Hierarchical Decoding for Discrete Speech Synthesis with Multi-Resolution Spoof Detection
For neural codec language model users, this provides a lightweight inference-time method to reduce artifacts and distributional drift in speech synthesis.
MSpoof-TTS improves zero-shot discrete speech synthesis by using a training-free, multi-resolution spoof detection framework to prune and re-rank codec sequences, enhancing perceptual realism without retraining.
Neural codec language models enable high-quality discrete speech synthesis, yet their inference remains vulnerable to token-level artifacts and distributional drift that degrade perceptual realism. Rather than relying on preference optimization or retraining, we propose MSpoof-TTS, a training-free inference framework that improves zero-shot synthesis through multi-resolution spoof guidance. We introduce a Multi-Resolution Token-based Spoof Detection framework that evaluates codec sequences at different temporal granularities to detect locally inconsistent or unnatural patterns. We then integrate the spoof detectors into a hierarchical decoding strategy, progressively pruning low-quality candidates and re-ranking hypotheses. This discriminator-guided generation enhances robustness without modifying model parameters. Experiments validate the effectiveness of our framework for robust and high-quality codec-based speech generation. Audio samples are available at https://danny-nus.github.io/MSpoofTTS.github.io/.