Understanding Voice Parameters

​

Overview

AI-generated speech can be customized using various parameters to achieve the desired speed, clarity, and similarity to reference audio. Below are the key parameters that allow you to fine-tune speech output.

​

Voice Parameters

​

1. Speed (speed)

• Description: Controls the rate at which speech is generated.
• Default Value: 1
• Allowed Range: 0.5 ≤ x ≤ 2
• Effects:
  • Decreasing speed (< 1) slows down the speech, making it clearer but longer.
  • Increasing speed (> 1) makes speech faster but may reduce clarity.

​

2. Consistency (consistency)

• Description: Manages word repetition and skipping to maintain speech fluency.
• Default Value: 0.5
• Allowed Range: 0 ≤ x ≤ 1
• Effects:
  • Lower values (< 0.5) reduce skipped words but may allow slight repetition.
  • Higher values (> 0.5) prevent repetition but may cause words to be skipped.

​

3. Similarity (similarity)

• Description: Adjusts how closely the generated speech matches the reference audio.
• Default Value: 0
• Allowed Range: 0 ≤ x ≤ 1
• Effects:
  • Higher values (> 0) make speech resemble the reference voice more closely.
  • Lower values (0) allow for more flexible voice generation.

​

4. Enhancement (enhancement)

• Description: Improves speech quality, with a trade-off in processing speed.
• Default Value: 1
• Allowed Range: 0 ≤ x ≤ 2
• Effects:
  • Increasing this value enhances speech clarity and naturalness.
  • Higher values may introduce latency due to additional processing.

​

Optimizing Speech Output

• Adjust Speed for fast or slow narration styles.
• Fine-tune Consistency to avoid missing or repeated words.
• Use Similarity to match a specific reference voice.
• Increase Enhancement for the best quality, but balance it with latency needs.

These parameters allow you to create a customized, natural-sounding AI voice experience. 🚀