An AI-driven podcast production platform
,allowExpansion)
THE CHALLENGE
Make the process of creating high-quality voice content easier and faster.
scenario
Renewing the audio publishing landscape with AI
GEDI, one of the most important publishing groups in Italy, is constantly looking for new models to involve its audience and offer personalized experiences for the use of its editorial content.
With the increasing distribution of podcasts as an engaging and accessible communication tool, GEDI wanted to experiment with speech synthesis powered by Artificial Intelligence in the process of creating audio articles so as to reduce production times and respond to the growing demand for quality vocal content.
)
THE SOLUTION
Generate podcasts with the support of AI: Voice Studio
The Gedi Group has thus chosen Voice Studio, Discovery Reply's solution that uses artificial intelligence to transform textual articles published on web newspapers into high-quality audio content. Using one of the best voice cloning engines in AI, Voice Studio allows you to automatically generate audio with synthetic neural voices or real sampled voices, ensuring the realism and naturalness necessary to create engaging narratives.
The podcasts generated with Voice Studio are already active for La Repubblica and La Stampa, two of GEDI's main newspapers. Thanks to Voice Studio, GEDI has improved its audio production capacity, creating up to 300 podcasts per day. This has greatly simplified the process, making it more in line with editorial needs. In addition, the possibility of using different voices and speech styles adapted to the contents enriches the listening experience, increasing user engagement.
HOW DID WE DO IT
Innovating podcasting with efficiency and automation
The project began with the creation of two synthetic voices, sampled from real voices, able to create a unique bond with the Gedi group. We then moved on to the automation of the audio production process, which, thanks to different text-to-speech engines integrated into Voice Studio, can quickly transform textual content into fluid and authentic speech.
Finally, some customized features have been activated for the GEDI editorial context, such as the adaptation of unusual foreign words, the contextual expansion of acronyms to improve the listening experience, the setting of intonation, rhythm and emphasis in accordance with the type of article (for example news, sport, culture,..)
Removing hallucinations
The system makes it possible to correct the 'hallucinations' of AI, with an editorial intervention on the pronunciation of the words.
Adaptation memory
The solution 'learns' from the corrections made by the publishers, adapting the pronunciation of subsequent generations.
Automatic audio enhancement
The musical background is automatically integrated and the final audio undergoes a quality control to filter out any distortions or anomalies.
With 5 million users who read and listen to us every day, GEDI is one of the most important publishing groups in Europe. Through the newspapers La Repubblica, La Stampa, and the various online portals, GEDI produces information and multimedia content that every day describe with professionalism, honesty and commitment how the world changes. The entertainment produced by the Group's radio stations (Radio Deejay, Radio Capital and M2o) and by One Podcast focuses on quality and originality to give listeners a wide range of offerings in step with new listening habits.
With deep experience in Cloud Computing, Digital Media and System Integration, Discovery Reply specializes in the development of advanced end-to-end Digital eXperience Management (DXM) solutions for the management and distribution of brand multimedia content and digital assets. A complete suite of enterprise-level vertical technologies and applications aimed at the digital media market that drive new business models and omnichannel strategies.