The NewConnect listed company, EKIPA Holding S.A, has announced that it is beginning the implementation of a project, “Development and Implementation of New Voice Synthesis Technology in a Foreign Language Using Artificial Intelligence.” The project results from a co-financing agreement signed by Ekipa Management and the Malopolska Centre for Entrepreneurship.
The project will be executed with the resources from Ekipa Management as well as European Union funds. The total value of the project is PLN 8.026.727,62, with eligible expenses determined at PLN 7.842.227,62, The total value of the co-financing will not exceed PLN 3.921.113,71.
The project that will be implemented by Ekipa Management involves developing and implementing new voice synthesis technology in a foreign language using artificial intelligence. The aim of the project is to automate the generation of dubbing for video content, create an advanced speech synthesis model capable of reflecting individual voice characteristics such as prosody, timbre, emotions, and other speech aspects. The main challenge is to develop an effective Neural Text To Speech architecture, focus on sounds specific to a foreign language and identification and use of available audio-video data sources. The project anticipates experiments with various voice synthesis methods, selection of the network model structure, and strives to minimize acoustic distance, aiming for speech synthesis quality close to a native English language user.
“We are pleased to receive the co-financing as it implements the development strategy of the entire holding. We see how AI is changing the world, and we want to use the referred technology in our activity. The created model will help us in the implementation of international projects, where there will be no language barrier,” indicated Łukasz Wojtyca, President of EKIPA Holding S.A.’s Management Board.
The project involves the development and implementation of new voice synthesis technology in a foreign language using artificial intelligence. It aims for automated dubbing generation for video content, creating an advanced speech synthesis model capable of reflecting individual voice traits such as prosody, timbre, emotions, and other speech aspects. An efficient Neural Text To Speech architecture, focusing on sounds specific to a foreign language, and identifying and utilizing available audio-video data sources constitute the key challenge. The project includes tests with various voice synthesis methods, seeking a network model structure and striving to minimize acoustic distance, aiming for speech synthesis quality close to that of a native English language user.
As a result, it will reduce the time needed for dubbing preparation, decrease movie production costs, and most importantly, it will be innovation in dubbing services through advanced NTTS technologies, ensuring the automation of this process while maintaining the faithfulness of the original voice traits representation.
This project responds to market needs such as multilingualism, cost-effectiveness, and high-quality dubbing, while simultaneously satisfying the global community’s needs for access to multimedia content in different languages. The proposed technology also aims to reduce the natural resources usage by shortening the time needed for dubbing and minimizing waste generated by traditional dubbing methods.
The project implementation is scheduled to conclude on March 31, 2027, or on the actual date of the project’s physical and financial completion, whichever is earlier.
Source: https://managerplus.pl/ekipa-otrzyma-blisko-4-mln-zl-dofinansowania-na-technologie-syntezy-glosu-z-wykorzystaniem-ai-98706
