Machine annotation
Build your own machine annotator (BYOMA) for your study by selecting target words and phones and configuring how articulatory targets are detected. BYOMA runs in your browser using kinematic and acoustic data folders.
Build your own machine annotator (BYOMA) for your study by selecting target words and phones and configuring how articulatory targets are detected. BYOMA runs in your browser using kinematic and acoustic data folders.
Review and correct annotated files. Requires kinematic and acoustic data folders (outputs from the preprocessing step) and runs in your browser.
BYOMA lets you define rules for automatically labeling articulatory achievements across subjects. The Hand Annotator lets you inspect annotations and adjust landmark positions by hand.
You need kinematic data describing articulator movement and acoustic alignment data identifying when words and speech sounds occur. Organize both into folders by subject, with matching filenames for each recording.
Here, acoustic data means tables of word and speech-sound labels with their start and end times. Audio recordings are a separate, optional input in the Hand Annotator.
Launch the tool you want to use and select your study folder and data folders. In BYOMA, choose the words and speech sounds to annotate and define your annotation rules. In the Hand Annotator, select a recording to review and correct its landmarks.
Your research data stays on your machine. Both tools run in your browser and read and write the folders you select.
Corrections are saved in your study folder as separate files. Your original input files are not changed.
Use Chrome or Edge so the tools can access the local folders you select.
Your browser may ask you to grant folder access again. Re-select your study folder and data folders when prompted.