You are tasked to fine tune text to speech using Speech service core features in Azure. Based on the requirements, you want to control certain aspects of speech output such as adjust pitch or add pauses. This will improve the quality of the synthesized content. You also want to define your own lexicons and switch between different speaking styles. 
Given the requirement, which core feature of text to speech service would you use?

Answer options:

A.Asynchronous Synthesis of Long Audio
B.Neural Voices
C.Visemes
D.Speech Synthesis Markup Language

Correct Answer: D
Option A is incorrect because the long audio, as the name suggests is used for the speech files that are longer than 10 minutes. They are meant for asynchronous responses.
Option B is incorrect because Neural voices are human like voices that utilize deep neural networks to overcome the limitations of traditional speech synthesis. They are used in synthesizing speech for the audio books, chatbots and voice assistants. 
Option C is incorrect because Visemes provide a visual description based on the position of face, lips, jaw and tongue. They are used in two dimensional and three dimensional models to match speech movement of the mouth.
Option D is correct. Speech Synthesis Markup Language or SSML is meant to tune the speech to text output. Using SSML you can change speaking rate, volume, adjust pitch and add pauses. You can also define your own lexicons and switch between different speaking styles
Reference:
To learn more about core features of text to speech, use the link given below: 

https://docs.microsoft.com/en-us/azure/cognitive-services/speech-service/text-to-speech