The XBMU-AMDO31 is a transcribed audio corpus of Tibetan Amdo dialect. It contains about 31 hours of recordings read by 66 native speakers. There are 22630 sentences. The speech recording was conducted in a quiet room. The recording devices are mobile phones. The sampling frequency of the audio files is 16 KHz and the quantization accuracy is 16 bits. The total number of words are 2754. The corpus was constructed by Li Guanyu’s research group of Northwest Minzu University, China.

You can cite the data using the following BibTeX entry:

@inproceedings{
  title={{XBMU-AMDO31:An open source of Amdo Tibetan speech
database and speech recognition baseline system}},
  author={Senyan Li, Guanyu Li, Jiewen Ning},
  booktitle={National Conference on Man-Machine Speech Communication,NCMMSC2022},
  year={2022},
}