No abstract

This paper describes a method to control prosodic features using phonetic and prosodic symbols as input of attention-based sequenceto-sequence (seq2seq) acoustic modeling (AM) for neural text-to-speech (TTS). The method involves inserting a sequence of prosodic symbols between phonetic symbols that are then used to reproduce prosodic acoustic features, i.e. accents, pauses, accent breaks, and sentence endings, in several seq2seq AM methods. The proposed phonetic and prosodic labels have simple descriptions and a low production cost. By contrast, the labels of conventional statistical parametric speech synthesis methods are complicated, and the cost of time alignments such as aligning the boundaries of phonemes is high. The proposed method does not need the boundary positions of phonemes. We propose an automatic conversion method for conventional labels and show how to automatically reproduce pitch accents and phonemes. The results of objective and subjective evaluations show the effectiveness of our method.

No abstract

This paper presents a portable real-time speech rate conversion system for compensating degeneration of hearing intelligibility of elderly listeners who are suffering from listening to rapid speech. This system enables an elderly user to convert a speech rate as desired by him/herself on real time, with invariance in pitch as well as small impairments in quality. Conventional hearing aids are focusing on compensation of decreased hearing abilities of peripheral auditory pathways in the frequency area, while the new system tries to compensate for them with central auditory pathways in the temporal area. When this system is applied to speech accompanied with a picture, temporal discrepancy between the converted voice and the picture can be absorbed by changing the speech rate at every pitch period along a monotone decreasing function from slow to fast. The system is 180×130×65 mm in size. Another feature allows a user to change the speech rate quickly whenever he/she wants to convert at any stage of conversion. This system can be applied to both the Japanese spoken language and foreign spoken languages, and can be used as a listening aid to learn foreign languages.

No abstract