Company contribution
Company contribution
Experts at Ai-Media reveal how the company is revolutionising broadcast captioning with secure , AI-powered technology
Broadcast captioning
Most broadcasters looking to utilise captioning have three main goals : 1 . They need to caption everincreasing amounts of content as they expand into over-the-top streaming . 2 . They are looking to reduce costs by shifting legacy workflows to the cloud and leveraging advancements in AI and ASR . 3 . They want solutions that provide world-class security and uptime , to ensure captions are always delivered to screens . As the global leader in captioning technology , Ai-Media ’ s end-to-end solutions allow broadcasters to tick all these boxes .
Introducing LEXI 3.0 – the future of automatic captioning Ai-Media is continually working to enhance its captioning technology with the latest advancements in AI . And at the 2023 NAB Show in Las Vegas , the company launched the latest breakthrough in AI-powered automatic captioning with LEXI 3.0 .
Trusted by most of the world ’ s leading broadcasters , this new and improved version of Ai-Media ’ s flagship live captioning solution leverages the power of AI to deliver results rivalling human captions , at a fraction of the cost .
12 EUROMEDIA
Independent audits have proven that LEXI 3.0 delivers an average accuracy of 98.7 % based on the globally recognised NER model . It also achieves up to 60 % fewer recognition errors , formatting , and punctuation errors than the previous version .
LEXI 3.0 also provides an enhanced viewer experience thanks to innovative new features , including automated speaker identification and AI-powered caption placement to avoid on-screen interference .
All this makes LEXI 3.0 the perfect solution for broadcasters looking to cost-effectively ensure compliance with accessibility requirements – whether to caption a linear TV broadcast or an OTT content platform with multiple channels .
LEXI 3.0 can also break down language barriers by delivering multilingual captions in over 30 languages .
Why is LEXI 3.0 so accurate ? One of the keys to LEXI 3.0 ’ s accuracy is its unique Topic Models – otherwise known as a custom dictionary . Powered by AI , this feature greatly improves automatic caption accuracy by allowing LEXI 3.0 to recognise unusual or domain-specific words , phrases , names , and context .
A Topic Model consists of text-training data such as common phrases expected to be heard during the audio programme , proper nouns which are desired to be recognised but would not be found in a standard dictionary , and pronunciation clues for any new words . When this data is entered into a Topic Model , LEXI 3.0 ’ s ASR engine learns to produce terms correctly when it ‘ hears ’ them , through the power of AI .
Broadcasters can easily add text-training data into a Topic Model , either manually or via LEXI 3.0 ’ s data scraping feature . This allows you to scan a website , for instance , and all custom terms and phrases are automatically uploaded into a Topic Model . Broadcasters can save even more time by requesting the assistance of Ai-Media ’ s team of captioning experts to manage custom dictionary curation as part of a managed service .
How the Portland Trail Blazers enhanced broadcast captioning with LEXI 3.0 LEXI 3.0 ’ s AI-driven capabilities have achieved particularly stellar results in the world of sports broadcasting , with worldfamous franchises like the Portland Trail Blazers making the most of the enhanced