TVEyes, the industry-leading provider of global audio and video intelligence, has announced a major upgrade to its Speech-to-Text (STT) technology, resulting in significant accuracy improvements against key global languages and surpassing industry leaders in performance.
Our dedicated team of developers, linguists and technologists at the TVEyes Language Technology Labs (TLT Labs) released these STT upgrades, integrating models generated from different training frameworks and third-party components. This leap forward represents two generations of technological advancement, and the results are impressive.
An analysis of four key global languages reveal significant accuracy improvements, such as a 13.13% overall accuracy boost for U.S. English and a 10.22% improvement for German.
TVEyes transcribes more than 2 million hours of TV, radio, podcast and online video content monthly. In our labs, we continually develop, test and train language models to transcribe audio and video streams, producing transcripts in multiple languages for end users. Our commitment to these models enables us to rapidly launch media monitoring and data solutions in new countries and languages, meeting evolving customer demands.
To date, we have developed nearly 40 different language models – from English to French and Spanish to Mandarin, Standard Arabic, Japanese, with many more in development.
SO, HOW DOES TVEYES STT COMPARE?
The team analyzed TV and Radio transcripts from selected television and radio channels for each of these languages. They selected a 4-minute clip per active TV and radio channels on different dates and at different times.
In each analysis, TLT Labs consistently met or exceeded the performance of industry leaders such as Whisper, Amazon and Google. In a review across four key language models, including U.S. English, Australian English, Swedish and German, TVEyes averaged an overall accuracy score that surpassed Amazon, Whisper and Google STT.
Pictured below: Australian English – 8.75% accuracy improvement from TVEyes Legacy STT to TVEyes Next Gen STT
Benefits:
Unmatched Accuracy and Keyword Recognition: Our upgraded STT technology significantly enhances overall accuracy, ensuring precise keyword recognition. This results in cleaner, more reliable outcomes for your media monitoring needs.
Reduce AI Hallucination: With customers using advanced algorithms and refined processing, TVEyes’ STT upgrades help minimize AI hallucinations, delivering more trustworthy and consistent transcriptions.
Comprehensive and Detailed Reports: The improvements in our STT technology lead to more complete and insightful reports, empowering users with the most accurate and actionable broadcast and podcast intelligence available.
TVEyes STT upgrades will be integrated into our entire data solutions product suite, encompassing Broadcast, Podcasts and Online Video Data Solutions. To learn more please contact us at partnerships@tveyes.com.
About TVEyes:
TVEyes powers real-time decision making at some of the world’s leading organizations. If it involves audio or video content, data or intelligence, chances are we are behind it. We offer powerful tools and applications that turn audio and video content into actionable intelligence. With global coverage and rich data assets, we leverage our expertise and technology to deliver valuable solutions for data-driven decision makers around the world.
