Closed Captioning Framework for Media
Media & EntertainmentBusiness Impacts
Reduced latency
83%
Transcription accuracy
Improved user experience
Customer Key Facts
- Location : North America
- Industry : TV Broadcasting
Problem Context
The process of overlaying closed captioning in News and Media has seen minimal advancements over the past decades, especially for live-streamed content. Live streaming still relies heavily on antiquated technologies, such as teleprompters, encoders, and other hardwired devices, which are prone to glitches and errors in transcriptions.
The customer, a holding company of eleven News broadcasting studios across the United States, wanted to have real-time transcriptions on their live broadcasted news feeds on the web and for pre-recorded video news segments.
Challenges
- Incomplete transcriptions on their web-based portal for recorded and live stream videos
- Random breakages and glitches in transcriptions during transmission
- Poor transcription accuracy numbers
Technologies Used
Google Cloud Platform
Kubernetes Engine
Cloud Storage
Cloud Functions
Cloud Speech-to-Text
Transforming Closed Captioning, One Live Stream at a Time
Solution
Quantiphi developed a scalable web-based user interface (UI) which consisted of closed captions obtained using Google Speech-to-Text API synced with the video feed. This enabled near real-time captioning and viewing of those captions on live news broadcasts. To transcribe recorded videos, Google Cloud Functions and Google Kubernetes Engine were used to extract the audio track from the videos and Cloud Storage was used to store these transcriptions.
Result
- Performance metrics for data-driven decision-making
- Improved data storage & data management
