DiDiSpeech: A Large Scale Mandarin Speech Corpus
arXiv:2010.09275
Abstract
This paper introduces a new open-sourced Mandarin speech corpus, called DiDiSpeech. It consists of about 800 hours of speech data at 48kHz sampling rate from 6000 speakers and the corresponding texts. All speech data in the corpus is recorded in quiet environment and is suitable for various speech processing tasks, such as voice conversion, multi-speaker text-to-speech and automatic speech recognition. We conduct experiments with multiple speech tasks and evaluate the performance, showing that it is promising to use the corpus for both academic research and practical application. The corpus is available at https://outreach.didichuxing.com/research/opendata/.
5 pages, 2 figures, 11 tables
References in corpus (15)
- Adam: A Method for Stochastic Optimization
- VoxCeleb2: Deep Speaker Recognition
- Attention-Based Models for Speech Recognition
- MelGAN: Generative Adversarial Networks for Conditional Waveform Synthesis
- FastSpeech: Fast, Robust and Controllable Text to Speech
- Transfer Learning from Speaker Verification to Multispeaker Text-To-Speech Synthesis
- Mockingjay: Unsupervised Speech Representation Learning with Deep Bidirectional Transformer Encoders
- THCHS-30 : A Free Chinese Speech Corpus
- Tacotron: Towards End-to-End Speech Synthesis
- Improving Transformer-based Speech Recognition Using Unsupervised Pre-training
- DurIAN: Duration Informed Attention Network For Multimodal Synthesis
- The Voice Conversion Challenge 2018: Promoting Development of Parallel and Nonparallel Methods
- A Further Study of Unsupervised Pre-training for Transformer Based Speech Recognition
- Transformer based unsupervised pre-training for acoustic representation learning
- A comparable study of modeling units for end-to-end Mandarin speech recognition