Name "Processor ⚑" "Phonetic alphabet πŸ”€" "Insta-clone πŸ‘₯" "Emotion control 🎭" "Prompting πŸ“–" "Speech control 🎚" "Streaming support 🌊" "Voice conversion 🦜" "Longform synthesis" Amphion CUDA πŸ‘₯ 🎭πŸ‘₯ ❌ Bark CUDA ❌ 🎭 tags ❌ E2/F5 TTS πŸ‘₯ ❌ πŸ“– EmotiVoice Glow-TTS GPT-SoVITS HierSpeech++ ❌ πŸ‘₯ 🎭πŸ‘₯ ❌ "speed / stability 🎚" 🦜 IMS-Toucan CUDA ❌ ❌ ❌ ❌ MahaTTS Matcha-TTS IPA ❌ ❌ ❌ "speed / stability 🎚" MeloTTS CPU / CUDA ❌ ❌ "speed 🎚" MetaVoice-1B CUDA πŸ‘₯ 🎭πŸ‘₯ ❌ "stability / similarity 🎚" Yes Neural-HMM TTS OpenVoice CUDA ❌ πŸ‘₯ "6-type 🎭 πŸ˜‘πŸ˜ƒπŸ˜­πŸ˜―πŸ€«πŸ˜Š" ❌ OverFlow TTS Parler CUDA ❌ πŸŽ­πŸ“– πŸ“– ❌ 🌊 pflowTTS Pheme CUDA ❌ πŸ‘₯ 🎭πŸ‘₯ ❌ "stability 🎚" Piper RAD-TTS Silero StyleTTS 2 CPU / CUDA IPA πŸ‘₯ 🎭πŸ‘₯ ❌ 🌊 Yes Tacotron 2 TorToiSe TTS ❌ ❌ ❌ πŸ“– 🌊 TTTS CPU / CUDA ❌ πŸ‘₯ VALL-E VITS/ MMS-TTS CUDA ❌ ❌ ❌ ❌ "speed 🎚" WhisperSpeech CUDA ❌ πŸ‘₯ 🎭πŸ‘₯ ❌ "speed 🎚" XTTS CUDA ❌ πŸ‘₯ 🎭πŸ‘₯ ❌ "speed / stability 🎚" 🌊 ❌ xVASynth CPU 1️⃣ / CUDA ARPAbet+ ❌ "4-type 🎭 πŸ˜‘πŸ˜ƒπŸ˜­πŸ˜― per‑phoneme" ❌ "speed / pitch / energy / 🎭 🎚 per‑phoneme" ❌ 🦜 ~15 secs max