This research evaluates the quality of text-to-speech (TTS) voices in delivering mindfulness-based therapies. It explores the potential of technology in enhancing these therapies, focusing on the user-perceived quality of TTS voices, their emotional expressiveness, and the effects of personalization. The study found that while human voices were rated higher than TTS voices, user-personalized TTS voices performed almost as well as human voices. This suggests that user personalization could significantly improve the perceived quality of TTS voices in mindfulness therapies.
Publication date: 9 Jan 2024
Project Page: https://doi.org/10.1145/3568162.3576987
Paper: https://arxiv.org/pdf/2401.03581
