Automated Vowel Extraction with Manual Transcriptions and Utterance Boundaries

Upload your audio files and manual transcriptions (with sentence boundaries annotated as described here) to run forced alignment followed by formant extraction.

Click here at any time to cancel or start over.

Upload a .wav or .mp3 file:
Longer recordings (of at least 2 minutes) are recommended. Your uploaded files are stored temporarily on the Dartmouth servers in order to process your job, and deleted after.

Manual transcription as a .TextGrid file:
Textgrid should contain a tier named "sentence" with sentence/breath group intervals.

Filter out stop-words? Yes No
This is the list of stop-words we remove. (Link opens in a new tab.)

Filter out unstressed vowels? Yes No

Filter out vowels whose F1 or F2 bandwidths are over 300? Yes No

Your e-mail address:
We will not store or distribute your address.