Self-Authored Multilingual Subtitle Alignment Samples for AI Video Translation

This dataset contains self-authored multilingual subtitle alignment samples for evaluation of video subtitle translation workflows. The package includes 24 scripted clips, 72 timestamped source subtitle segments, and 144 aligned translation rows across English, Spanish, and Chinese (Simplified), along with a schema, data dictionary, sample SRT files, and methodology notes. All text was authored specifically for this release to avoid third-party rights issues. Related application context: https://aitranslatevideo.org/