Dubbing & TTS Channels

Dubbing (TTS) is the third step in video translation. It converts translated subtitle text into speech audio. pyVideoTrans supports 30+ dubbing channels.

You can also use the dubbing feature independently via the left panel: Batch Dub Subtitles. It supports importing multiple SRT subtitle files or TXT files.

The clone voice role uses the speaker's original voice timbre from the video to perform voice cloning. This role is only available in the main interface's Video Translation feature.

All channels that support the clone role also support custom reference audio. This allows you to clone a voice from a 3–10 second audio clip you provide. See Create and Use Reference Audio below for details.

---

Ready to Use (Free)

No complex configuration required—ideal for beginners.

Channel	Description	Rating
Edge-TTS(Free)	Microsoft's free service, natural voices, supports all languages	⭐⭐⭐ Default Recommended
gTTS(Free)	Google TTS, basic quality, requires a proxy in mainland China	⭐⭐

⚠️ Heavy usage of Edge-TTS in a short period may trigger rate limiting. We recommend setting concurrency to 1 and pause duration to 5–10 seconds in Advanced Options.

---

Built-in (Free)

Models will be downloaded automatically upon first use.

Channel	Description	GPU Acceleration	Supports Cloning (clone role)	Rating
Qwen3-TTS(Built-in)	Supports Chinese, English, Japanese, Korean, German, French, Russian, Portuguese, Spanish, and Italian	✅	✅	⭐⭐⭐ Recommended
F5-TTS(Built-in)	Chinese, English, Japanese, French, German, Russian, Italian, Spanish, Hindi, Arabic	✅	✅	⭐⭐⭐
OmniVoice-TTS	Supports 600+ languages (Built-in from v4.05)	✅	✅	⭐⭐⭐ Recommended
Confucius-TTS(Built-in)from v4.06	Chinese, English, Japanese, Korean, German, French, Spanish, Indonesian, Italian, Thai, Portuguese, Russian, Malay, Vietnamese	✅	✅	⭐⭐⭐
MOSS-TTS-Nano(Built-in)	Chinese, English, German, Spanish, French, Japanese, Italian, Hungarian, Korean, Russian, Persian, Arabic, Polish, Portuguese, Czech, Swedish, Greek, Turkish	❌	✅	⭐⭐
ZipVoice(Built-in)	Chinese and English	✅	✅	⭐⭐⭐ Recommended
Piper(Built-in)	Lightweight, supports 20 languages	❌	❌	⭐⭐
ChatterBox(Built-in)	Arabic, German, Greek, English, Spanish, Finnish, French, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Dutch, Norwegian, Polish, Portuguese, Russian, Swedish, Swahili, Turkish, Chinese	✅	✅	⭐⭐⭐ Recommended
Supertonic3(Built-in)	English, Korean, Japanese, Arabic, Czech, German, Greek, Spanish, French, Hindi, Hungarian, Indonesian, Italian, Dutch, Polish, Portuguese, Romanian, Russian, Swedish, Turkish, Ukrainian, Vietnamese	❌	❌	⭐⭐
VITS(Built-in)	Chinese and English dubbing	❌	❌	⭐⭐
Higgs-audio-v3(Built-in)	All built-in languages; CPU-only requires 20GB RAM, GPU acceleration requires >10GB VRAM	✅	✅	⭐⭐

Due to network conditions and large model sizes, automatic downloads might occasionally fail. If a download fails, click here to view model download links and manual setup instructions.

---

Self-Hosted / Local Deployment (Advanced)

Channel	Description	Supports Cloning (clone role)	Rating
GPT-SoVITS	Supports Chinese, English, Japanese, Korean, Cantonese	✅	⭐⭐⭐ Recommended
Index-TTS	Chinese, English	✅	⭐⭐⭐ Recommended
VoxCPM-TTS	Arabic, Burmese, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Indonesian, Italian, Japanese, Khmer, Korean, Lao, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, Tagalog, Thai, Turkish, Vietnamese	✅	⭐⭐⭐
CosyVoice	Chinese, English, Japanese, Korean, German, Spanish, French, Italian, Russian	✅	⭐⭐
ChatTTS	Supports Chinese and English	—	⭐⭐
Fish-TTS	Supports all built-in languages	—	⭐
Kokoro-TTS	Chinese, English, Korean, Italian, Portuguese, German, French, Hindi	—	⭐
Spark-TTS	Chinese, English	✅	⭐⭐
clone-voice	No longer maintained	✅	⭐

Create and Use Reference Audio

Any dubbing channel that supports voice cloning (where the role list includes clone) can use reference audio. Once configured, simply select your reference audio as the voice role, and the software will automatically dub the audio using that voice timbre. Reference audio is configured under Menu → TTS Settings → Reference Audio Settings.

Step-by-Step Guide

0. Record or trim a 3–10s audio clip from existing audio. Save it as a .wav file, making sure the pronunciation is clear and free of background noise.
1. Open the Menu → TTS Settings → Reference Audio Settings window.
2. Enter your reference audio info in the text box using the following format:

audio_filename#matching_text_content

3. Place the reference audio file inside the pyVideoTrans_folder/f5-tts folder (create this folder manually if it doesn't exist).

Example

Suppose you have an audio file named nverguo.wav, and the spoken words in that audio are "The text content corresponding to the audio". You would enter:

nverguo.wav#The text content corresponding to the audio

Reference Audio Requirements

Item	Requirement
Format	WAV format (recommended); MP3 and other common formats are also supported
Duration	3–10 seconds
Audio Quality	Clear speech, no background noise
Text	Must match the spoken words in the audio exactly

---

Professional Cloud Services (API Key Required)

Channel	Description	Rating
Azure TTS	Microsoft's enterprise-grade voice service	⭐⭐⭐
OpenAI TTS	Industry-leading audio technology	⭐⭐⭐
ByteDance TTS 2.0	Authentic Chinese pronunciation	⭐⭐⭐
Alibaba Qwen-TTS	Alibaba Cloud text-to-speech	⭐⭐⭐
Gemini TTS	Google TTS	⭐⭐
Elevenlabs.io	High-end AI voice generation	⭐⭐⭐
302.AI	API aggregation platform	⭐⭐
Minimaxi	MiniMax TTS (requires paid balance)	⭐⭐
Xiaomi TTS	Xiaomi AI Open Platform	⭐⭐
X.AI TTS	x.ai platform	⭐⭐

---