v1.2.4 - Automatic Text Chunking

[Enemyx-net]

📝 Automatic Text Chunking

• Long Text Support: Single Speaker node now automatically handles texts of any length
• Audio Quality Fix: Prevents audio acceleration issues that occurred with long texts
• Configurable Chunking: New max_words_per_chunk parameter for fine-tuning
• Seamless Output: Automatically concatenates chunks for smooth, natural speech

✨ What's New

Smart Text Processing

• Automatic Splitting: Texts longer than 250 words are intelligently divided
• Consistent Voice: Uses the same seed across all chunks for uniform voice characteristics
• Natural Breaks: Splits at sentence boundaries when possible
• Transparent Process: Works automatically without user intervention

New Parameter

• max_words_per_chunk: Customize chunk size
  • Range: 100-500 words
  • Default: 250 words
  • Lower values: More stable but more processing time
  • Higher values: Faster but may risk acceleration on very long texts

🎯 Benefits

• ✅ No More Audio Acceleration: Fixes the common issue of sped-up audio on long texts
• ✅ Unlimited Text Length: Process entire documents, books, or scripts
• ✅ Consistent Quality: Maintains voice characteristics throughout
• ✅ Smooth Playback: Seamless audio concatenation with no audible breaks

💡 Use Cases

Perfect for:

• 📖 Audiobook generation
• 📚 Long-form content narration
• 📝 Document reading
• 🎭 Extended monologues

🔧 Usage

The chunking happens automatically. To adjust:

1. Find the max_words_per_chunk parameter in Single Speaker node
2. Adjust based on your needs:
   • Stability Priority: Use 150-200 words
   • Balanced (default): 250 words
   • Speed Priority: Use 300-500 words

💾 Installation

Install via ComfyUI Manager or manually:

git clone https://github.com/Enemyx-net/VibeVoice-ComfyUI

📋 Notes

Multi-Speaker node unaffected (already processes by speaker segments)
Chunking is transparent - you'll receive a single audio output
Processing time scales linearly with text length
Voice consistency is maintained using deterministic seeding

---

This discussion was created from the release v1.2.4 - Automatic Text Chunking.