← Back to papers
2022 INTERSPEECH INTERSPEECH 2022

Opencpop: A High-Quality Open Source Chinese Popular Song Corpus for Singing Voice Synthesis

Abstract

This paper introduces Opencpop, a publicly available high-quality Mandarin singing corpus designed for singing voicesynthesis (SVS). The corpus consists of 100 popular Mandarinsongs performed by a female professional singer. Audio filesare recorded with studio quality at a sampling rate of 44,100 Hzand the corresponding lyrics and musical scores are provided.All singing recordings have been phonetically annotated withphoneme boundaries and syllable (note) boundaries. To demon-strate the reliability of the released data and to provide a baselinefor future research, we built baseline deep neural network-basedSVS models and evaluated them with both objective metrics andsubjective mean opinion score (MOS) measure. Experimentalresults show that the best SVS model trained on our databaseachieves 3.70 MOS, indicating the reliability of the providedcorpus. Opencpop is released to the open-source community WeNet, and the corpus, as well as synthesized demos, can befound on the project homepage.

🌉 Interdisciplinary Bridge — Machine Learning and Speech & Audio
🧭 Keyword Pioneer — phoneme boundaries
🐝 Cross-Pollinator — Artificial Intelligence, Computer Science, Computer Vision, Data Science & Analytics, Deep Learning, Healthcare & Medicine, Interdisciplinary, Knowledge & Reasoning, Machine Learning, Mathematics & Optimization, Natural Language Processing, Reinforcement Learning, Robotics, Security & Privacy, Speech & Audio

Authors

Yu Wang , Xinsheng Wang , Pengcheng Zhu , Jie Wu , Hanzhao Li , Heyang Xue , Yongmao Zhang , Lei Xie , Mengxiao Bi

Topics

Machine Learning > Core Methods > Regression Machine Learning > Optimization & Theory > Statistical Learning Machine Learning > Application Areas > Knowledge Distillation Speech & Audio > Synthesis

Keywords

speech synthesis deep neural network singing voice synthesis mean opinion score phoneme boundaries phoneme annotation
Download PDF

Related papers

Example-based Explanations with Adversarial Attacks for Respiratory Sound Analysis 2022
Which Model is Best: Comparing Methods and Metrics for Automatic Laughter Detection in a Naturalistic Conversational Dataset 2022
Evidence of Onset and Sustained Neural Responses to Isolated Phonemes from Intracranial Recordings in a Voice-based Cursor Control Task 2022
Pre-trained Speech Representations as Feature Extractors for Speech Quality Assessment in Online Conferencing Applications 2022
Exploring the influence of fine-tuning data on wav2vec 2.0 model for blind speech quality prediction 2022