AI powered captioning

Closed captioning is typically a manual, error-prone process that costs you time, money, and resources. And mishaps can result in hefty FCC fines. Watson Captioning leverages AI to automate and simplify the process for live broadcasts and on-demand video.

Watson Captioning is designed to be

Automated. Accurate. Editable. Fast.

Watson Captioning Live

Watson Captioning Live

Highly accurate real-time captioning powered by Watson AI

Watson Captioning for On-demand

Watson Captioning for On-demand

Add captions to existing video libraries and add automated captioning to your publishing workflows.

  • Deliver highly accurate closed captions after training Watson which will continuously improve over time
  • Use existing Newsroom Control System scripts to pre-train Watson
  • Leverage features such as custom corpora and custom vocabulary to enhance accuracy of first-run caption scripts
  • Increase recognition accuracy through domain-specific language model customization
Highly accurate captions
  • Leverage Watson automated speech-recognition capabilities to ingest spoken and audial elements of video assets
  • Generate captions using a smart layout algorithm, which segments caption cues at natural breaking points, resulting in greater readability
  • Support captioning for on-demand video assets via an entirely web-based solution
  • Support captioning for live broadcasts via a hybrid solution that prioritizes speed
  • Reduce time to generate captions to a fraction of the length of the video
  • Use existing station hardware (for TV stations generating live captions)
Automated caption generation
  • Quickly review and correct low-confidence words using the caption editor
  • Optimized for efficiency with an intuitive UI for caption corrections
  • Edited words are automatically time-synchronized to make manual correction quick and painless
  • Designed for both experts and non-professionals

* Caption editor is available for on-demand only

Quality control with the caption editor
  • Apply a range of AI functionality to assess and act on the interpreted data
  • Leverage self-learning from every correction; recognition accuracy improves with every use
  • Extract names and proper nouns automatically from reviewed captions and set as glossary words for future use
  • Add additional “sounds like” terms to the glossary at any time
Self-learning

Watson Captioning in action

Watson Captioning in action

Watson works the U.S. Open

Watson transcribed over 660 U.S. Open match videos using the Watson Speech to Text API.

The cognitive era is here

Conquer the captioning challenge

Put Watson Captioning to work for your video business.