AI-900Domain 4: Describe Features of Natural Language Processing Workloads on AzureNLP & SpeechmediumSingle choice
An accessibility app reads articles aloud in natural-sounding voices. Which capability — and which Speech feature makes the voices sound human?
A.Speech-to-text with custom models
Reverse direction — that's audio in, text out.
B.Text-to-speech with neural voices
TTS + neural voices. Correct.
C.Speaker verification
Speaker verification identifies a speaker — doesn't generate speech.
D.Real-time speech translation
Translation changes language — unrelated.
Why this answer
Text-to-speech synthesizes audio from text; neural voices (deep-learning-based) produce humanlike prosody vs older concatenative voices. Voice assistants, audiobooks, and accessibility all use this.
Correct answer: B · Last reviewed 2026-02-14
Official AWS references
Practice this interactively
Answer without seeing the solution first in the NLP & Speech practice session.
Practice NLP & Speech