Text-to-speech (TTS)

This page describes how to use sherpa-onnx for text-to-speech (TTS).

Please first follow Installation and/or Install the Python Package to install sherpa-onnx before you continue.

• All-in-one
• Huggingface space
• Pre-trained models
  • RTF of pre-trained models
  • Matcha
    • matcha-icefall-en_US-ljspeech (American English, 1 female speaker)
      • Download the model
      • Generate speech with executables compiled from C++
      • Generate speech with Python script
      • RTF on Raspberry Pi 4 Model B Rev 1.5
    • matcha-icefall-zh-baker (Chinese, 1 female speaker)
      • Download the model
      • Generate speech with executables compiled from C++
      • Generate speech with Python script
      • RTF on Raspberry Pi 4 Model B Rev 1.5
  • Kokoro
    • kokoro-multi-lang-v1_1 (Chinese + English, 103 speakers)
    • kokoro-multi-lang-v1_0 (Chinese + English, 53 speakers)
      • Download the model
      • Map between speaker ID and speaker name
      • Generate speech with executables compiled from C++
        • Audio samples
      • Generate speech with Python script
      • RTF on Raspberry Pi 4 Model B Rev 1.5
    • kokoro-en-v0_19 (English, 11 speakers)
      • Download the model
      • Map between speaker ID and speaker name
      • Generate speech with executables compiled from C++
      • Generate speech with Python script
      • RTF on Raspberry Pi 4 Model B Rev 1.5
  • vits
    • All models in a single table
    • vits-melo-tts-zh_en (Chinese + English, 1 speaker)
      • Download the model
      • Generate speech with executables compiled from C++
      • Generate speech with Python script
      • RTF on Raspberry Pi 4 Model B Rev 1.5
    • vits-piper-en_US-glados (English, 1 speaker)
      • Download the model
      • Generate speech with executables compiled from C++
      • Generate speech with Python script
      • RTF on Raspberry Pi 4 Model B Rev 1.5
    • vits-piper-en_US-libritts_r-medium (English, 904 speakers)
      • Download the model
      • Generate speech with executables compiled from C++
      • Generate speech with Python script
      • RTF on Raspberry Pi 4 Model B Rev 1.5
    • ljspeech (English, single-speaker)
      • Download the model
      • Generate speech with executables compiled from C++
      • Generate speech with Python script
      • RTF on Raspberry Pi 4 Model B Rev 1.5
    • VCTK (English, multi-speaker, 109 speakers)
      • Download the model
      • Generate speech with executables compiled from C++
      • Generate speech with Python script
      • RTF on Raspberry Pi 4 Model B Rev 1.5
    • csukuangfj/sherpa-onnx-vits-zh-ll (Chinese, 5 speakers)
      • RTF on Raspberry Pi 4 Model B Rev 1.5
    • csukuangfj/vits-zh-hf-fanchen-C (Chinese, 187 speakers)
      • RTF on Raspberry Pi 4 Model B Rev 1.5
    • csukuangfj/vits-zh-hf-fanchen-wnj (Chinese, 1 male)
      • RTF on Raspberry Pi 4 Model B Rev 1.5
    • csukuangfj/vits-zh-hf-theresa (Chinese, 804 speakers)
      • RTF on Raspberry Pi 4 Model B Rev 1.5
    • csukuangfj/vits-zh-hf-eula (Chinese, 804 speakers)
      • RTF on Raspberry Pi 4 Model B Rev 1.5
    • aishell3 (Chinese, multi-speaker, 174 speakers)
      • Download the model
      • Generate speech with executables compiled from C++
      • Generate speech with Python script
      • RTF on Raspberry Pi 4 Model B Rev 1.5
    • en_US-lessac-medium (English, single-speaker)
      • Download the model
      • Generate speech with executables compiled from C++
      • Generate speech with Python script
      • RTF on Raspberry Pi 4 Model B Rev 1.5
• WebAssembly
  • Install Emscripten
  • Build
  • Huggingface Spaces (WebAssembly)
    • English TTS
    • German TTS
• Piper
  • Install dependencies
  • Find the pre-trained model from piper
  • Download the pre-trained model
  • Add meta data to the onnx model
  • Download espeak-ng-data
  • Test your converted model
• MMS
  • Install dependencies
  • Download the model file
  • Download MMS source code
  • Convert the model
  • Use the converted model
• Frequently Asked Question (FAQs)
  • Is there a colab notebook
  • How to enable UTF-8 on Windows
  • How to install sherpa-onnx for TTS
    • For Python users
    • Build from source
  • Where to get pre-trained TTS models
  • How to handle OOVs