The Gold Standard of Audio Description Services | Meaningful Visual Access and Faithful Narration

Audio description makes important visual information understandable through spoken narration for people who cannot see a video adequately. Its gold standard is faithful, useful access—not nonstop commentary or a pleasant voice that misses the lesson’s main point.

This guide follows the fictional educational video Rainwater in Our City, where a crucial water-filtration diagram is visible but unexplained. It complements Audio Post-Production for finished soundtracks, Voiceover Production for recording performance and Educational Publishing for instructional meaning.

Find the part you need

Choose a chapter group or follow the process from content analysis to verified accessible playback.

Chapter index

1. An educational video can leave its main lesson invisible

A fictional community learning centre publishes a short video called Rainwater in Our City. The presenter explains that a storm’s runoff is moving through a diagram, then says, ‘This is where the most important change occurs.’ The animation shows water flowing into a filter and a gauge changing colour, but neither action is explained in the narration. A viewer who cannot see the diagram hears a confident conclusion without learning what actually happened.

The gold standard of audio description services is to make necessary visual information accessible through accurate, useful spoken description without distorting the work’s meaning. It is not simply reading every visible object aloud or making the soundtrack louder. The centre and video are fictional. Their problem illustrates how audiovisual information can remain inaccessible even when the narration is clear and the presenter communicates warmly.

Back to chapters · Next chapter

2. Determine whether additional description is actually needed

Audio description supports people who are blind or cannot adequately perceive visual information in a video. It provides actions, on-screen text, relationships and other details needed for comprehension when those details are not already communicated in the main sound. The requirement and best method depend on the material and target platform.

W3C’s guidance on describing visual information notes that when the original audio already conveys all important visual content, separate description may be unnecessary. The centre should first review what the presenter already says and what a listener can infer. Adding redundant commentary risks overwhelming the story instead of improving access. The service begins with a content assessment, not the assumption that every video needs a constant second narrator.

Back to chapters · Next chapter

3. Identify the actual information the viewer needs

Not every colour, gesture or background object has equal importance. The describer should understand the video’s purpose and identify visual events essential to the audience’s decisions or learning. In a drama, a meaningful glance may change character relationships; in a science explanation, a graph axis or step of a process can be the core information.

For Rainwater in Our City, the path of water into a filtration unit and the changing gauge may be crucial. The presenter waving a hand beside the chart may not need mention unless that gesture conveys something otherwise unavailable. A good script ranks visual information by significance and relevance. The objective is comprehension equivalent to the intended material, not a competition to describe every frame.

Back to chapters · Next chapter

4. Distinguish integrated and separate audio description

An educational video can sometimes include necessary descriptions naturally in its main presentation. This is integrated description: the presenter says what changes in the diagram as part of teaching. A finished film may instead need an optional described audio track or a separately described version.

W3C’s media planning guidance recommends considering visual description during scripting and production. It can be easier to build into a training video before filming than to squeeze into narrow gaps afterward. Decide what the actual player can support and which experience best serves users. Do not advertise selectable audio description when the delivered platform has no means to select or play it.

Back to chapters · Next chapter

5. Write in plain language that preserves context

A useful description names the relevant action or visual relationship without inventing motives. It should be concise, specific and understandable alongside the existing audio. Present tense and straightforward wording can help describe the unfolding image, while complicated technical material may need carefully prepared terminology.

The fictional gauge changes from one labelled condition to another. The describer should explain the label or meaningful value rather than announce that ‘everything becomes safe’ if the video never established that conclusion. W3C’s description guidance advises objective statements focused on needed information. A good writer checks factual meaning with the original content owner rather than substitute assumptions about why a person moved or what a graph proves.

Back to chapters · Next chapter

6. Name on-screen text and graphical meaning appropriately

Titles, speaker names, dates, statistics, diagram labels and links sometimes appear only in the picture. If they matter for understanding or action, an accessible presentation must make their information available through a usable channel. The words need not always be reproduced letter for letter, but material meaning must not disappear.

Rainwater in Our City displays a diagram labelled ‘Runoff’, ‘Filter’ and ‘Outflow’ while the narration discusses ‘this stage’. A reader of audio alone needs enough relationships to understand the process. If the video shows a contact address or registration deadline, include the practical information rather than merely say ‘text appears on screen’. Validate important numbers and units against the approved visual material.

