The DKU-JNU-EMA Electromagnetic Articulography Database
| Item Name: | The DKU-JNU-EMA Electromagnetic Articulography Database |
| Author(s): | Xiaoyi Qin, Xinzhong Liu, Zexin Cai, Ming Li |
| LDC Catalog No.: | LDC2019S14 |
| ISBN: | 1-58563-894-3 |
| ISLRN: | 147-070-436-975-2 |
| DOI: | https://doi.org/10.35111/61f9-zg33 |
| Release Date: | July 15, 2019 |
| Member Year(s): | 2019 |
| DCMI Type(s): | Sound, Text |
| Sample Type: | pcm |
| Sample Rate: | 16000 |
| Data Source(s): | microphone speech |
| Application(s): | language identification, phonetics, pronunciation modeling |
| Language(s): | Yue Chinese, Hakka Chinese, Min Nan Chinese, Mandarin Chinese |
| Language ID(s): | yue, hak, nan, cmn |
| License(s): |
LDC User Agreement for Non-Members |
| Online Documentation: | LDC2019S14 Documents |
| Licensing Instructions: | Subscription & Standard Members, and Non-Members |
| Citation: | Qin, Xiaoyi, et al. The DKU-JNU-EMA Electromagnetic Articulography Database LDC2019S14. Web Download. Philadelphia: Linguistic Data Consortium, 2019. |
| Related Works: | View |
Introduction
The DKU-JNU-EMA Electromagnetic Articulography Database was developed by Duke Kunshan University and Jinan University and contains approximately 10 hours of articulography and speech data in Mandarin, Cantonese, Hakka, and Teochew Chinese from two to seven native speakers for each dialect.
Electromagnetic articulography (EMA) is a method of measuring the position of parts of the mouth and their movement over time during speech and swallowing. Measurements are made from sensors placed in the mouth to capture real-time vocal tract variable trajectories. EMA is used in linguistics and language-related research to study phonetics, in particular, articulation (how sounds are made).
Data
Articulatory measurements were made using the NDI electromagnetic articulography wave research system. Subjects had six sensors placed in various locations in their mouth and one reference sensor was placed on the bridge of their nose. For simultaneous recording of speech signals, subjects also wore a head-mounted close-talk microphone.
Speakers engaged in four different types of recording sessions: one in which they read complete sentences or short texts, and three sessions in which they read related words of a specific common consonant, vowel or tone.
Audio data is presented as single channel, 16kHz, 16-bit flac compressed wav files. Articulography data is stored as UTF-8 plain text files.
Samples
Please view the following samples:
- Cantonese Speech
- Cantonese Sensor Data
- Cantonese Parameter Data
- Mandarin Speech
- Mandarin Sensor Data
- Mandarin Parameter Data
Updates
Chinese and English README files were updated on April 22, 2024.