GELP: GAN-Excited Liner Prediction for Speech Synthesis from Mel-spectrogram

By Lauri Juvela, Bajibabu Bollepalli, Junichi Yamagishi, Paavo Alku. 2019

18 citations recorded by the source corpus.

Recent advances in neural network -based text-to-speech have reached human level naturalness in synthetic speech. The present sequence-to-sequence models can directly map text to mel-spectrogram acoustic features, which are convenient for modeling, but present additional challenges for vocoding (i.e., waveform generation from the acoustic features). High-quality synthesis can be achieved with neural vocoders, such as WaveNet, but such autoregressive models suffer from slow sequential inference. Meanwhile, their existing parallel inference counterparts are difficult to train and require increasingly large model sizes. In this paper, we propose an alternative training strategy for a parallel neural vocoder utilizing generative adversarial networks, and integrate a linear predictive synthesis filter into the model. Results show that the proposed model achieves significant improvement in inference speed, while outperforming a WaveNet in copy-synthesis quality.

Subfields: Artificial Intelligence, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Xu Tan
2. Tao Qin
3. Tie-Yan Liu
4. Frank K. Soong
5. Lei He
6. Sheng Zhao
7. Kai Yu
8. Zhijun Liu

In Signal Processing

1. Xu Tan
2. Tao Qin
3. Tie-Yan Liu
4. Frank K. Soong
5. Lei He
6. Zhijun Liu
7. Kai Yu
8. Yang Cui

Citations in the corpus

• A Survey on Neural Speech Synthesis (2021, methodology): “Other Vocoders Some works leverage neural-based source-filter model for waveform generation [375, 374, 371, 207, 146, 145, 75, 305, 407], aiming to achieve high voice quality while maintaining controllable speech generation.”
• Conditional Spoken Digit Generation with StyleGAN (2020, methodology): “GANbased waveform synthesis models that are capable of fast parallel inference have recently been proposed as alternative to autoregressive WaveNets for the latter task [11, 12, 13].”
• An Efficient Subband Linear Prediction for LPCNet-Based Neural Synthesis (2020, methodology): “Based on the source-filter model, some neural vocoders can get natural speech with smaller model and higher efficiency, such as LPCNet[8], LP-WaveNet[9], neural source-filter model[10], and glottal neural vocoders[11, 12, 13, 14].”
• Neural Homomorphic Vocoder (2020, background): “Many models aim to improve source signal modeling in a source-filter model, including LPCNet [12], GELP [13], GlotGAN [14].”