Accessibility

Topics:

  • Progress on video description? The FCC is going to start requiring shows to provide video descriptions. When the shows move online, we will need to support them, too.
  • JWPlayer? - lots of accessibility improvements for screen readers, alt info, keyboard control; has anyone encountered issues e.g. mouse vs keyboard events
  • Using textracks for things other than captions
  • Kaltura is about to release a new version of the player; what should we be paying attention to?
  • Audio description and visual descriptive services; Dolby codes provide support for that, in supporting broadcaster through government mandates for audio description; here to learn about gaps we can help with.
  • Ideas for improving a11 in YouTube? players
  • Chapter support (useful for navigation)
  • Audio description, chapters, text track APIs?, making the player controls accessible
  • Any progress on Standardization of keyboard control of player, how to move focus to player? Any standardization among screen readers
  • Descriptive audio file formats

Player accessibility:

  • We discussed standardization last year, but no one seems to have picked it up at all.
  • Certainly usage seems to have become common (space bar for pause, left and right arrow keys for forward and backward.
  • Tooltips
  • YouTube? Player, JWPlayer?, VideoJS? as good examples
  • Focus highlight expectations.
  • Catering for keyboard vs mouse users - how to tell?
  • Media accessibility user requirements has some indication of what is expected, but no in depth technical details of how. Maybe this level of detail should be added? Document is already released as a standard.. SIlvia? suggests it would be better to write a best practices document.
  • Any blindness disability groups involved? They participated in Media Accessibility Requirements..
  • Someone needs to start an effort with a standards organization. Maybe as part of Video Dev github? Volunteers from JWPlayer?, Chrome, VideoJS?

Audio descriptions:

  • This could be audio snippets that get played, or text that is read out via TTS
  • WebVTT targets the latter approach. Chrome extension demonstrates this approach.
  • Broadcast media is likely to be the former approach.
  • Providing text allows you to control the voice, volume, can translate, etc.
  • Does anyone have examples of videos with audio descriptions? Silvia gave a talk, and the video is uploaded to YouTube? (one with audio description and one without). Search for Silvia and Audio Description on YouTube?
  • The biggest challenge is content authoring. It is very expensive. This is one thing that the text approach helps with.
  • Is there a format for representing the audio approach? WebM? and mpeg4 have containers for signalling this usage.
  • Netflix just has an alternate audio track. It doesn't try to merge them. It is extra metadata that signals that the other track is an audio description.
  • A quick scan of the specs didn't turn up anything; it may not be in the core specs. In Metroshka?
  • With web material, you don't need the separate audio track; you would select the presentation and it would be mixed.
  • What happens if someone skips to the middle of the video? Will they miss the description? No, it is synchronized / mixed in to the same timeline.
  • There are authoring challenges and skills in producing audio descriptions.
  • Extended audio descriptions will pause the video if necessary to extend the pauses.
  • Action: make sure the underlying container formats have provisions for multi-;language descriptive audio tracks, particularly in audio representation.
  • Silvia is confident that WebM? spec contains this support for WebVTT text tracks.
  • Media controller.js library somewhere will implement the spec for merging audio track.
  • VideoJS? exposes text track to screen readers and lets them read it out. But by default, captions are hidden from screen reader. Ask browser vendors to make captions visible to screen readers if kind is audio description.
  • Caldera used to support audio descriptions, but now must support multiple audio tracks where one includes audio descriptions. Studio provides a track with the audio descriptions mixed in. (No mixing in the player.)
  • JWPlayer? - has seen (already mixed) audio tracks for audio descriptions.
  • None of the browsers implement support for text audio descriptions; unless it is a standard feature in browsers, won't be adopted by authors.
  • Discussion about whether extension support is sufficient as an accessibility solution.
  • Dialog enhancement is ability to adjust the dialog independent of the rest of the scene. e.g. listening to commentators over the noise of a game.
  • In the studio production flow, the broadcast content is premixed. Customers who are less professional are asking them to do the mixing.
  • If tracks are separate, the volume can be controlled separately.
  • IMF contains multiple tracks. It can be used any way you want, but it can support audio description, enhanced dialog, etc. It will be a solved problem on the content creation side.
  • Curious how companies got their players to synchronize the separate tracks.

Metadata

  • Spec doesn't say when to trigger events related to track start and track end. Different behaviors in different browsers. Underspecified - please file a bug.

Chapters

  • Chapters is not a requirement from media companies. Requested by training companies.
  • For vision impaired, chapters help with navigation on the timeline.
  • Last year we had a discussion around thumbnails, and lack of standardization. Thumbnails and chapters would be helpful, but no standardization push,
  • In W3C? specs, if features aren't used, they will be removed.
  • VideoJS? just shows a menu that can be used for navigation.
  • Metadata parts of WebVTT are underspecified, which leads to different uses/implementations.
  • Chapters could be specified more fully. When first introduced, different chapter features were very different. But it seems to be converging at this point, and we could specify the format of a chapter track.
  • How about a thumbnail kind? Probably a WebVTT v2 feature - please get involved!

Screen reader standards

  • Keyboard control of the browser. Last year, discussed desire to be able to move focus to the video element, instead of needing to TAB to it.
  • Screen readers support jumping to the <main> element. Put that mark-up around player?
  • These suggestions probably belong in a best practices document.
  • Should screen readers be visible to applications? It is currently tricky to handle events in JS.
  • Chrome extension to play WebVTT audio description tracks using Text to Speech:
      https://chrome.google.com/webstore/detail/html5-audio-description-v/jafenodgdcelmajjnbcchlfjomlkaifp?utm_source=chrome-app-launcher-info-dialog
  • Test web site containing a video with an audio description track:

https://gingertech.net/external-videos/test-video-for-audio-descriptions-elephants-dream/