DeepL, renowned for its text translation tools, has announced the launch of a voice-to-voice translation suite aimed at enhancing communication across various settings, such as meetings, mobile or web interactions, and group discussions for frontline workers via tailored applications. Additionally, the company is introducing an API that enables developers and businesses to leverage DeepL’s technology for specific applications, including call centre operations.
In an interview with TechCrunch, DeepL’s CEO, Jarek Kutylowski, highlighted the company’s evolution from text to voice translation, expressing confidence that their deep expertise in text translation has positioned them well for this next step. He pointed out that while advancements have been made in text and document translation, a robust real-time voice translation solution was notably missing from the market.
Kutylowski acknowledged the complexities this new product entails, particularly in minimizing latency – the time lag during speech and playback – while ensuring translation accuracy. DeepL is also rolling out add-ons for popular platforms like Zoom and Microsoft Teams, allowing users to receive real-time translations when others speak in their native languages, or follow along with translated text on their screens. Currently in early access, organisations are encouraged to join a waitlist for the program. The platform also facilitates in-person or remote conversations among multiple participants, with the ability to join discussions through QR codes.
A unique feature of DeepL’s voice technology is its capacity to adapt to specific vocabularies, accommodating industry-specific jargon and personal or company names. Kutylowski foresees that advancements in AI will redefine customer service paradigms, suggesting that translation tools will play a vital role in enabling companies to provide support in multiple languages, particularly where hiring qualified personnel is both challenging and costly.
DeepL maintains control over its voice-to-voice translation stack. Currently, its approach involves converting speech into text, translating it, and then transforming it back into speech. However, DeepL aims to advance towards an end-to-end model, potentially bypassing the text conversion phase entirely.
This new venture arrives amidst growing competition from well-funded rivals. Companies like Sanas, which recently secured $65 million in funding, are using artificial intelligence to modify accents in real-time, specifically targeting call centre applications. Similarly, Camb.AI, based in Dubai, is engaged in speech synthesis and translation for entertainment and media sectors. Palabra, supported by Alexis Ohanian’s firm Seven Seven Six, is also developing a real-time translation engine that aims to maintain both meaning and the speaker’s voice, drawing direct competition with DeepL’s initiatives.
In summary, DeepL’s new voice translation suite represents a significant expansion of its services, with promising potential applications in diverse communication scenarios, catering to the evolving demands of global interactions.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
