Is it possible to transcribe audio in D-Lab3?

We have recorded eye-tracking data with a participant wearing mascara, and D-Lab has wrongly identified her eyelashes as her pupil. How do we fix that without going frame-by-frame?