Minutes from Media Timed Events call 27 September 2021

Dear all,

The minutes from the Media & Entertainment IG, Media Timed Events call on unbounded cues in WebVTT are available [1].

Summary:

1. Unbounded cues in segmented WebVTT can be represented as a sequence of bounded cues, repeated in each segment. In the case where the cue spans an entire segment, the cue start and end time matches the corresponding segment..

  * Write this up as a worked example

  * Any special considerations for low latency DASH?

2. We can write client processing rules that describe how to combine the sequence of bounded cues into one contiguous cue, so that there's no discontinuity between segments. This could be a Working Group Note.

3. There may need to be a way of identifying the same cue across different segments, which are separate WebVTT documents. Further discussion needed.

4. Are there requirements for other delivery formats for WebVTT aside from segmented media that may still need a syntax for unbounded cues, e.g., real-time delivery over WebSockets?

Next meeting date: Monday 11 October at 15:00 UTC

Many thanks,

Chris (for the MEIG co-chairs)

[1] https://www.w3.org/2021/09/27-me-minutes.html