PROFESSIONAL PROJECTS

Previous Work

Automatic tone alignment for any Mandarin text.

This tool was created to better analyze corpora or text. The tool allows you to create classes of words based on Mandarin tones. There are two separate outputs. The primary output is a list of vectors that contain all of the words of each tonal category to allow for acoustic analysis of particular tones. The secondary output is a multi-colored text which shows the of a text or corpora through color. The multicolored text is for ease of ensuring the tool works and also allows you to see an interesting output.

Web Crawling

This project was created in order to contribute to the Pinyin Tutor on Brian McWhinney's TalkBank. I used web crawling to pull open sources audio files for all possible Pinyin pronunciations of standard Mandarin. This input is very interesting because it allows for a more HVPT like training model (high variable phonetic training). I hope this tool could be useful for anyone trying to pull audio files for language training or machine learning NLP of real language sounds.