How do I turn word-level ASR or forced-alignment timestamps into caption events that meet Netflix and BBC timing rules?

Problem. ASR returns words with start/end times, not caption events. Conditions. Word timestamps available (see C3-B636-12, C3-B844). Method. Inference-based pipeline anchored on documented rules: (1) group words into sentences/clauses by punctuation and long pauses; BBC prefers a single sentence per subtitle and clause-boundary splits for long ones. (2) Fill each event up to the line limit (for example 2 x 42 characters) but cut it earlier if the reading speed would exceed the cps limit (20 cps adult Netflix, 17 children) or if the event would exceed 7 s. (3) Set in-time at the first word start: Netflix puts it on the first frame of audio, 1-2 frames tolerance. (4) Set out-time to the last word end plus about half a second where no subtitle follows (Netflix); BBC allows up to 1.5 s hang-over. (5) Chain: where the gap to the next event is under half a second, extend the previous out-time to 2 frames before the next in-time (Netflix 'chaining'). (6) Enforce minimum 5/6 s by extending the out-time into free time, never by starting earlier than the audio. (7) If the gap between events is a real pause, leave it; BBC says a gap should be at least 1 s. (8) Validate cps and re-split or condense text if over. Example. Netflix timing page: 'a run of subtitles with even gaps' by raising out-time to two frames before the next in-time. Failure modes. Aligner word boundaries can be early or late by tens of milliseconds, so tolerance of 1-2 frames matters; ASR often drops disfluencies that verbatim rules require; BBC says viewers prefer verbatim so condensing is a last resort. Limitations. Steps 1-2 and 8 are inference; the sources do not define an ASR algorithm. Version/date. Netflix timing page (frame values for 24 fps; half second = 15 frames at 30 fps). Curator qualifications: All cited documentation excerpts located on the live pages by curator GET; no video sources. Limitations: Untested procedure. BBC and Netflix timing are made for human-timed files, so applying them to aligner output assumes aligner error is within 1-2 frames, which I did not measure. Evidence status: statements located in the cited sources by a correlated desk review (Claude); not executed; not independently reviewed.

{
  "citations": [
    {
      "url": "https://partnerhelp.netflixstudios.com/hc/en-us/articles/360051554394-Timed-Text-Style-Guide-Subtitle-Timing-Guidelines",
      "note": "1: Timing to audio; 1: Timing to audio; 1: Timing to audio; 5.2; 5.4; 3.2; researcher reasoning"
    },
    {
      "url": "https://www.bbc.co.uk/accessibility/forproducts/guides/subtitles",
      "note": "1: Timing to audio; 1: Timing to audio; 1: Timing to audio; 5.2; 5.4; 3.2; researcher reasoning"
    }
  ]
}