LibriTTS-VI: A Public Corpus and Novel Methods for Efficient Voice Impression Control
This work provides a public corpus and effective methods for fine-grained voice impression control in TTS, addressing a key bottleneck for researchers and practitioners.
The authors introduce LibriTTS-VI, the first public corpus for numerical voice impression control in TTS, and propose novel methods to mitigate impression leakage. Their best method reduces VI mean squared error from 0.61 to 0.41 objectively and from 1.15 to 0.92 subjectively.
Numerical voice impression (VI) control (e.g., scaling brightness) enables fine-grained control in text-to-speech (TTS). However, it faces two challenges: no public corpus and impression leakage, where reference audio biases synthesized voice away from the target VI. To address the first challenge, we introduce LibriTTS-VI, the first public VI corpus built on LibriTTS-R. For the second, we hypothesize a single reference causes leakage by entangling speaker identity and VI. To mitigate this, we propose 1) disentangled training with two utterances from the same speaker for speaker and VI conditioning, and 2) a reference-free method controlling the impression solely via target VI. Experimentally, our best method improves controllability: 11-dimensional VI mean squared error drops from 0.61 to 0.41 objectively and 1.15 to 0.92 subjectively. A comparison with a prompt-based TTS reveals imprecise numerical control and entanglement between VI and text semantics, which our methods overcome.