Greek text to speech system improves with small curated audiobook data

Deterministic Prompting for Speaker-Stable Low-Resource Greek TTS

SoundComputation and LanguageMachine Learning

Summary

Making computers speak in Greek sounds nearly as good as a human even when there's only a little clean speech to learn from. The authors turned audiobook recordings into good quality training data using special alignment tools. They adapted a large speech model trained on multiple languages to Greek, fixing issues where the voice would change unexpectedly by using consistent prompts and a fine-tuning step that keeps the speaker’s identity. Their system speaks clearly and naturally, close to human voices, using only a few hours of single-speaker recordings.

What this means in practice

Authors

Georgios Syllas, Efthymios Georgiou, Kosmas Kritsis, Alexandros Potamianos

Abstract

Modern TTS systems approach human quality for high-resource languages but degrade when clean speech data is scarce. Modern Greek exemplifies this, lacking the curated corpora behind state-of-the-art synthesis. We propose a data curation recipe that transforms audiobook recordings into TTS-ready data via WhisperX alignment and filtering. Then we fine-tune Parler-TTS (880M), a prompt-based multilingual model whose pre-training encodes phonetic priors transferable to Greek. During development, we find that LLM-generated style prompts introduce speaker drift at inference. Replacing them with deterministic prompts resolves this, and a speaker-specific LoRA stage trained on 3.5 h of single-speaker data anchors identity while updating ~5% of parameters. Our system achieves WER 10.7% (2.9 above the ASR floor), MOS-I 4.00 (vs. 4.36 human speech), and near-human speaker consistency (MOS-C 4.24 vs. 4.30), showing that robust single-speaker Greek TTS is achievable with limited curated data.