DeepL Expands Voice Translation Suite

DeepL just launched Voice-to-Voice. Real-time speech translation is now enterprise software. 𝐃𝐞𝐞𝐩𝐋 — 𝐥𝐨𝐧𝐠 𝐤𝐧𝐨𝐰𝐧 𝐚𝐬 𝐭𝐡𝐞 𝐩𝐫𝐞𝐦𝐢𝐮𝐦 𝐌𝐓 𝐞𝐧𝐠𝐢𝐧𝐞 — has just announced DeepL Voice-to-Voice: real-time spoken translation that integrates with 𝐌𝐢𝐜𝐫𝐨𝐬𝐨𝐟𝐭 𝐓𝐞𝐚𝐦𝐬 𝐚𝐧𝐝 𝐙𝐨𝐨𝐦. 𝐓𝐡𝐢𝐬 𝐢𝐬 𝐧𝐨𝐭 𝐚 𝐧𝐢𝐜𝐡𝐞 𝐩𝐫𝐨𝐝𝐮𝐜𝐭. 𝐓𝐡𝐢𝐬 𝐢𝐬 𝐞𝐧𝐭𝐞𝐫𝐩𝐫𝐢𝐬𝐞 𝐢𝐧𝐟𝐫𝐚𝐬𝐭𝐫𝐮𝐜𝐭𝐮𝐫𝐞. 𝐖𝐡𝐚𝐭 𝐝𝐨𝐞𝐬 𝐭𝐡𝐢𝐬 𝐦𝐞𝐚𝐧 𝐟𝐨𝐫 𝐢𝐧𝐭𝐞𝐫𝐩𝐫𝐞𝐭𝐞𝐫𝐬 𝐚𝐧𝐝 𝐥𝐨𝐜𝐚𝐥𝐢𝐳𝐚𝐭𝐢𝐨𝐧 𝐩𝐫𝐨𝐟𝐞𝐬𝐬𝐢𝐨𝐧𝐚𝐥𝐬? → Real-time interpretation demand will shift. Live events are affected. → The human premium moves to high-stakes, high-consequence contexts. → New skills: voice QA, AI speech review, prompt calibration. 𝐓𝐡𝐞 𝐭𝐨𝐨𝐥𝐬 𝐚𝐫𝐞 𝐞𝐯𝐨𝐥𝐯𝐢𝐧𝐠 𝐟𝐚𝐬𝐭𝐞𝐫 𝐭𝐡𝐚𝐧 𝐦𝐨𝐬𝐭 𝐩𝐫𝐨𝐟𝐞𝐬𝐬𝐢𝐨𝐧𝐚𝐥𝐬 𝐫𝐞𝐚𝐥𝐢𝐬𝐞. Are you watching? Follow CITLoB for weekly localization industry updates. 👥  #DeepL #SpeechTranslation #Interpreting #Localization #AITools #CITLoB Biraj Rath Rushikesh Atre Sandeep Nulkar Sunil Kulkarni Sudheen M Tanushree Choudhury Prithvi Jain Sanjeev Das Sankeshwari Deo Sanket Joshi Shikhar Saran Srivastava Shilpa Saiya Soham Kakade

Can Earbuds Break the Language Barrier? “AI wants to help you in a foreign land.” ——- The latest wave of “smart translator” earbuds—like those from Smartech Gadgets—promise something that feels almost science fiction: real-time conversation across languages,… Read more: https://lnkd.in/gs42Fx4D #travel #tech #travelgram #translation 

Another update to our new live AI captions and translation system. Alongside real-time captions and multilingual translation for live events, we now also support session transcript recording with downloadable TXT and SRT outputs. So in addition to live on-screen captions, relay screens, and translated output, clients can now retain: • Full plain text transcript (.TXT) • Timecoded subtitle file (.SRT) for recorded content and post-production workflows Useful for accessibility deliverables, post-event editing, content repurposing, compliance records, and archived session transcripts. Built with real live event workflows in mind. #LiveCaptions #LiveTranslation #EventTech #HybridEvents #VirtualEvents #Accessibility #EventProduction #ConferenceTech #Webcasting #LiveStreaming #Subtitling #AITranslation

