Voice Styles & Emotions¶
Many voices in Speech Studio support multiple speaking styles and emotional tones. This guide explains how to use these styles to make your audio more expressive and engaging.
What Are Voice Styles?¶
Voice styles change the way a voice delivers text. The same words can sound cheerful, serious, sad, or excited depending on the selected style. Styles are powered by Azure's neural voice technology and go far beyond simple pitch or speed adjustments.
Available Styles¶
Not every voice supports every style. The table below lists the most common styles and which types of voices typically support them.
| Style | Description | Best For |
|---|---|---|
cheerful |
Upbeat and positive tone | Marketing, announcements |
sad |
Somber and mournful delivery | Dramatic content, storytelling |
angry |
Forceful and intense | Character dialog, dramatic reads |
excited |
High energy and enthusiastic | Promotions, sports |
friendly |
Warm and approachable | Customer service, tutorials |
terrified |
Fearful, trembling voice | Storytelling, audiobooks |
whispering |
Soft, quiet delivery | ASMR, intimate narration |
shouting |
Loud, projected voice | Announcements, alerts |
newscast |
Professional news anchor style | News reading, reports |
narration |
Storytelling delivery | Audiobooks, documentaries |
customer-service |
Polite and helpful | IVR systems, support bots |
How to Apply Styles¶
Method 1: Style Dropdown¶
- Select a voice that supports styles (indicated by a style icon next to the voice name)
- Open the Style dropdown that appears below the voice selector
- Choose a style from the list
- The entire text will be spoken in that style
Method 2: SSML Tags¶
For more granular control, use SSML to apply different styles to different parts of your text:
<mstts:express-as style="cheerful">
Welcome to our show!
</mstts:express-as>
<mstts:express-as style="serious">
Now let's discuss the important details.
</mstts:express-as>
See SSML Support for the complete SSML reference.
Style Intensity¶
Some voices support adjustable style intensity using the styledegree attribute in SSML. Values range from 0.01 to 2.0, where 1.0 is the default:
<mstts:express-as style="cheerful" styledegree="2">
This is extremely cheerful!
</mstts:express-as>
Finding Styled Voices
In the voice selection panel, use the Has Styles filter to show only voices that support multiple speaking styles. The most expressive voices are typically the newer neural voices for major languages like English, Chinese, and Japanese.