End-to-end 비자기회귀식 가속 음성합성기
End-to-end non-autoregressive fast text-to-speech
1
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Autoregressive Text-to-Speech (TTS) models suffer from inference instability and slow inference speed. Inference instability occurs when a poorly predicted sample at time step t affects all the subsequent predictions. Slow inference speed arises from a model structure that forces the predicted samples from time steps 1 to t-1 to predict the sample at time step t. In this study, an end-to-end non-autoregressive fast text-to-speech model is suggested as a solution to these problems. The results of this study show that this model's Mean Opinion Score (MOS) is close to that of Tacotron 2 - WaveNet, while this model's inference speed and stability are higher than those of Tacotron 2 - WaveNet. Further, this study aims to offer insight into the improvement of non-autoregressive models.
Abstract
Autoregressive Text-to-Speech (TTS) models suffer from inference instability and slow inference speed. Inference instability occurs when a poorly predicted sample at time step t affects all the subsequent predictions. Slow inference speed arises from a model structure that forces the predicted samples from time steps 1 to t-1 to predict the sample at time step t. In this study, an end-to-end non-autoregressive fast text-to-speech model is suggested as a solution to these problems. The results of this study show that this model's Mean Opinion Score (MOS) is close to that of Tacotron 2 - WaveNet, while this model's inference speed and stability are higher than those of Tacotron 2 - WaveNet. Further, this study aims to offer insight into the improvement of non-autoregressive models.
참고문헌 (18)
- "Deep voice: Real-time neural text-to-speech" (2017)[URL]
- "Boltzmann machines and denoising autoencoders for image denoising" (2013)[URL]
- Dvorak, J. L., "Moving wearables into the mainstream : Taming the Borg" (2011), Springer
- Griffin, D., "Signal estimation from modified short-time Fourier transform", Proceedings of the 8th International Conference on Acoustics, Speech, and Signal Processing, pp.804~807 (1983)
- Holmes, J., "Speech synthesis and recognition" (2002), CRC Press
- "MelGAN:Generative adversarial networks for conditional waveform synthesis" (2019)[URL]
- "FastSpeech: Fast, robust and controllable text to speech" (2019)[URL]
- "Natural TTS synthesis by conditioning Wavenet on mel spectrogram predictions" (2017)[URL]
- "Sequence to sequence learning with neural networks" (2014)[URL]
- "Efficiently trainable text-to-speech system based on deep convolutional networks with guided attention" (2017)[URL]
- "Flowtron: An autoregressive flow-based generative network for text-to-speech synthesis" (2020)[URL]
- "Attention is all you need" (2017)[URL]
- "WaveNet: A generative model for raw audio" (2016)[URL]
- "Parallel WaveNet: Fast high-fidelity speech synthesis" (2017)[URL]
- Wang, T., "Non-autoregressive end-to-end TTS with coarse-to-fine decoding", Proceedings of the 21st Annual Conference of the International Speech Communication Association, pp.3984~3988 (2020)
- "Tacotron: Towards end-to-end speech synthesis" (2017)[URL]
- "Parallel WaveGAN: A fast waveform generation model based on generative adversarial networks with multi-resolution spectrogram" (2019)[URL]
- Yarrington, D., "Audio anecdotes:Tools, tips and techniques for digital audio (Vol. 3)", pp.143~155 (2007), AK Peters