Maestro-EVC: Controllable Emotional Voice Conversion Guided by References and Explicit Prosody

TL;DR

Maestro-EVC is a novel emotional voice conversion framework that independently controls speech content, speaker identity, and emotion, effectively modeling temporal emotional dynamics and prosody for high-quality, expressive speech synthesis.

Contribution

It introduces a method to disentangle content, speaker, and emotion attributes from separate references and models temporal emotional dynamics explicitly.

Findings

Achieves high-quality, controllable emotional voice conversion

Effectively disentangles speech attributes for independent control

Models temporal emotional dynamics robustly

Abstract

Emotional voice conversion (EVC) aims to modify the emotional style of speech while preserving its linguistic content. In practical EVC, controllability, the ability to independently control speaker identity and emotional style using distinct references, is crucial. However, existing methods often struggle to fully disentangle these attributes and lack the ability to model fine-grained emotional expressions such as temporal dynamics. We propose Maestro-EVC, a controllable EVC framework that enables independent control of content, speaker identity, and emotion by effectively disentangling each attribute from separate references. We further introduce a temporal emotion representation and an explicit prosody modeling with prosody augmentation to robustly capture and transfer the temporal dynamics of the target emotion, even under prosody-mismatched conditions. Experimental results confirm that Maestro-EVC achieves high-quality, controllable, and emotionally expressive speech synthesis.