Datasets
Video
Broadcast News
On-air and field output produced on a daily news cycle, including anchor reads, reported packages, live field shots, interviews, and raw B-roll. Polished studio delivery sits alongside unscripted field capture from the same news days, in both finished and raw form.
Hours408K+
Segments8+
Topics10+
Training Use Cases
✓Talking-head and reporter-style video generation
✓Avatar and digital human generation for news anchors
✓Video summarization and news segment retrieval
✓Speaker identification and diarization
Key Highlights
✓8+ segment types including anchor reads, packages, live shots, interviews, pressers, and B-roll
✓10+ news topic areas covering politics, business, weather, sports, technology, health, and culture
✓Multilingual coverage from local-market through national and international broadcast outlets
✓Mix of finished on-air segments and unedited raw field material from the same news cycles
Metadata Fields
durationLength of clip in HH:MM:SS
resolutionPixel dimensions (e.g., 1920x1080, 3840x2160)
frame_rateFrames per second (e.g., 24, 30, 60, 120)
contains_audioWhether the clip carries an audio track (boolean)
primary_categoryDominant content category assigned to a video
stylestudio | field | mixed | archival
segment_typeanchor_read | reported_package | live_shot | interview | presser | breaking_cutin | weather | b_roll
topic_areapolitics | business | weather | sports | technology | health | culture | human_interest
languagePrimary spoken language (ISO 639-1 code, e.g., en, es, zh)
outlet_scalelocal | regional | national | international