Another update to our new live AI captions and translation system. Alongside real-time captions and multilingual translation for live events, we now also support session transcript recording with downloadable TXT and SRT outputs. So in addition to live on-screen captions, relay screens, and translated output, clients can now retain: • Full plain text transcript (.TXT) • Timecoded subtitle file (.SRT) for recorded content and post-production workflows Useful for accessibility deliverables, post-event editing, content repurposing, compliance records, and archived session transcripts. Built with real live event workflows in mind. #LiveCaptions #LiveTranslation #EventTech #HybridEvents #VirtualEvents #Accessibility #EventProduction #ConferenceTech #Webcasting #LiveStreaming #Subtitling #AITranslation

Can Earbuds Break the Language Barrier? “AI wants to help you in a foreign land.” ——- The latest wave of “smart translator” earbuds—like those from Smartech Gadgets—promise something that feels almost science fiction: real-time conversation across languages,… Read more: https://lnkd.in/gRea5aZg #travel #tech #travelgram #translation 

One of the things that really stuck with me from ISE 2025 was sitting at the Televic Conference stand while Didier Rosez demonstrated the tabletop UniCOS demo system run live AI transcription and translation through the conference workflow. It wasn't golden firmware — it was the actual delegate experience, and the latency and accuracy was good that we could speak in Afrikaans and immediately see our words transcribed and translated into English and French in real time! More recently Televic has been showing the same kind of workflow flowing through Televic Unite as the delivery layer, which opens it up to rooms that don't have dedicated delegate units at every seat. This is going to fundamentally change how mission-critical meetings get run. Board votes. Council sessions. Tribunals. Disciplinary hearings. International negotiations. The kind of room where a misheard word in the minutes can become a legal problem six months later. And this is where it stops being a software story and starts being an audio story: AI transcription and translation are only as good as the audio you feed them. A generic boundary mic picking up HVAC rumble, two delegates talking over each other, mobile phone RF buzzing into the cable, the room's reflections smearing the consonants — you can throw the best AI model on earth at that signal and it will still hallucinate names, drop negatives, and confidently translate "we cannot approve" as "we can now approve." In a casual meeting that's a chuckle. In a mission-critical one it's a lawsuit. This is exactly the problem Televic has been solving for decades, long before AI was in the conversation. One delegate, one shielded microphone, immune to mobile phone interference, with a frequency response engineered specifically around speech intelligibility, gated and routed cleanly through the Televix Plixus bus. The signal that reaches the AI engine is already clean, already attributed to the right speaker, already isolated from the next person at the table. That's not something you can retrofit onto a generic ceiling-array room — it has to be designed in from the microphone up. Which is why I think the Televic + AI translation pairing is going to age really well. Everyone is going to spend the next two years discovering that their AI captions only work when one person speaks slowly into a perfect mic in a treated room. Televic users wont need to worry about this. We're bringing this end-to-end at Mediatech Africa | Pro AV, UC, Broadcast & Event Tech Expo in June. Come and watch it run live, with real overlapping speakers, real background noise, and see for yourself what the difference between getting AI translation and trusting AI translation actually looks like. More to follow closer to the show. 👀 https://lnkd.in/dcxNuqKe Register for Mediatech here using the free access code "TELEVIC": https://lnkd.in/dYYminB9 #Mediatech2026 #ProAV #Televic #AIinAV

Peter Argondizzo calls out the "transparency problem" he sees in the language services industry. Peter, who founded Argo Translation after working as a translation buyer, has watched the problem reshape itself over decades. The old version sounded like agency owners claiming their translators were so good they didn't need proofreaders. The new version is harder for buyers to catch. Some providers build elaborate AI systems and claim translators handle the work, but the content never actually sees a human. The fix is the same in either era. Buyers should ask how their translation was produced, and a trustworthy provider will give a clear answer.

OpenAI releases three new audio models for the API: GPT-Realtime-2 for real-time conversations, Translate for translations, and Whisper for live transcription.

OpenAI releases three new audio models for the API: GPT-Realtime-2 for real-time conversations, Translate for translations, and Whisper for live transcription.

Published a research preprint titled: “GloChat: A Real-Time Multilingual Chat System with AI-Powered Translation” The work explores multilingual conversational systems and real-time AI-powered contextual translation. Access the paper here: https://lnkd.in/gWiptmuX