๐Ÿ—ฃ๏ธ Indonesian Text-to-Speech with Voice Conversion

Convert Indonesian text to speech using fine-tuned CSM-1B (GPU) or CPU fallback.

๐ŸŽค Voice Cloning: Upload a reference voice to clone its characteristics (gender, tone, pitch)!

โš™๏ธ System Settings

If GPU not available, run model on CPU

๐ŸŽค Voice Cloning Settings

Clone the voice characteristics from reference audio

โš™๏ธ Generation Settings

0 10
50 400
0.1 1
0.1 2

๐Ÿ“ How to use Voice Cloning:

  1. Upload reference audio: Use a clear 3-10 second sample
  2. Check "Enable Voice Cloning": This will apply the conversion
  3. Gender conversion: Works automatically (male โ†” female)
  4. Best results: Use clean audio with minimal background noise

โœจ What gets cloned:

  • ๐ŸŽต Pitch (voice height)
  • ๐ŸŽญ Gender (formants)
  • ๐ŸŽจ Timbre (voice color/character)
  • ๐Ÿ”Š Energy (speaking volume/intensity)