Back to chapters · Next chapter

7. Place description without covering speech or meaningful sound

Standard audio description often uses natural pauses in dialogue and relevant sound effects. Timing can be difficult when important visual events occur during continuous narration or music. Inserting narration over every available millisecond may make the soundtrack harder to follow for the people the service intends to support.

W3C Technique G78 describes user-selectable audio containing visual description in pauses, while Failure F113 addresses omitting description when suitable pauses are available. The service should plan scripts and timing deliberately, preserving important original audio. If a complex chart cannot be explained in the available gap, the producer should consider a different appropriate format instead of speaking impossibly fast.

Back to chapters · Next chapter

8. Know when extended description is the sensible option

Some sequences contain continuous speech with crucial visual information. Extended audio description may pause video playback to provide additional time for explanation, then resume. This can be useful for detailed educational processes, particularly where a graph or demonstration is essential and cannot be summarised accurately in a brief interval.

WCAG 2.2 Success Criterion 1.2.7 describes extended audio description at Level AAA for defined cases. That is different from the Level AA requirement for prerecorded video under Success Criterion 1.2.5. Not every media player supports extended description, and users should understand when playback will pause. A quality provider checks platform compatibility rather than promising a technique that the actual viewers cannot use.

Back to chapters · Next chapter

9. Cast and direct narration for clarity, not personality

Audio description may be voiced by a separate narrator or integrated into a presenter’s speech. Where a separate narrator is used, the voice should be distinguishable from other speakers, paced appropriately and guided by the video’s intended tone. It should not compete with the performers for attention or invent emotional conclusions.

The fictional science video could use a calm narrator who describes the diagram in short gaps. For drama, description must respect performance and avoid announcing a character’s private feelings without visible evidence. The W3C audio-description resource recommends a distinct voice and objective narration. A recognizable voice alone is insufficient if the script fails to convey the important information or the mix obscures it.

Back to chapters · Next chapter

10. Build description from an approved picture version

A visual change can invalidate an earlier description even when the audio script remains plausible. A chart may be revised, a shot removed or an animation timed differently. The describer needs the accepted picture version and a clear process for recording and implementing changes.

Rainwater in Our City receives an updated diagram in which the gauge labels are reordered. A description recorded for the previous cut could now teach the wrong sequence. Track picture identity, approved wording and relevant timings through the production handoff. When the author corrects a scientific statement, notify the description and caption teams together. A current audio file should never be assumed to match the newest video merely because both were uploaded on the same day.

Back to chapters · Next chapter

11. Mix description so that it is audible and respectful

Description needs suitable balance with dialogue, music and meaningful effects. Making it dramatically louder than everything else can be uncomfortable, while placing it too low can render it ineffective. The actual listening experience depends on the media, supported devices and any alternate-track or integrated-mix arrangement.

W3C’s combining-audio guidance describes adjusting original programme audio so description remains understandable when necessary. Such treatment should preserve important context and not erase the soundtrack’s artistic purpose. Review on suitable headphones, small speakers or other relevant devices. A clean solo description recording is not proof that the final accessible mix can be followed.

Back to chapters · Next chapter

12. Verify player behaviour, not just the exported sound file

A separate audio description file may be technically valid while the player does not expose a selection control, mislabels the language or fails to remain synchronised with picture. Accessibility must be validated at the user-facing interface. A claim of compliance should match the actual delivered experience.

W3C Technique G78 specifically considers a user-selectable described soundtrack. It also notes that the presence of an audio file alone is insufficient; the user must be able to select it. The centre should test playback with relevant assistive technologies and devices where possible. A player that starts the correct video but defaults to an undescribed audio track without a discoverable alternative may defeat the service’s purpose.

Back to chapters · Next chapter

13. Work with people who use description in real situations

Technical and editorial teams can produce a script that seems complete while omitting information a blind viewer actually needs. Representative user feedback can reveal uncertainty, confusing timing or unnecessary repetitions. A useful review asks what the person understood rather than whether they liked the narrator’s voice.

