Advanced Vocal Editing
PronunciationPitchVibratoBreathingFalsettoTensionStrengthEmotion
AI Singing Voice Generation
Conversational Speech Generation
Dataset Size
1 million hours
Genre Variety
PopSoulLatinoCinematicOperaChild VoiceHip hopBalladR&BLatin PopR&B/FunkSoul/FunkLatin Folk
Highly Editable AI Vocals
Language/Country Support
EnglishSpanishChineseJapanese
Model Sizes
Tiny: 1B backbone, 100M decoderSmall: 3B backbone, 250M decoderMedium: 8B backbone, 300M decoder
Multiple Speaker Handling
No Custom Model Training for VoiceMix
Objective Metrics
Word Error RateSpeaker SimilarityHomograph DisambiguationPronunciation Consistency
Partial Multilingual Support Planned
Planned for 20+ languages
Royalty-Free Commercial Use
Subjective Metrics
Comparative Mean Opinion Score
Voice Designer / VoiceMix