In the LinkedIn live event, Gio’s speech was transcribed and translated in real-time, leveraged by the Speechmatics engine. This feature has gained popularity due to the rise in live events, benefiting those who do not understand the spoken language, as they can view translated text or hear it in another language. Although the voices are not the originals, these advancements bring us closer to enabling global comprehension of any presentation. 

Q: What's the real value add that Syncwords is providing on top of the Speechmatics engine? How would you define that? What's your unique selling point (USP)? 

A: Our mission at SyncWords is to leverage the accuracy of Speechmatics and focus on delivering live video accessibility and localization translations to every player, whether it's on Zoom, a live stream, or live television. We’re unique because we’re able to take the engine that Speechmatics has and deliver it so that the audience can see it on top of their live video. 

Q: Where do you see the major use cases and markets? Where have you seen success and where do you see growth? 

A: SyncWords initially found success in the live television industry, mainly live news and sports to do real-time transcriptions within their live programming. During the pandemic, there was a surge in demand from the corporate sector for virtual conferences and international business meetings.  

This highlighted the opportunity for growth in the future, as virtual conferences allow people from all over the world to attend and share information.  

Q: Have you seen particular geo markets that have grown in that area? Where are you focusing your efforts and seeing a pull for demand? 

A: SyncWords is seeing a lot of demand in Asia and Europe. Companies want to increase their revenue and reach and that means growing outside of the United States. In some cases, it is driven by legislation but also for smaller countries like, Romania, who to grow beyond their boundaries they must do translation for their product videos and live streams.

Q: What are the cost implications of not using translation and transcription services like SyncWords? What drives businesses to invest in these services? 

Talking about automation and a live engine, there's two ways to look at it. If you don't do it, you're missing an opportunity. Where someone doesn’t understand your messaging or speaks another language, they’re going to lose interest or go with another vendor that speaks their language. 

When it comes to live translations, the usual approach is to hire interpreters. However, this may only sometimes be feasible, especially when specific languages are involved or when there's limited time. Additionally, there are logistical considerations and costs associated with hiring interpreters, including infrastructure setup and ensuring reliable internet connection and audio quality. The expense becomes significant when multiple languages are involved. This is where our platform stands out by automating the translation process, reducing costs, and providing a deliverable solution. 

Q: From a technology standpoint, how does SyncWords ensure effective dubbing? What role does punctuation play in the process? 

A: SyncWords combines Speechmatics with a dictionary for accuracy, which is crucial for effective dubbing. Speechmatics gives us the ability to add a dictionary, which has accurate punctuation where other engines do not. Punctuation is essential in translations as it determines the meaning of sentences. 

Proper punctuation allows translation automation to be more accurate and ensures the correct interpretation of the spoken word. SyncWords' platform provides options for choosing different accents, catering to the preferences of specific audiences, such as Spanish-speaking countries where different accents are prevalent, making the content more friendly and welcoming.  

Q: What are the challenges that might arise when translating content for different audiences? How does SyncWords overcome them? 

A: Audio quality plays a crucial role in speech-to-text technology. Live news and sports utilize professional equipment and techniques to ensure high-quality audio. Virtual live streams pose challenges, but standardizing AV practices and microphone setups can help. In-person events also require standardized AV practices to ensure accurate audio deliverables.

Synchronization of the audio and subtitles presents other challenges although can be adjusted on the platform, by introducing delays or using latency-hiding techniques. Language switches and cultural language structures present ongoing challenges that require research and development from Speechmatics to overcome. 

Q: Building foreign vocabulary into language packs is common for Speechmatics. Non-English packs often include English elements, which affects sentence structure and translation. 

What level of transcription accuracy is needed for effective translation?   

A: SyncWords prioritizes high-accuracy transcription for effective translation. Benchmarks demonstrate faster translation compared to human translators. Optimizing variables like good audio and utilizing Speechmatics with an added dictionary for preconfigured names, brands, and phrases ensures high accuracy. As soon as that sentence is complete, the system starts to trigger the translated phrase. Then moves on to the next phrase. Whereas a human interpreter must paraphrase and keep up.  

When doing live streaming events it’s critical to capture as much of the spoken word. Using Speechmatics increased accuracy we can deploy faster translation and cover more of the message.  

Q: What does the future hold for transcription/translation/dubbing?  

A: The future vision is to make real-time transcription and translation accessible to all. SyncWords aims to provide developers with an API that seamlessly integrates Speechmatics' translation workflow into various platforms, including Zoom, phones, and live TV channels. 

Collaboration with major vendors in live streaming and content delivery networks aims to standardize this workflow. The ultimate goal is for real-time transcription and translation to become the go-to solution whenever there is a need for it. 

Q: How many languages can be supported at once for live voice translation? 

A: Syncwords relies on Speechmatics for the languages available. Speechmatics supports translation into over 40 languages, including major languages, Eastern European languages, and several Asian languages, with the potential for further expansion. 

Q: How do you foresee the Language Service Providers (LSP’s) of the world accepting or rejecting this technology? 

A: I believe they're really going to embrace this technology. Language service providers, do not like turning down business, oftentimes they do have to due to availability constraints. They want a simple way to close business and a way to scale up offerings but also to always have a plan B.  

This article is an edited and summarized version of a LinkedIn Live conversation.