Speechmatics Helps Broadcasters Get Names Right

Language is nowhere near as static as you’d think. Speechmatics’ Custom Dictionary is smart enough to recognize context and pairs of words to ensure it gets the spelling right.

Speechmaticshas launched Custom Dictionary (CD), a new feature for its automatic speech recognition (ASR) speech-to-text transcription product. Live broadcast subtitling has a recognised problem when it comes to both people's names and locations. Custom Dictionary (CD) allows users to add context- specific words such as player names or breaking news locations in real-time to increase their accuracy.

Traditionally, transcription providers have to train each specific model for a customer to include special words. However, setting up a new model is time-consuming, usually requiring an overnight rebuild. Custom Dictionary streamlines the entire process by putting the customer in control and allowing them to add new words in plain text to “teach” the model, regardless of use case or language context.

Speechmatics' CD enables users to define the context of a conversation in advance, and input those difficult terms into the dictionary, so that when they do come up they are transcribed without error. For example, sports subtitling is dramatically improved with prior input of player names, venues and club names.

While developing CD, accuracy remained a top priority, as customers indicated this to be key to improving their value propositions. Speechmatics has successfully improved text accuracy rates and already comprehensive dictionary by dynamically adapting to the context of individual use cases. When compared directly to Google’s Cloud Speech, and with both Speechmatics’ CD and Google’s API “hints” features enabled, Speechmatics produced a transcript that was >12% more accurate.

Benedikt von Thüngen, Speechmatics CEO explained, “Accuracy is still the key metric for speech recognition systems and it has remained top of mind throughout the development of our solutions at Speechmatics. Understanding what has been said is key and everyone speaks using different acronyms and terminology that is bespoke to them. With CD we now offer the first step towards personalized transcription.”

Ian Firth, VP of Products at Speechmatics said, “In case of the broadcast industry, closed captioning is an area that desperately needs improvement. From the outdated re-speaking approach to very awkward blunders, the lack of innovation can cause a lot of errors. Custom Dictionary simplifies and improves closed captioning for broadcast."

You might also like...

FCC Expands Caption Mandate As Automated Processing Takes Center Stage

On October 27, 2020 The Federal Communications Commission issued an order to expand its captioning mandate for broadcasters to include audio description requirements for 40 designated market areas (DMAs) over the next four years. The move came after the Twenty-First Century Communications and…

The Resurrection Of Live Linear TV And How To Playout From The Cloud

One of the surprises from the latest research published by Nielsen was the significant rise in audiences watching live linear TV. Lockdown has not only sent SVOD viewing soaring through the roof but linear TV is expanding rapidly. One reason…

At Facebook, AI Tackles Automated Captioning For Online Video

Online video captioning is critical for the deaf community at any time, but during a public health emergency like COVID-19, it has taken on a new significance, particularly as people stay at home.

Articles You May Have Missed – August 15, 2018

The standards for moving video over IP are all decided, right? Not yet. Even so, the innovation presents unprecedented opportunities and empowers broadcasters to deliver flexibility, scalability, and more efficient workflows. Consultant and The Broadcast Bridge technology editor, Tony Orme,…

Automating Titling and Graphics Creation for Multiplatform Distribution

Aesthetically pleasing 3D titles and graphics are integral to providing the wow factor that keeps today’s broadcast viewers glued to the screen. These visual elements—including 3D and 2D titles, animated graphics and real-time data-driven overlays—provide the vital conte…