Effective Spectral and Excitation Modeling Techniques for LSTM-RNN-Based Speech Synthesis Systems

By Eunwoo Song, Frank K. Soong, Hong-Goo Kang. 2017

61 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Experimental and Cognitive Psychology, 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

In Experimental and Cognitive Psychology

In Signal Processing

Citations in the corpus

Parallel Wavegan: A Fast Waveform Generation Model Based on Generative Adversarial Networks with Multi-Resolution Spectrogram (2019, background): “Deep generative models in text-to-speech (TTS) frameworks have significantly improved the quality of synthetic speech signals [1–3].”
Probability density distillation with generative adversarial networks for high-quality parallel waveform generation (2019, background): “However, compared with traditional parametric vocoders [11–14], the WaveNet’s inference speed is inherently slow owing to its autoregressive model structure.”
Improved Parallel Wavegan Vocoder with Perceptually Weighted Spectrogram Loss (2021, methodology): “To enhance spectral clarity, an LSF-sharpening filter was applied to the spectral parameters [23].”
Neural Text-to-Speech with a Modeling-by-Generation Excitation Vocoder (2020, methodology): “The acoustic features were extracted using the improved time-frequency trajectory excitation vocoder at analysis intervals of 5 ms [19], and these features included 40-dimensional line spectral frequencies (LSFs), fundamental frequency (F0), energy, voicing ﬂag (v/uv), 32-dimensional slowly…”
TTS-by-TTS: TTS-Driven Data Augmentation for Fast and High-Quality Speech Synthesis (2020, methodology): “In the source TTS system, of the improved time-frequency trajectory excitation vocoder were extracted every 5 ms [22], which included 40-dimensional line spectral frequencies, fundamental frequency, energy, voicing flag, 32-dimensional slowly evolving waveform, and 4-dimensional rapidly evolving wav”
Improving LPCNET-Based Text-to-Speech with Linear Prediction-Structured Mixture Density Network (2020, methodology): “CONTENTS Introduction • LPCNet-based neural vocoding [1] Proposed system • Linear prediction-structured mixture density network • Improved LPCNet vocoder • Effective training and generation methods Experiments • Performance evaluations Summary & Conclusion”
Deep Learning for Mandarin-Tibetan Cross-Lingual Speech Synthesis (2022, background): “The main network architectures of deep learning include deep neural network (DNN) [7]–[9], long short-term memory (LSTM) and bi-directional long shortterm memory (BLSTM) [10].”
ExcitNet Vocoder: A Neural Excitation Model for Parametric Speech Synthesis Systems (2018, methodology): “As the spectral structure represented by LP coefficients, or by their equivalents such as line spectral frequencies (LSFs), is changing relatively slowly, it is easy to model with a simple deep learning framework [4].”