Text-to-Speech (TTS) Overview & Voice Selection Matrix
In DuRT 3.0, Text-to-Speech (TTS) powers automated video dubbing and voiceover creation. Combined with the DAG workflow's textToSpeech step, you can transform transcribed text or multilingual translations into expressive speech tracks.
3 Main TTS Engines Comparison Matrix
| TTS Engine | Architecture | Models | Voice Design (Prompt-based) | Voice Cloning | Voices & Languages | Best For |
|---|---|---|---|---|---|---|
| 🎙️ Xiaomi MiMo | Cloud Neural Voice | mimo-v2.5-ttsmimo-v2.5-tts-voicedesignmimo-v2.5-tts-voiceclone | ✅ Exclusive (Design voice via text prompts) | ✅ Supported (Few-second audio clone) | Hyper-realistic human cadence, rich emotional nuances | Video storytelling, character voiceovers, personal voice clones |
| 🔊 OpenAI TTS | Official Cloud API | tts-1tts-1-hd | ❌ Fixed official voices | ❌ Not supported | 6 standard voices (Alloy, Echo, Fable, Onyx, Nova, Shimmer) | Multilingual global podcasts, educational courses |
| 📦 OneTTS Gateway | Open-Source Gateway | edgettskokorodoubaof5tts | ✅ Depends on engine | ✅ F5-TTS fast clone | Comprehensive multi-language library | Zero-cost bulk generation, private offline workflows |