An English tutoring app asks students to read a passage aloud and must score how accurately each sound was produced, how smoothly the student spoke, and whether every word was said. Which Speech feature should the developer use?
Show answer and explanation
Answer: D
Pronunciation assessment compares spoken audio against the reference text and returns scores for accuracy, fluency, and completeness, along with word-level feedback. Those measures are exactly what the tutoring app needs. Speaker diarization separates voices in a recording, batch synthesis generates speech from text, and a custom neural voice creates a synthetic voice rather than evaluating a learner.
Why the other options are wrong
- A. Speaker diarization identifies which participant spoke each segment of a recording. It does not score pronunciation or fluency against a reference passage.
- B. Batch synthesis converts many text inputs into audio asynchronously. It produces spoken output for a listener, not an evaluation of the learner's speech.
- C. A custom neural voice is a synthetic voice model used for text to speech output. It does not measure how well a student pronounces the words in a passage.