio content in synchronised media.
What does a compliant caption actually contain? The W3C WAI is precise: captions are "a text version of the speech and non-speech audio information needed to understand the content." That goes well beyond a transcript of dialogue.
A compliant caption file must include:
- All spoken dialogue, accurately transcribed and synchronised with the audio
- Speaker identification - who is speaking, especially when speakers are off-screen or multiple speakers overlap
- Relevant non-speech sounds - sound effects, music cues, laughter, applause, alarms, and other audio information that carries meaning
- Audio cues that affect comprehension, such as
[tense music]or[door slams]
Captions that contain only dialogue - the default output of many auto-generated tools - do not meet the standard. Captions that are accurate but poorly synchronised do not meet the standard either.
Subtitles ≠ captions. Standard subtitles translate or transcribe dialogue only. Captions for accessibility (sometimes called SDH — Subtitles for the Deaf and Hard of Hearing) must also include speaker identification and non-speech audio information. If your platform only offers translation subtitles, you have a compliance gap.
Language coverage
The EAA does not enumerate required languages, but the obligation is clear in practice: if you publish content in a specific language for a given EU market, the captions must be in that language. English-only captions do not satisfy the requirement for non-English-speaking markets.
Audio description: making visuals accessible to blind and low-vision viewers
Audio description (AD) is narrated commentary - delivered in the natural pauses between dialogue - that describes key visual actions, scene changes, settings, and on-screen text that are not conveyed by the main audio track. It is the primary access service for blind and low-vision viewers.
Under WCAG 2.1 AA, SC 1.2.5 Audio Description (Prerecorded) requires audio description for all pre-recorded video content in synchronised media. The W3C WAI describes the scope precisely: audio description provides "information about actions, characters, scene changes, and on-screen text that are important and are not described or spoken in the main sound track."
Audio description is delivered during existing pauses in dialogue; where those pauses are insufficient, extended audio description (WCAG SC 1.2.7, Level AAA) requires the video to be paused to allow the description to complete.
Practically, this means:
- A product video that shows a user interface without narrating what is on screen requires audio description of the UI elements
- A talking-head interview where the speaker describes everything they are doing may not require additional description
- On-screen text - titles, captions, data labels, URLs - must be read aloud or described if they carry information not conveyed in the audio
For in-scope services, audio description tracks must be transmittable through the full delivery chain. Captions and audio description must be preserved whether users access content via a web browser, mobile app, smart TV application, or other platform.
Transcripts
For audio-only content (podcasts, audio guides), a text transcript is required at WCAG Level A. For video content, a descriptive transcript - one that combines the caption text with descriptions of the visual information - is strongly recommended and is required to serve users who are both Deaf and blind. Descriptive transcripts are straightforward to produce once you have caption files and audio description scripts: the two can be merged into a single accessible document.
The accessible media player: the layer most teams overlook
Having excellent captions and audio description is not enough if users cannot access or control them. The EAA requires that the player interface itself be accessible - perceivable, operable, understandable, and robust (POUR), the four principles of WCAG.
Concretely, a compliant media player must allow users to:
- Turn captions on and off without requiring a mouse
- Adjust caption display characteristics - text colour, background colour, font size, font family, and text borders - so users can adapt captions to their visual needs
- Access audio description tracks through a clearly labelled, keyboard-operable control
- Operate all player controls (play, pause, seek, volume, full-screen) by keyboard alone and with a screen reader
- Perceive all controls - buttons must have accessible names, focus indicators must be visible, and controls must not disappear on hover in a way that traps keyboard users
The caption and audio description controls must be at the same level of prominence and accessibility as the primary playback controls. A buried settings menu that requires a mouse to open does not meet the requirement.
Choosing a player: Not all video players meet these requirements out of the box. YouTube's embedded player, for example, does not natively support audio description tracks and has known keyboard navigation inconsistencies in embedded mode. Players with documented accessibility support include AblePlayer and OzPlayer. If you use a commercial player, request a VPAT (Voluntary Product Accessibility Template) or equivalent conformance report before committing.
The EAA also requires that accessibility features be "fully transmitted" through the entire content delivery chain. Captions and audio description must survive encoding, CDN delivery, and playback across every platform you support - not just on your primary web player.
Live vs pre-recorded: different challenges, same obligation
The WCAG 2.1 AA requirement for live captions (SC 1.2.4) applies to synchronised live media. For organisations streaming live events, webinars, product launches, or news content to EU audiences, this is an active obligation.
Live captioning presents a different quality challenge from pre-recorded work. The NER (Network Error Rate) model - the standard academic framework for measuring caption accuracy - classifies accuracy as follows:
In the largest study of English human and ASR broadcast captions across the US, UK, and Canada over four years (2018-2022), human captioners achieved 98.9%-99.4% accuracy, while ASR systems achieved 95.7%-96.3%. The practical implication: AI-only live captioning, without human review or correction, typically falls below the "acceptable" threshold under the NER model. For high-stakes live content, professional CART (Communication Access Realtime Translation) captioners or a human-in-the-loop ASR workflow is the safer path to compliance.