Give reviewers a task appropriate to the video: explain the route water follows, identify what a gauge reading means or state the action recommended at the end. Avoid coaching them through the diagram in advance. Record where the description failed and revise the real cause. A small test cannot certify universal accessibility, but it can demonstrate concrete barriers and support a stronger service than internal approval alone.

Back to chapters · Next chapter

14. Make obligations and levels of conformance precise

Accessibility requirements are shaped by applicable law, organisational policy and the media platform. The Web Content Accessibility Guidelines distinguish media types, prerecorded versus live content and different conformance levels. A provider should not reduce them to a statement that every video must use the same described track.

WCAG 2.2 Success Criterion 1.2.5 requires audio description for prerecorded synchronized video at Level AA when the important information is not already in the audio. At Level A, Criterion 1.2.3 provides a defined choice involving description or an alternative. Extended description has a separate Level AAA criterion. WCAG 3.0 remains draft work and should not be represented as a final replacement for WCAG 2.2. The applicable legal duty still depends on jurisdiction and service.

Back to chapters · Next chapter

15. Protect licences, performers and sensitive source material

Described media may contain copyrighted film, private scenes, identifiable people or unreleased information. An audio-description provider usually receives picture and dialogue for an agreed purpose, not unrestricted permission to distribute it or train a synthetic voice. The narrator’s recording may also have its own contractual terms.

Keep a clear rights record for original video, source files, voice performances, translations and any described versions. Secure unpublished media and share only what another production team legitimately needs. WIPO’s copyright overview provides general rights context, with actual permissions governed by law and contracts. Accessibility is a legitimate goal, but it does not automatically erase rights or personal-data obligations.

Back to chapters · Next chapter

16. Deliver content, evidence and an accountable correction route

The accepted package may include a described soundtrack, a described video version, a script, timing data, language metadata and related accessible text, depending on the platform and agreement. A professional handoff should state what was tested and what remains limited. An incorrect clip, missed on-screen deadline or inaccessible selection control needs a practical correction owner.

For the fictional video, the learning centre should receive a master linked to the right picture, evidence that the described playback is usable and a way to update narration if the science diagram changes. Preserve the prior approved version and identify the new release clearly. A file called ‘AD_final.wav’ without a version, source or playback instruction does not establish a complete accessible service.

Back to chapters · Next chapter

17. The gold standard is the same meaningful lesson through another route

Return to Rainwater in Our City. A listener who cannot see the animation can now follow the runoff to the filter, understand which gauge value changes and hear the essential visual text. The words are integrated at an appropriate pace or provided through a functioning alternate version. The centre has checked understanding and can trace which video the description serves.

Audio Description Services earn their gold standard by providing meaningful visual access without turning a film or lesson into something it never was. Accuracy, timing, dignity, effective playback and feedback matter together. The achievement is not extra narration for its own sake: it is allowing more people to understand the work as its creators intended.

Back to chapters

Worked example: the water moves, but the listener cannot see where

This fictional learning video separates the intended visual lesson from what its original soundtrack actually communicates.

Audio description acceptance checklist

• Analyse what essential visual meaning is absent from the existing audio.
• Choose integrated, separate or extended description appropriate to the media.
• Write concise, objective statements with correct on-screen facts.
• Protect dialogue and meaningful sound when timing extra narration.
• Record identifiable, clear performance suited to the material.
• Test user-selectable playback, captions and the actual visual-version match.
• Gather representative feedback from people who use description.
• Document conformance scope, permissions, changes and verified delivery.

Official accessibility guidance and connected eduKate articles

• W3C WAI — Description of Visual Information
• W3C WCAG 2.2 — Success Criterion 1.2.5 Audio Description (Level AA)
• W3C WCAG 2.2 — Extended Audio Description (Level AAA)
• W3C — User-Selectable Audio Description Track, G78
• W3C — Planning Accessible Audio and Video
• W3C — Accessible Media Player Considerations
• eduKate — Audio Post-Production
• eduKate — Voiceover Production
• eduKate — Educational Publishing
• Gold Standard Sectors and Services Directory

WCAG 2.2 has defined media types and conformance levels; a technique is not a mandatory implementation method, and local legal accessibility obligations differ. Audio description should not be confused with captions, dubbing or transcripts. This article explains responsible service design rather than certifying a specific video or media player.