Attention network based on duration information for text-to-speech analysis

摘要

The method and apparatus include receiving a text input comprising a sequence of text components. Each time duration of the text component is determined using a duration model. A first set of spectra is generated based on the sequence of text components. A second set of spectra is generated based on the first set of spectra and the respective temporal durations of the text component sequence. A spectrogram frame is generated based on the second set of spectra. The audio waveform is generated based on the spectrogram frame. An audio waveform is provided as an output.