Sorry, you need to enable JavaScript to visit this website.

Transformer-based text-to-speech with weighted forced attention

Citation Author(s):
Takuma Okamoto, Tomoki Toda, Yoshinori Shiga, Hisashi Kawai
Submitted by:
Takuma Okamoto
Last updated:
6 May 2020 - 9:36pm
Document Type:
Presentation Slides
Document Year:
Takuma Okamoto
Paper Code:

This paper investigates state-of-the-art Transformer- and FastSpeech-based high-fidelity neural text-to-speech (TTS) with full-context label input for pitch accent languages. The aim is to realize faster training than conventional Tacotron-based models. Introducing phoneme durations into Tacotron-based TTS models improves both synthesis quality and stability. Therefore, a Transformer-based acoustic model with weighted forced attention obtained from phoneme durations is proposed to improve synthesis accuracy and stability, where both encoder–decoder attention and forced attention are used with a weighting factor. Furthermore, FastSpeech without a duration predictor, in which the phoneme durations are predicted by another conventional model, is also investigated. The results of experiments using a Japanese female corpus and the WaveGlow vocoder indicate that the proposed Transformer using forced attention with a weighting factor of 0.5 outperforms other models, and removing the duration predictor from FastSpeech improves synthesis quality, although the proposed weighted forced attention does not improve synthesis stability.

IEEE Xplore:

Presentation video:

Demo samples:

0 users have voted: