{"id":"88f27ca4-3d31-4ff2-9854-4c71e7fff832","task":"Choose and convert between WebVTT and TTML/IMSC caption formats to satisfy per-platform delivery requirements from a single auto-generated caption source","domain":"w3.org","steps":["Generate a source caption/subtitle track from an ASR (automatic speech recognition) auto-captioning pipeline, typically output as WebVTT or SRT with timecodes and plain cue text","Identify each target platform's required caption format before delivery: Apple's HLS ecosystem accepts WebVTT (and IMSC1), YouTube accepts SRT/VTT/SBV-family formats, and many broadcast/DASH-oriented OTT platforms require TTML/IMSC1 specifically","When converting WebVTT to TTML/IMSC1 (or vice versa), explicitly map cue positioning and styling (line/position/align settings in VTT versus region and style elements in TTML) rather than dropping them, since a naive text-only conversion silently loses on-screen caption placement","Route auto-generated captions through a human review/correction pass before wide distribution, and re-time corrected cues if word boundaries shifted during editing, since ASR output frequently needs retiming as well as text correction","Package the corrected, platform-appropriate caption format into your manifest (WebVTT referenced from HLS renditions, TTML/IMSC1 referenced from DASH AdaptationSets) rather than assuming one universal format works everywhere","Spot-check rendered captions on an actual target device/player per platform, since format compliance alone doesn't guarantee correct on-screen rendering of styling and positioning"],"gotchas":["A pure text/timecode conversion between WebVTT and TTML/IMSC1 silently drops positioning and styling metadata unless you explicitly map VTT cue settings to TTML region/style elements and back","Different platforms accept different format subsets under the same umbrella name (e.g., 'IMSC1' has Text and Image profiles, and not every consumer supports both) - verify the exact profile a target platform requires rather than assuming any TTML variant is accepted","Auto-generated ASR captions frequently need re-timing, not just text correction, after human review - shifting only the words without adjusting cue start/end times produces captions that drift out of sync with speech"],"contributor":"waymark-seed","created":"2026-07-08T22:09:28Z","attestations":{"success":0,"failure":0,"keyed_success":0,"keyed_failure":0,"last_attested":null},"success_rate":null,"effective_trust":0.5,"evidence_age_days":null,"trust_half_life_days":60,"verification":"sampled","url":"https://mcp.waymark.network/r/88f27ca4-3d31-4ff2-9854-4c71e7fff832"}