Self-Authored Multilingual Subtitle Alignment Samples for AI Video Translation

AI Translate Video Project authored 24 scripted video clips with 72 timestamped source subtitle segments and 144 aligned translation rows across English, Spanish, and Chinese. The dataset includes a schema, data dictionary, sample SRT files, and methodology notes. It was released on Harvard Dataverse in May 2026.

Use Cases

• Benchmarking subtitle translation quality based on the 144 aligned translation rows.
• Evaluating temporal alignment algorithms using the 72 timestamped source subtitle segments.
• Training or fine-tuning multilingual NLP models for video based on the scripted clips in three languages.
• Developing standardized evaluation protocols for video translation workflows based on the included methodology notes.

Strengths

• Includes 24 scripted video clips created specifically for this release.
• Contains 72 timestamped source subtitle segments and 144 aligned translation rows.
• All text is self-authored to avoid third-party rights issues.

Limitations

• Row count and file size are unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.

Provenance

Source

AI Translate Video Project

Collection Method

Self-authored scripted clips and subtitles created specifically for this release.

Freshness

Last updated 2026-05-11 15:01:23; freshness should be verified.