Measure caption readability through one injected, identity-bearing runtime
and judge each cue against its language's profile.
The grapheme runtime is injected because the host owns it: production binds
Node's installed Intl.Segmenter and reports that exact identity. Each cue is
measured on its canonical presentation: graphemes, lines, the longest line,
duration, the gap since the previous cue of the same language identity in
track order, and graphemes per second at timeline.fps. A cue whose language
has no profile, or whose profile names a segmentation identity other than the
runtime's, keeps its measurement with a not-run outcome instead of a
verdict. Profiles are matched by ASCII case-insensitive language identity;
design validation refuses two profiles for one identity, so none is shadowed.
The track must be in start order with positive durations, as film assembly
guarantees; a zero-length cue would report an infinite rate.
Measure caption readability through one injected, identity-bearing runtime and judge each cue against its language's profile.
The grapheme runtime is injected because the host owns it: production binds Node's installed
Intl.Segmenterand reports that exact identity. Each cue is measured on its canonical presentation: graphemes, lines, the longest line, duration, the gap since the previous cue of the same language identity in track order, and graphemes per second attimeline.fps. A cue whose language has no profile, or whose profile names a segmentation identity other than the runtime's, keeps its measurement with anot-runoutcome instead of a verdict. Profiles are matched by ASCII case-insensitive language identity; design validation refuses two profiles for one identity, so none is shadowed. The track must be in start order with positive durations, as film assembly guarantees; a zero-length cue would report an infinite rate.