For pre-recorded content, there is no excuse for accuracy below 99%. The captioner can review and correct before publication.
Key differences between live and pre-recorded obligations:
| Requirement | Pre-recorded | Live |
|---|---|---|
| Captions | Required (WCAG SC 1.2.2, Level A) | Required (WCAG SC 1.2.4, Level AA) |
| Audio description | Required (WCAG SC 1.2.5, Level AA) | Not required at AA; AAA only |
| Transcript | Required for audio-only (Level A) | Not required at AA |
| Accuracy target | 99%+ (achievable with review) | Aim for ≥98.5%; use professional CART for critical content |
| Synchronisation | Tight sync required | Minimal delay; some latency is accepted given live nature |
Prioritised action checklist
If you are starting from zero, work through this in order. Compliance is not binary - demonstrating systematic progress matters to enforcement authorities.
List every video asset by platform, publication date, and content type. Separate new content (post-28 June 2025) from legacy content. Identify which assets are in scope under the EAA and which fall under AVMS-specific rules. Prioritise high-traffic, public-facing assets and any content in the critical user journey (onboarding, checkout, product demos).
Check whether existing captions include speaker identification and non-speech audio cues — not just dialogue. Run a sample through an accuracy check. Auto-generated captions from hosting platforms rarely meet the 99% accuracy standard and almost never include non-speech information. If you serve multiple EU markets, verify language coverage for each.
Test your player with a keyboard only (no mouse). Can you reach and operate every control? Can you turn captions on before the video starts? Can you adjust caption display settings? Can you activate an audio description track? Test with a screen reader (NVDA on Windows, VoiceOver on macOS/iOS). Request a conformance report from your player vendor if you use a third-party solution.
Identify which videos carry visual information not conveyed in the audio track. Commission professional audio description for those assets. Build AD production into your video production workflow so new content ships with description already recorded. Ensure your player and delivery pipeline can carry and surface the AD track.
For live video streams to EU audiences, decide between professional CART captioning, a human-in-the-loop ASR service, or — for lower-stakes content — a reviewed ASR workflow. Document your accuracy monitoring process. After each live event, review caption quality and retain records.
For podcasts, audio guides, and other audio-only content, publish a text transcript alongside the media. For video content, consider publishing a descriptive transcript — it serves Deaf-blind users and improves SEO as a side effect.
Add video accessibility acceptance criteria to your definition of done. Require captions, audio description, and player testing before any new video goes live. Schedule periodic audits of your video library, especially after player updates or CDN changes that could break caption delivery.
A note on the AVMSD and how it interacts with the EAA
For organisations that are audiovisual media services - broadcasters, streaming platforms, VOD providers - the regulatory picture involves two overlapping instruments.
The AVMSD (Directive 2010/13/EU as amended in 2018) requires media service providers to make their services "continuously and progressively more accessible" to persons with disabilities. Its Article 7 encourages subtitling, audio description, and accessible menu navigation, but its wording is broad and has historically led to divergent national implementation. The AVMSD does not impose specific quantitative targets or timelines for access service coverage.
The EAA adds a harder, more specific layer. It requires that the services used to find, choose, and watch video content - the apps, EPGs, and web interfaces - meet EN 301 549 / WCAG 2.1 AA. It also requires that access services (captions, audio description, spoken subtitles, sign language) be transmittable in high quality and properly synchronised, and that users have control over how those features display.
The practical result: AVMS providers face obligations under both instruments simultaneously. The AVMSD governs the content accessibility obligation (progressively more captioned and described content); the EAA governs the interface and delivery accessibility obligation (the player, the EPG, the app). Neither instrument alone is sufficient.
Disability rights organisations have called for the AVMSD revision - currently under evaluation by the European Commission - to introduce specific quantitative targets and timelines for access service coverage, which the current text lacks. That revision is expected to be informed by the Commission's evaluation due by December 2026. Watch this space.
Video accessibility under the EAA is not a single checkbox. It is a stack of obligations - accurate captions with non-speech information, audio description for visual content, an accessible player, and a delivery chain that preserves all of the above. The good news is that each layer is well-defined in EN 301 549 and the W3C WAI guidance. The work is real, but it is knowable - and starting now, with a clear inventory and a prioritised plan, is far less costly than remediation under enforcement pressure.
Related reading
Accessibility Conformance Reports (ACR/VPAT): The EU Procurement Guide for Vendors and Buyers
What a VPAT is, how it becomes an ACR, which edition to use for EU procurement under EN 301 549, and how to write one that survives scrutiny - plus a buyer-side checklist.
Web Accessibility Directive vs European Accessibility Act: Two EU Laws, One Compliance Programme
The EU has two distinct digital-accessibility regimes: the WAD for public sector bodies and the EAA for private-sector products and services. Many organisations are subject to both. Here's how to tell them apart - and how to satisfy both at once.
Cognitive Accessibility and the EAA: Why WCAG Compliance Isn't Enough
WCAG 2.1 AA is the EAA's legal floor - but it only partially covers people with cognitive and learning disabilities. Here's what teams must do beyond the standard.