piper-tts-training

Piper TTS Voice Training

Train custom text-to-speech voices compatible with Piper's lightweight ONNX runtime.

Overview

Piper produces fast, offline TTS suitable for embedded devices. Training involves:

1. Corpus preparation (text covering phonetic range)
2. Audio generation or recording
3. Quality validation via Whisper transcription
4. Fine-tuning from existing checkpoint (recommended) or training from scratch
5. ONNX export for deployment

Fine-tuning vs from-scratch:

• Fine-tuning: ~1,300 phrases + 1,000 epochs (days on modest GPU)
• From scratch: ~13,000+ phrases + 2,000+ epochs (weeks/months)

Workflow

1. Corpus Preparation

Gather 1,300-1,500+ phrases covering broad phonetic range:

• Use piper-recording-studio corpus as base
• Add domain-specific phrases for your use case
• Include varied sentence structures and lengths

Critical for non-US English: Ensure corpus uses correct regional spelling. See Localisation.

2. Audio Generation

Generate or record training audio at 22050Hz mono WAV.

If using voice cloning (e.g., Chatterbox TTS):

• Generate at source sample rate (often 24kHz)
• Convert to 22050Hz: sox -v 0.95 input.wav -r 22050 -t wav output.wav
• The -v 0.95 prevents clipping during resampling

Recording requirements:

• Consistent microphone position and room acoustics
• Minimal background noise
• Natural speaking pace (not reading voice)

3. Quality Validation with Whisper

Automate quality checks rather than manual listening:

import whisper
from piper_phonemize import phonemize_text

model = whisper.load_model("base")

def validate_sample(audio_path, expected_text):
    result = model.transcribe(audio_path)
    transcribed = result["text"].strip()

    # Compare phonemically to handle spelling/punctuation differences
    expected_phonemes = phonemize_text(expected_text, "en-gb")
    transcribed_phonemes = phonemize_text(transcribed, "en-gb")

    return expected_phonemes == transcribed_phonemes