Speaker info for WebVTT subtitles

[ashiagr]

We're using WebvttParser() to parse WebVTT transcripts, which internally uses WebvttCueParser() to parse cues.
While WebvttCueParser.parseCueText(...) provides the payload with style formatting, it removes the speaker identifiers or the TAG_VOICE.

Can we get an in-built support for speaker info within the cues of WebVTT transcripts?

[icbaker]

I guess you're referring to this? https://www.w3.org/TR/webvtt1/#webvtt-cue-voice-span

We do have some code already that parses <v[voice="foo"]> tags in order to resolve the associated styles: https://github.com/search?q=repo%3Aandroidx%2Fmedia+path%3Avtt+voice&type=code

But the voice information isn't directly exposed in the resulting Cue object.

I suspect this would need to be exposed using a 'custom span' in Cue.text, like we do for Japanese rubies: https://github.com/androidx/media/blob/release/libraries/common/src/main/java/androidx/media3/common/text/RubySpan.java

I'll mark this as an enhancement. I'm afraid we are unlikely to work on this ourselves soon, but we would consider a high quality PR implementing this.

[icbaker]

Closing this because #1652 has been merged - thanks!