Voice Change

Convert voice characteristics in audio using .GENVoiceChange().

Basic Usage

AudioClip original = Resources.Load<AudioClip>("VoiceRecording");
AudioClip changed = await original
    .GENVoiceChange()
    .SetTargetVoice(ElevenLabsVoice.Adam)
    .ExecuteAsync();

Input Types

AudioClip Input

AudioClip original = Resources.Load<AudioClip>("Voice");
AudioClip changed = await original
    .GENVoiceChange()
    .SetTargetVoice(newVoice)
    .ExecuteAsync();

File Input

var audioFile = new File<AudioClip>(audioClip, "voice.wav");
AudioClip changed = await audioFile
    .GENVoiceChange()
    .SetTargetVoice(targetVoice)
    .ExecuteAsync();

Configuration

Target Voice

AudioClip changed = await originalVoice
    .GENVoiceChange()
    .SetTargetVoice(ElevenLabsVoice.Rachel)
    .ExecuteAsync();

Voice Settings

AudioClip changed = await originalVoice
    .GENVoiceChange()
    .SetTargetVoice(ElevenLabsVoice.Adam)
    .SetStability(0.5f)
    .SetSimilarityBoost(0.8f)
    .ExecuteAsync();

Unity Integration Examples

Example 1: Character Voice System

public class CharacterVoiceSystem : MonoBehaviour
{
    [SerializeField] private Voice maleVoice = ElevenLabsVoice.Adam;
    [SerializeField] private Voice femaleVoice = ElevenLabsVoice.Rachel;
    
    public async UniTask<AudioClip> ApplyCharacterVoice(
        AudioClip recording, 
        CharacterGender gender)
    {
        Voice targetVoice = gender == CharacterGender.Male 
            ? maleVoice 
            : femaleVoice;
        
        return await recording
            .GENVoiceChange()
            .SetTargetVoice(targetVoice)
            .ExecuteAsync();
    }
}

public enum CharacterGender { Male, Female }

Example 2: NPC Voice Customization

public class NPCVoiceCustomizer : MonoBehaviour
{
    private Dictionary<string, Voice> npcVoices = new();
    
    void Start()
    {
        npcVoices["shopkeeper"] = ElevenLabsVoice.Antoni;
        npcVoices["guard"] = ElevenLabsVoice.Arnold;
        npcVoices["child"] = ElevenLabsVoice.Rachel;
    }
    
    public async UniTask<AudioClip> GetNPCVoice(string npcType, AudioClip baseAudio)
    {
        if (!npcVoices.ContainsKey(npcType))
            return baseAudio;
        
        return await baseAudio
            .GENVoiceChange()
            .SetTargetVoice(npcVoices[npcType])
            .ExecuteAsync();
    }
}

Example 3: Voice Effects System

public class VoiceEffectsSystem : MonoBehaviour
{
    public async UniTask<AudioClip> ApplyRobotVoice(AudioClip humanVoice)
    {
        // Use a deep, mechanical-sounding voice
        return await humanVoice
            .GENVoiceChange()
            .SetTargetVoice(ElevenLabsVoice.Adam)
            .SetStability(0.2f)  // More robotic
            .ExecuteAsync();
    }
    
    public async UniTask<AudioClip> ApplyGhostVoice(AudioClip normalVoice)
    {
        // Use ethereal, whispery voice
        return await normalVoice
            .GENVoiceChange()
            .SetTargetVoice(ElevenLabsVoice.Serena)
            .SetStability(0.8f)  // More consistent
            .ExecuteAsync();
    }
}

Example 4: Voice Aging System

public class VoiceAgingSystem : MonoBehaviour
{
    public async UniTask<AudioClip> AgeVoice(AudioClip youngVoice, int targetAge)
    {
        Voice targetVoice = targetAge switch
        {
            < 18 => ElevenLabsVoice.Rachel,    // Young
            < 40 => ElevenLabsVoice.Antoni,    // Adult
            < 60 => ElevenLabsVoice.Adam,      // Middle-aged
            _ => ElevenLabsVoice.Arnold         // Elderly
        };
        
        return await youngVoice
            .GENVoiceChange()
            .SetTargetVoice(targetVoice)
            .ExecuteAsync();
    }
}

Example 5: Voice Anonymizer

public class VoiceAnonymizer : MonoBehaviour
{
    private List<Voice> anonymousVoices = new()
    {
        ElevenLabsVoice.Adam,
        ElevenLabsVoice.Antoni,
        ElevenLabsVoice.Arnold
    };
    
    public async UniTask<AudioClip> AnonymizeVoice(AudioClip originalVoice)
    {
        // Pick random voice
        Voice randomVoice = anonymousVoices[Random.Range(0, anonymousVoices.Count)];
        
        return await originalVoice
            .GENVoiceChange()
            .SetTargetVoice(randomVoice)
            .ExecuteAsync();
    }
}

