Knowledge-and-Data-Driven Amplitude Spectrum Prediction for Hierarchical Neural Vocoders

Abstract

In our previous work, we have proposed a neural vocoder called HiNet which recovers speech waveforms by predicting amplitude and phase spectra hierarchically from input acoustic features. In HiNet, the amplitude spectrum predictor (ASP) predicts log amplitude spectra (LAS) from input acoustic features. This paper proposes a novel knowledge-and-data-driven ASP (KDD-ASP) to improve the conventional one. First, acoustic features (i.e., F0 and mel-cepstra) pass through a knowledge-driven LAS recovery module to obtain approximate LAS (ALAS). This module is designed based on the combination of STFT and source-filter theory, in which the source part and the filter part are designed based on input F0 and mel-cepstra, respectively. Then, the recovered ALAS are processed by a data-driven LAS refinement module which consists of multiple trainable convolutional layers to get the final LAS. Experimental results show that the HiNet vocoder using KDD-ASP can achieve higher quality of synthetic speech than that using conventional ASP and the WaveRNN vocoder on a text-to-speech (TTS) task.

arXiv commentSubmitted to Interspeech 2020

References in corpus (5)

• TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems
• WaveNet: A Generative Model for Raw Audio
• SampleRNN: An Unconditional End-to-End Neural Audio Generation Model
• Waveform Modeling and Generation Using Hierarchical Recurrent Neural Networks for Speech Bandwidth Extension
• Neural Harmonic-plus-Noise Waveform Model with Trainable Maximum Voice Frequency for Text-to-Speech Synthesis

Cited by in corpus (2)

• WaveGrad: Estimating Gradients for Waveform Generation
• Reverberation Modeling for Source-Filter-based Neural Vocoder