A Comparison of Recent Neural Vocoders for Speech Signal Reconstruction

By Prachi Govalkar, Johannes Fischer, Frank Zalkow, Christian Dittmar. 2019

31 citations recorded by the source corpus.

In recent years, text-to-speech (TTS) synthesis has benefited from advanced machine learning approaches. Most prominently, since the introduction of the WaveNet architecture, neural vocoders have exhibited superior performance in terms of the naturalness of synthesized speech signals in comparison to traditional vocoders. In this paper, a fair comparison of recent neural vocoders is presented in a signal reconstruction scenario. That means we use such techniques to resynthesize speech waveforms from mel-scaled spectrograms, a compact and generally non-invertible representation of the underlying audio signal. In that context, we conduct listening tests according to the well established MUSHRA standard and compare the attained results to similar studies. Weighing off the perceptual quality to the computational requirements, our findings shall serve as a guideline to both practitioners and researchers in speech synthesis.

Subfields: Artificial Intelligence, Speech & Audio, 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. Sheng Zhao
6. Junichi Yamagishi
7. Erica Cooper
8. Yi Ren

In Speech & Audio

1. Xu Tan
2. Tao Qin
3. Tie-Yan Liu
4. Frank K. Soong
5. Junichi Yamagishi
6. Sheng Zhao
7. Erica Cooper
8. Yi Zhao

In Signal Processing

1. Xu Tan
2. Tao Qin
3. Tie-Yan Liu
4. Frank K. Soong
5. Junichi Yamagishi
6. Erica Cooper
7. Yi Zhao
8. Sheng Zhao

Citations in the corpus

• You Do Not Need More Data: Improving End-To-End Speech Recognition by Text-To-Speech Data Augmentation (2020, neutral): “It was chosen as a good compromise between quality and inference speed for a multispeaker TTS system [30].”
• A Survey on Neural Speech Synthesis (2021, background): “[88] conduct a comprehensive study on different kinds of vocoders.”
• Pretraining Strategies, Waveform Model Choice, and Acoustic Configurations for Multi-Speaker End-to-End Speech Synthesis (2020, background): “We can see that fine-tuning from the cleaner data is always better, significantly so in most cases (with exceptions for both naturalness and similarity of seen speakers using WaveNet, and similarity of test speakers using WaveRNN).”
• HUI-Audio-Corpus-German: A high quality TTS dataset (2021, background): “LJ Speech is a well-known audio dataset, that achieves good results in state-of-the-art TTS models [11].”