A ship-caulker records a long take that includes the phrases “oakum, then pitch.” One annotator marks when each phrase starts; another writes a single mood note on the whole file. Which annotation helps the speech recognizer learn from the clip?

Short Explanation

Imagine sheet music with bar lines versus a sticky note that just says “energetic.” The recognizer needs words pinned to time, not a floating mood remark.

Full Explanation

Time-aligned transcript annotation ties spoken content to moments in the audio so ASR customization can learn from it. An unanchored mood note, an image caption, or a deploy setting does not provide that speech timing signal.

×