Xiaomi has announced its newest accessibility application, which uses powerful algorithms and self-developed voice technology. The Xiaomi AI Lab’s spontaneous style Text-To-Speech technology is utilized to create a unique and personalised voice for a user with speech disorders.
Instead of an usual repetitive artificial speech, this user may now converse with others using “his own voice.” This successful attempt, as part of the “Own My Voice” pre-research initiative conducted by Xiaomi Technical Committee, proves Xiaomi’s dedication to “Tech for Good” and achieving its purpose of “letting everyone in the world live a better life via creative technology.”
Why did Xiaomi initiate this project?
Xiaomi cares about people and strives to meet their various demands via technological innovation. It realized many users with speech difficulties wanted to own their distinctive voices for everyday communication and formed the “Own My Voice” project team to ask a user with speech issues to be the voice receiver. “We are thrilled to investigate numerous values that technology innovation provides to us, such as responding to users’ desires for self-identification and identity creation,” said Zhu Xi, Technology Committee topic convener on Tech for Good, Xiaomi Corporation.
How did Xiaomi design the task?
The project team enlisted over 200 volunteers from Xiaomi to lend their voices in order to create the most appropriate and individualized voice for the recipient. The voiceprint matching algorithm was used to match the attributes of the contributed voices of volunteers with those of the recipient’s voice. They discovered the best voice to serve as the recipient’s primary voice reference using this method. In order to safeguard individuality and anonymity, the chosen genuine voice was altered using complicated acoustic alteration to create a new and unique vocal sound.
They then employed spontaneous style Text-To-Speech technology to train the AI model, gradually gaining a natural rhythm and intonation that accurately reflects human emotion and tone.
The “Own My Voice” project combines a range of cutting-edge algorithms with Xiaomi’s in-house speech technology to assure the synthesized voice’s specificity, safety, and high authenticity, resulting in a novel approach to tailored speech synthesis for users with speech impairments.
What is the project's significance?
This initiative is supported by a team of voice technology professionals from Xiaomi AI Lab. They have published 37 articles on speech in the proceedings of premier international conferences such as the International Conference on Acoustics, Speech, and Signal Processing since 2017. (ICASSP). The popularity of “Own My Voice” is mostly due to the spontaneous style Text-To-Speech technology that they created.
The spontaneous style Text-To-Speech technology makes the synthesized voice seem like a genuine human in terms of intonation, pause, pace, and other characteristics. This replaces the synthetic voice’s repetitive and unnatural vibe with a more natural one. This technology is currently available on several smart gadgets that include Xiaomi, Xiaomi’s AI voice assistant. The “Own My Voice” project demonstrates how naturalistic Text-To-Speech technology may be extensively used in accessibility areas to improve user experience.
Zhu Xi said, “If we can identify and fulfill the requirements of minority groups early on, the process of technological diffusion might be substantially accelerated. This enables the advantages of new technology to be made available to individuals with special needs as soon as possible.”
Moving ahead, Xiaomi will continue to get input from the voice recipient and will investigate the viability of this initiative in a broader context. Xiaomi will continue to encourage accessibility through cutting-edge technology, attempting to meet people’s different requirements through technological innovation.
