Open Speech and Language Resources



Contact
dpovey@gmail.com
Phone: 425 247 4129
(Daniel Povey)

HI-MIA

Identifier: SLR85

Summary: A far-field text-dependent speaker verification database for AISHELL Speaker Verification Challenge 2019

Category: Speech

License: Apache License v.2.0

Downloads (use a mirror closer to you):
train.tar.gz [36G]   (Training set with speaker dependent sub folders )   Mirrors: [China]  
dev.tar.gz [5.1G]   (Dev set with speaker dependent sub folders )   Mirrors: [China]  
test.tar.gz [4.7G]   (Test set with target/non-target answer )   Mirrors: [China]  
test_v2.tar.gz [4.7G]   (Updated test set fixing corrupted audio files )   Mirrors: [China]  
filename_mapping.tar.gz [5.9M]   (Filename mapping rules for multi-channel information )   Mirrors: [China]  

About this resource:

The data is used in AISHELL Speaker Verification Challenge 2019. It is extracted from a larger database called AISHELL-WakeUp-1.

The contents are wake-up words "Hi, Mia" in both Chinese and English. The data is collected in real home environment using microphone arrays and Hi-Fi microphone. The collection process and development of a baseline system was described in the paper below. The data used in the challenge is extracted from 1 Hi-Fi microphone and 16-channel circular microphone arrays for 1/3/5 meters. And the contents are the Chinese wake-up words. The whole set is divided into train (254 people), dev (42 people) and test (44 people) subsets. Test subset is provided with paired target/non-target answer to evaluate verification results.

You can cite the data using the following BibTeX entry:


@misc{himia,
    title={HI-MIA : A Far-field Text-Dependent Speaker Verification Database and the Baselines},
    author={Xiaoyi Qin and Hui Bu and Ming Li},
    year={2019},
    eprint={1912.01231},
    archivePrefix={arXiv},
    primaryClass={cs.SD}
}

External URL: http://aishelltech.com/wakeup_data