Get Voice from Description Result

​

Understanding Voice Generation Results

​

How to Access Your Voice Interpretations

GET /text-to-voice-result/{run_id}

GET /text-to-voice-result/8827

​

Understanding the Response

​

Preview Variations Explained

1. Preview 1: Primary Interpretation This first interpretation closely adheres to the core parameters of your description. It represents the system’s primary understanding of your requested voice characteristics, focusing on the fundamental attributes you specified such as gender, age range, and basic tonal qualities.
2. Preview 2: Nuanced Alternative The second interpretation explores alternative expressions of your description with modified pacing and emphasis patterns. This version might adjust factors like speaking rate, rhythmic patterns, and the relative emphasis placed on different syllables or words, creating a subtly different feel while still honoring your core description.
3. Preview 3: Boundary Exploration The third interpretation takes a more experimental approach, testing the boundaries of your description parameters. This version might intensify certain characteristics or introduce complementary vocal qualities that weren’t explicitly stated but might enhance the overall effect of the voice.

• Emotional cadence variations: How emotional qualities rise and fall throughout the speech
• Prosody adjustments: Alterations in intonation patterns, rhythm, stress, and tonal qualities
• Phonetic emphasis patterns: Changes in how individual sounds are articulated and emphasized

​

From Preview to Permanent Voice

​

Creating a Permanent Voice Profile

​

Implementation Example

import requests

BASE_URL = "https://client.camb.ai/apis"

def create_voice_from_preview(run_id, preview_index=0):
    """
    Converts temporary preview into permanent custom voice

    This function performs a complete workflow:
    1. Retrieves the specified preview audio using the run_id
    2. Downloads the selected preview to a local file
    3. Creates a permanent voice clone from this sample

    Parameters:
        run_id (int): The unique identifier of your voice generation task
        preview_index (int): Which preview to use (0, 1, or 2, corresponding to the three previews)

    Returns:
        dict: Response from the custom voice creation API
    """
    # Retrieve preview URL from the results endpoint
    preview_res = requests.get(f"{BASE_URL}/text-to-voice-result/{run_id}")
    preview_url = preview_res.json()['previews'][preview_index]

    # Download the selected audio sample to a local file
    audio_file = requests.get(preview_url).content
    with open("voice_sample.mp3", "wb") as f:
        f.write(audio_file)

    # Create a permanent custom voice using the downloaded sample
    files = {'file': open('voice_sample.mp3', 'rb')}
    data = {
        'voice_name': 'My Campaign Voice',  # Choose a descriptive name for your voice
        'gender': 1,  # Male (matches preview characteristics)
        'age': 35     # Estimated vocal age
    }

    creation_res = requests.post(
        f"{BASE_URL}/create-custom-voice",
        files=files,
        data=data,
        headers={"x-api-key": API_KEY}
    )

    return creation_res.json()

# Example usage
voice_profile = create_voice_from_preview(run_id=8827, preview_index=1)
print(f"New voice ID: {voice_profile['voice_id']}")

​

Best Practices for Voice Selection

1. Content context: Different voices may be more suitable depending on whether your content is informational, narrative, promotional, or instructional.
2. Audience expectations: Consider what vocal qualities would resonate best with your target audience based on their demographic and psychographic characteristics.
3. Brand alignment: If the voice will represent your brand, assess which interpretation best embodies your brand personality and values.
4. Emotional impact: Listen for how each voice conveys emotional subtext and choose the one that elicits the desired emotional response.
5. Technical qualities: Consider practical aspects like clarity, intelligibility across different listening environments, and how well the voice handles specialized vocabulary.