System,Training Data,🌐 Multilingual,📚 Training Amount (k hours),🧠 Num. Parameters (M),🎯 Target Repr.,📖 LibriVox Only,🔄 NAR,🔁 AR,🔡 G2P,🧩 Language Model,🎵 Prosody Prediction,🌫 Diffusion,⏱️ Delay Pattern
Bark,Unknown,✅,Unknown,240,Audio Codec Code,❌,✅,✅,❌,✅,❌,❌,❌
E2-TTS,Emilia,✅,95,330,Mel Spectrogram,❌,✅,❌,✅,❌,✅,❌,❌
F5-TTS,Emilia,✅,95,330,Mel Spectrogram,❌,✅,❌,✅,❌,❌,✅,❌
Fish (1.4),"LibriLight, PlayerFM, StarRail...?",✅,700,500,Audio Codec Code,❌,❌,✅,❌,✅,❌,❌,❌
GPT-SoVITS,"Chinese, English, Japanese (1000+700+300 hrs)",✅,2,200,Audio Codec Code,❌,✅,✅,✅,✅,❌,❌,❌
Hierspeech++,"LibriTTS, LibriLight, Expresso, MSSS, NIKL",✅,2.7,97,Waveform,❌,✅,❌,✅,❌,✅,❌,❌
MetaVoice,Unknown,❌,100,1000,Audio Codec Code,❌,✅,✅,❌,✅,❌,❌,✅
NaturalSpeech2,LibriTTS,❌,0.58,380,Audio Codec Code,✅,✅,❌,✅,❌,✅,✅,❌
OpenVoice,Unknown,❌,0.6,73,Mel Spectrogram,❌,✅,❌,✅,❌,✅,❌,❌
ParlerTTS,"MLS, LibriTTS",❌,23,2200,Audio Codec Code,✅,❌,✅,❌,✅,❌,❌,✅
Pheme,"GigaSpeech, MLS, LibriTTS",❌,11,300,Audio Codec Code,❌,✅,✅,✅,✅,❌,❌,❌
SpeechT5,"LibriTTS, LibriSpeech",❌,1.4,144,Mel Spectrogram,✅,❌,✅,❌,❌,❌,❌,❌
StyleTTS2,LibriTTS,❌,0.24,191,Mel Spectrogram,✅,✅,❌,✅,❌,✅,✅,❌
TorToiSe,"LibriTTS, HifiTTS, Podcasts + Audiobooks",❌,49.5,960,Mel Spectrogram,❌,✅,✅,❌,✅,❌,✅,❌
VallEv1,LibriLight,❌,6,370,Audio Codec Code,✅,❌,✅,✅,✅,❌,❌,❌
VoiceCraft,GigaSpeech,❌,9,830,Audio Codec Code,❌,❌,✅,✅,✅,❌,❌,✅
WhisperSpeech,MLS,✅,80,2054,Audio Codec Code,✅,✅,✅,❌,✅,❌,❌,❌
XTTSv2,"LibriTTS, Common Voice, Internal",✅,27,456,Audio Codec Code (Own),❌,❌,✅,❌,✅,❌,❌,❌
xVASynth,Unknown,✅,Unknown, , , , , , , , , ,