Post

Replies

Boosts

Views

Activity

WebVTT positioning cues
Hello!I'm trying to implement creation of HLS packages with one or more WebVTT tracks that are created from upstream subtitle / closed caption masters that contain positioning data.Simple positioning works fine in Safari. However, if we have multiple on-screen subtitles with the same timing information but different positions, subtitle playback breaks - specifically some subtitles are skipped, and subtitles frequently go out of sync.This type of content is common when two actors are haing a conversation. It is handled properly by Android's ExoPlayer and Chrome when playing the same WebVTT files via MPEG DASH.Simply discarding the positioning data and collapsing these formerly separate subtitles does work, but then we lose the nice user experience of having separately positioned subtitles.e.g.00:06:25.468 --> 00:06:28.722 align:start line:69% position:26% size:22%Mmm. I hate itwhen that happens.00:06:25.468 --> 00:06:28.722 align:start line:85% position:76% size:9%Me too.Works in Android and Chrome, but breaks in Safari / native iOS. If I dumb it down to the following it definitely does work00:06:25.468 --> 00:06:28.722Mmm. I hate itwhen that happens.Me too.Does Apple support multiple simultaneous subtitles with WebVTT? If not, then it seems like I just have to discard positioning data.
4
0
2.8k
Apr ’17