- The “Own My Voice” project enables voice recipients to own unique and customized voices
Xiaomi, a consumer electronics and smart manufacturing company with smartphones and smart hardware connected by an IoT platform at its core, has announced its latest accessibility application of powerful algorithms and self-developed speech technology. The Xiaomi AI Lab’s spontaneous style Text-To-Speech technology is utilised to create a unique and personalised voice for a user with speech difficulties.
Instead of a monotonous electronic voice, this user may now communicate with others using “his own voice.” This successful attempt, which was part of the Xiaomi Technical Committee’s “Own My Voice” pre-research project, demonstrates Xiaomi’s dedication to “Tech for Good” and achieving its mission of “let everyone in the world enjoy a better life through innovative technology”.
Why did Xiaomi launch this project?
Xiaomi is concerned about people and strives to meet their various demands via technological innovation. It realised that many users with speech issues wanted to own their own voices for everyday communication, therefore it formed the “Own My Voice” project team to find a user with speech disorders to be the voice recipient. “We are excited to explore multiple values that technology innovation brings to us, such as responding to users’ demands for self-identity and the construction of identity,” said Zhu Xi, Xiaomi Corporation’s Technology Committee topic convener on Tech for Good.
How did Xiaomi carry out the project?
The project team recruited more than 200 volunteers from Xiaomi to offer their voices in order to create the most suitable and individualised voice for the recipient. They used a voiceprint matching algorithm to match the characteristics of donated voices with the characteristics of the recipient’s voice. They discovered the most suitable voice as the recipient’s basic sound of voice reference using this approach. The chosen real voice was altered with complex acoustic modification to form a new and original sound of voice in consideration of personalisation and privacy protection.
They then used spontaneous Text-To-Speech technology to train AI models, eventually giving this new voice a natural rhythm and intonation that can accurately reflect human emotion and tone.
The “Own My Voice” project combines a range of advanced algorithms with Xiaomi’s self-developed speech technology to assure the synthesised voice’s specificity, safety, and high genuineness, resulting in a new concept in personalised speech synthesis for users with speech disorders.
What is the significance of the project?
A group of Xiaomi AI Lab voice technology experts is at the heart of this project. Since 2017, they’ve published 37 articles on speech in top international conferences’ proceedings, such as the International Conference on Acoustics, Speech, and Signal Processing (ICASSP). The popularity of “Own My Voice” is mainly due to the spontaneous Text-To-Speech technology Xiaomi developed.
In terms of intonation, pause, speed, and other qualities, the spontaneous style Text-To-Speech technology essentially make the synthesised voice seem like a real human. This replaces the synthetic voice’s repetitive and unnatural sound with a more natural one. This technique is currently applicable to various smart devices that include Xiaomi’s AI voice assistant, Xiaoai. The “Own My Voice” project demonstrates how spontaneous Text-To-Speech technology may be used to improve user experience in accessibility areas.
Zhu Xi added, “If we notice and address the needs of minority groups at an early stage, the process of technology diffusion could be greatly shortened. This allows the benefits of new technologies to become accessible to users with special needs without delay.”
Xiaomi will continue to get feedback from the voice receiver and will explore the feasibility of this project on a wider scale. Xiaomi will continue to empower accessibility through cutting-edge technology, striving to meet the diverse needs of people through technological innovation.