Example 6: Real-time Voice Chat Modifier

public class VoiceChatModifier : MonoBehaviour
{
    [SerializeField] private Voice desiredVoice;
    private Queue<AudioClip> processingQueue = new();
    
    public void OnMicrophoneInput(AudioClip recording)
    {
        ProcessVoiceAsync(recording).Forget();
    }
    
    async UniTaskVoid ProcessVoiceAsync(AudioClip recording)
    {
        AudioClip modified = await recording
            .GENVoiceChange()
            .SetTargetVoice(desiredVoice)
            .ExecuteAsync();
        
        // Play modified voice
        AudioSource.PlayClipAtPoint(modified, Camera.main.transform.position);
    }
}

Provider Support

ElevenLabs

AudioClip changed = await audioClip
    .GENVoiceChange()
    .SetTargetVoice(ElevenLabsVoice.Rachel)
    .SetStability(0.5f)
    .SetSimilarityBoost(0.8f)
    .ExecuteAsync();

Features:

• ✅ Natural voice conversion

• ✅ Multiple target voices

• ✅ Voice characteristic control

• ✅ Preserves speech timing

Note: Currently, ElevenLabs is the primary provider for voice change.

Use Cases

Use Case	Example
Character Voices	Convert player voice to NPC voice
Privacy	Anonymize voice in recordings
Localization	Match voice to character ethnicity
Effects	Robot, ghost, monster voices
Accessibility	Convert to clearer voice
Age Matching	Make voice sound older/younger

Best Practices

✅ Good Practices

// ✅ Cache converted voices
Dictionary<string, AudioClip> cache = new();

// ✅ Choose appropriate target voice
Voice targetVoice = character.Gender == Gender.Male 
    ? ElevenLabsVoice.Adam 
    : ElevenLabsVoice.Rachel;

// ✅ Preserve original for comparison
AudioClip original = audioClip;
AudioClip modified = await original.GENVoiceChange()...;

// ✅ Clean up unused clips
Destroy(audioClip);

❌ Bad Practices

// ❌ Don't convert in Update()
void Update()
{
    await audio.GENVoiceChange().ExecuteAsync();  // NO!
}

// ❌ Don't convert silent audio
if (audioClip.length == 0) return;  // Check first

// ❌ Don't forget cleanup
// Memory leak if clips aren't destroyed

// ❌ Don't block main thread
AudioClip clip = audio.GENVoiceChange().ExecuteAsync().Result;  // Blocks!

Audio Requirements

Supported formats:

• WAV, MP3, M4A, FLAC, OGG

Quality requirements:

• Clear speech (minimal background noise)

• Good recording quality

• No heavy distortion

Limits:

• Max file size: 25 MB

• Duration: varies by provider

Error Handling

try
{
    AudioClip changed = await audioClip
        .GENVoiceChange()
        .SetTargetVoice(targetVoice)
        .ExecuteAsync();
    
    if (changed == null || changed.length == 0)
        throw new Exception("Voice change failed");
    
    audioSource.clip = changed;
    audioSource.Play();
}
catch (AIApiException ex)
{
    Debug.LogError($"Voice change failed: {ex.Message}");
}
catch (Exception ex)
{
    Debug.LogError($"Unexpected error: {ex.Message}");
}

Performance Tips

// ✅ Good - cache results
Dictionary<string, AudioClip> voiceCache = new();

async UniTask<AudioClip> GetCachedVoiceChange(AudioClip source, Voice target)
{
    string key = $"{source.name}_{target}";
    if (!voiceCache.ContainsKey(key))
    {
        voiceCache[key] = await source
            .GENVoiceChange()
            .SetTargetVoice(target)
            .ExecuteAsync();
    }
    return voiceCache[key];
}

// ✅ Good - parallel processing
var tasks = audioClips.Select(clip =>
    clip.GENVoiceChange().SetTargetVoice(voice).ExecuteAsync()
);
await UniTask.WhenAll(tasks);

Limitations

1. Speech Only: Works best with clear speech, not music or effects

2. Quality Dependent: Input quality affects output quality

3. Timing Preserved: Cannot change speech speed significantly

4. Language: Best results with same language as target voice

Next Steps

• Audio Isolation - Clean up audio

• Text to Speech - Generate speech

• Speech to Text - Transcribe audio