Yandex Cloud
Search
Contact UsGet started
  • Pricing
  • Customer Stories
  • Documentation
  • Blog
  • All Services
  • System Status
    • Featured
    • Infrastructure & Network
    • Data Platform
    • Containers
    • Developer tools
    • Serverless
    • Security
    • Monitoring & Resources
    • AI for business
    • Business tools
  • All Solutions
    • By industry
    • By use case
    • Economics and Pricing
    • Security
    • Technical Support
    • Start testing with double trial credits
    • Cloud credits to scale your IT product
    • Gateway to Russia
    • Cloud for Startups
    • Center for Technologies and Society
    • Yandex Cloud Partner program
  • Pricing
  • Customer Stories
  • Documentation
  • Blog
© 2025 Direct Cursus Technology L.L.C.
Yandex SpeechKit
  • SpeechKit technology overview
    • About the technology
    • List of voices
      • Overview
      • Data for SpeechKit Brand Voice Lite
      • Data for SpeechKit Brand Voice Call Center
      • Tips for recording roles
  • Supported audio formats
  • IVR integration
  • Quotas and limits
  • Access management
  • Pricing policy
  • Audit Trails events

In this article:

  • Neutral role (default voice)
  • Friendly role
  • Good role
  • Strict role
  • Evil role
  1. Speech synthesis
  2. SpeechKit Brand Voice
  3. Tips for recording roles

Tips for recording roles

Written by
Yandex Cloud
Updated at July 10, 2025
  • Neutral role (default voice)
  • Friendly role
  • Good role
  • Strict role
  • Evil role

You can create multiple roles for your voice. SpeechKit uses the following roles:

  • neutral
  • friendly
  • good
  • strict
  • evil
  • whisper

All roles are different in terms of manner, pitch, tempo, volume, and pauses. The recommendations below are based on descriptions of natural live human speech aiming to express particular emotions. Try to follow these suggestions when recording for a particular role.

When creating a role, make sure the audio sounds fairly different from your neutral speech but still naturally enough. Base your manner mainly on the delivery description; the suggested speech parameters are but useful hints.

Neutral role (default voice)Neutral role (default voice)

Delivery: Speak as you normally would with someone you do not know, aiming to clearly communicate your point while keeping the tone neutral and friendly.

Pitch: Go with the pitch that feels most natural and comfortable for you in daily conversations.

Intonation: Follow the logical flow of the text without expressing emotions. Emphasize points central to the message of the sentence. Avoid emphasizing emotionally loaded words or adding dramatic tones to the phrase.

Tempo: Ideally, moderate. Pick a comfortable tempo you can maintain throughout your recording session for the role. Avoid accelerating or decelerating as you navigate through the phrase or based on sentiment.

Pauses: Pause where it makes sense, just like you would in everyday speech.

Volume: Even. Pick a volume you can maintain throughout the recording session without losing your voice and breath. Avoid varying volume based on the sentiment of the phrase.

Friendly roleFriendly role

Delivery: Speak as if you feel tenderness for the listener and want them surrounded with care and attention.

Pitch: You will probably be more comfortable speaking at or below your neutral pitch.

Intonation: Use more pronounced intonations with stronger and more frequent emphases, more tonal variations, and more downward inflections than neutral.

Tempo: Likely slower than neutral.

Pauses: Pause where it makes sense, just like you would in everyday speech. You will likely use longer logical pauses than neutral.

Volume: Lower than neutral. Keep your volume consistent throughout your recording session for the friendly role.

Good roleGood role

This role should sound fairly different from your neutral speech but still feel relatively natural to you. Base your manner mainly on the delivery description; the suggested speech parameters are but useful hints.

Delivery: Speak as if you want to uplift the listener, share your enthusiasm, and inspire confidence in what you say and what lies ahead.

Pitch: You will probably want to speak at or above your neutral pitch.

Intonation: Use more pronounced intonations with more upward inflections, stronger and more frequent emphases, more tonal variations than neutral.

Tempo: Most likely, the overall tempo will remain moderate as in your neutral speaking style. The speech rate will increase, but logical pauses will become longer.

Pauses: Make logical pauses more pronounced than neutral; feel free to skip the less significant ones.

Volume: Higher than neutral. Keep your volume consistent throughout your recording session for this role.

Strict roleStrict role

Delivery: Speak as if you are talking about something too serious for a smile, or as if you are talking to someone who is in a joyless situation or someone you feel no connection with.

Pitch: Lower than neutral.

Intonation: More reserved than neutral, with only the key logical points emphasized. Emphasized logical points have stronger tonal variations; less important ones may be skipped. Use fewer upward inflections than neutral.

Tempo: Most likely, the overall tempo will remain moderate as in your neutral speaking style. The speech rate will increase, but logical pauses will become longer.

Pauses: More pronounced than neutral. Perhaps with longer logical pauses.

Volume: Perhaps a little softer than neutral.

Evil roleEvil role

Delivery: Speak as if you need to make a willful person to comply while staying within the bounds of politeness and decency.

Pitch: You will probably be more comfortable speaking at or slightly below your neutral pitch. Lowering your pitch too far may lead to unwanted effects.

Intonation: Use more pronounced intonations with stronger and more frequent emphases, more tonal variations than neutral.

Tempo: Definitely faster than neutral. Slowing down too much may lead to unwanted effects.

Pauses: More pronounced than neutral.

Volume: Higher than neutral.

Was the article helpful?

Previous
Data for SpeechKit Brand Voice Call Center
Next
About the technology
© 2025 Direct Cursus Technology L.L.C.