RTL>
13 April 2022 – 12:08 look
Sometimes you see information or a message on the Internet and you do not have the possibility to read the text. It would be great if someone was there to just read the text. With us, for example, you can engage the RTL presenter Maik Meuser. With your voice and a good dose of technical innovation, you can make more and more text content sound.
We develop so-called “text-to-speech” offers
RTL presenter Maik Meuser and podcast host Inken Wriedt
Speech synthesis enables a variety of new customizable media offerings. RTL Germany is currently leading the development of so-called “text-to-speech” offerings, which can convert text into realistic speech using computationally intensive artificial intelligence (AI). The project is part of a global partnership for technological innovations in the fields of media and education between Bertelsmann and the technology company Microsoft and was created in cooperation with the software agency Appsfactory. The project is funded by the Journalism Laboratory of the National Media Authority of North Rhine-Westphalia.
Maik Meuser and Inken Wriedt lend their voices
For a showcase, RTL Deutschland recorded the voices of RTL presenter Maik Meuser and podcast host Inken Wriedt and synthesized them via an artificial neural network, which now enables RTL’s new text-to-speech applications. The project created realistic, natural versions of the voices from four hours of spoken text, which are indistinguishable from the human original. All voices familiar to users can be used in many areas.
TIP: Here you can directly try the new voice function. Just make you read the article.
Possible applications are, for example, customizable audio news information or the read-aloud function now available for the text-based online news offerings of RTL and ntv. It can be called up via smart speakers, in-car entertainment systems or RTL Germany’s online services. AI can also provide support in the field of digital education, for example with functional texts in training videos. Here, RTL Germany is in close contact with Relias, a company of the Bertelsmann Education Group, which is one of the world’s leading providers of digital education specifically for the healthcare sector.
We are making progress in text-to-speech translation
Isabella Thissen, Senior Vice President Editorial Products & Innovation at RTL Germany
Isabella Thissen, Senior Vice President Editorial Products and Innovation at RTL Deutschland: “The translation of text into realistic speech is currently making huge progress through the use of artificial intelligence and machine learning. This opens up new possibilities for user-centric and trustworthy offerings, especially in the information sector.We are currently gaining valuable experience with our innovation case and actively participating in shaping this development. “The future of AI voices, as we are currently developing them, will result in products that give our users a new experience through voices.”
Improve the accessibility of media offers
Holger Meinzer is director of industry relations, media and telecommunications at Microsoft Germany
“We are very pleased with this cooperation, as innovative media companies like RTL are now discovering the great opportunities of artificial intelligence technologies,” says Holger Meinzer, director of industry relations, media and telecommunications at Microsoft Germany. “Artificial intelligence can, for example, convert content automatically and almost in real time to other formats, improve the accessibility of media offers and allow a user-centric and personalized approach to the audience. It thus becomes a factor decisive factor in the competitiveness of media companies thanks to attractive exclusive platforms and innovative offers The responsible use of these technologies is at least as important.”
Strict ethical standards apply
The pilot project follows strict ethical guidelines set out in Microsoft’s guidelines for responsible use of artificial intelligence. This way, the synthetic voices are clearly marked for future use by users. There are also clear guidelines in terms of content. For example, any content that could manipulatively influence the formation of opinions is excluded from synthesis using AI. The project team is in contact with the Medienanstalt NRW on these issues.