Standards for Video Accessibility

Session Notes

Session Type: Working group

Session Category: Open Media Developers

Session Leader: Silvia Pfeiffer (Google), Philip Jagenstedt (Opera)

Day: Saturday
Time: 4:00pm – 5:30pm
Room: Faculty Commons


Video and audio have become native Web citizens through their inclusion into HTML5 and implementation in modern Web browsers. Media accessibility approaches have been specified in HTML5, but not yet been included into any released browser. This includes subtitles and captions as much as audio descriptions and sign language video tracks.

In this session we will discuss the new WebVTT file format which provides timed text for video and audio elements. We will discuss limitations that implementers have come across and new challenges that haven’t yet been resolved.

HTML5 video accessibility also has implications on accessibility features of desktop video applications. So, another focus of the session will be to analyze requirements and limitations that desktop (and more generally non-web) applications face in the new HTML5 video environment. One issue to consider in this context will be the authoring of WebVTT files.

We will also discuss the HTML5 multitrack specification which will allow publication of audio descriptions and sign language video in sync with a main video and determine obstacles for its implementation.

Another topic of interest could be about whether it makes sense to add WebVTT as an in-band track to the open WebM and Ogg file formats.

Finally we will discuss other requirements that users of open video have for video accessibility and see if technical solutions can be found.


This session will make progress on the technical realization of video accessibility on the Web.

One possible outcome could be the registration of new bugs on the HTML5 specification, and of bugs on browsers.

Another possible outcome could be the creation of a list of further requirements for video accessibility and video accessibility applications.

Other documents of recommendation or best practices could also result from this session.


Potential topics:

  • WebVTT syntax ‒ anything missing that we should add in a hurry?
  • WebVTT Style and rendering ‒ is it any good? Overkill to support multiple tracks natively?
  • <track> UI ‒ standardized icons for indicating availability? [CC] icon?
  • UI for enabling multiple tracks?
  • <track> Chapters ‒ should the end point be significant?
  • Transcript ‒ can it be built into browsers in a sane way?
  • Descriptions ‒ synthesis?
  • WebVTT in WebM


  • searchability?
  • chaptering and subchaptering, also useful for metadata-based access
  • url link to chapter
  • thumbnails are great for navigation
  • audio descriptions delivered as text descriptions
  • concerns of desktop players
  • why was a new format introduced? Why not leveraging what is already available (e.g. matroska external entity support and ASS subtitle format)
  • need to support not just subtitles & captions, but also chapters, metadata & descriptions
  • better to have a separate file to the video to support many video formats
  • needs to be independent from media format
  • simplicity to translate between new format and e.g. SRT
  • current state of <track> element
  • multitrack API: @mediagroup
  • run scripts at certain time: @kind=metadata track
  • security considerations: CORS & same-origin; no cross-origin xml file like Flash?
  • events are great
  • voice tag: how to expose special styling? apply class to <v> tag
  • why have a "WEBVTT" magic string at the file start? file type identification
  • WebVTT in WebM
  • Create tracks later: add <track> later or use MutableTextTrack
  • @srclang is not on audio or video, so missing for multitrack API