KeSpeech: An Open Source Speech Dataset of Mandarin and Its Eight Subdialects.
Zhiyuan TangDong WangYanguang XuJianwei SunXiaoning LeiShuaijiang ZhaoCheng WenXingjun TanChuandong XieShuran ZhouRui YanChenjia LvYang HanWei ZouXiangang LiPublished in: NeurIPS Datasets and Benchmarks (2021)
Keyphrases
- open source
- speech recognition
- broadcast news
- emotion recognition
- spoken document retrieval
- open source software
- speech signal
- speaker independent
- case study
- prosodic features
- benchmark datasets
- speaker identification
- audio visual
- automatic speech recognition
- speech recognizer
- hidden markov models
- dialogue system
- noisy environments
- open source projects
- multi modal
- text to speech
- speech synthesis
- language model
- speaker diarization
- real world