Task-Vector Arithmetic for Emotional Expressivity Control in Language-Model-Based Text-to-Speech

Abstract

We investigate whether task-vector arithmetic, successful for cross-speaker emotional intensity control in modular text-to-speech (TTS), transfers to large-scale TTS systems built on language-model backbones with in-context learning (LM-TTS). Through a systematic elimination study over four progressively narrower operands on Qwen3-TTS-12Hz-1.7B - model weights via LoRA fine-tuning, continuous codec embeddings, discrete codec tokens, and the speaker embedding (x-vector) produced by an ECAPA-TDNN encoder jointly trained with the synthesis backbone - we localize the dominant carrier of emotional prosody to the x-vector. Building on this finding, we propose a training-free method based on centroid arithmetic in x-vector space: an emotion direction $\ddot{I}„={\mathbb{E}}_i[x(s_i,\text{emo})]-{\mathbb{E}}_i[x(s_i,\text{neutral})]$ applied to an unseen target speaker as $x_{\text{new}}=x(\text{target},\text{neutral})+\hat{I}\pm \cdot \ddot{I}„$. Using ESD (English) as the $\ddot{I}„$ source and emoUERJ (Brazilian Portuguese) as a cross-lingual ground-truth target, we observe average gains of $+0.29$ in emotion2vec cosine over the ICL baseline on English held-out speakers and $+0.09$ on Brazilian Portuguese held-out speakers, while largely preserving identity (WavLM SECS $\gtrsim 0.88$ for the multi-speaker $\ddot{I}„$ variant) and intelligibility (WER $\approx 0$ in PT-BR). These results offer initial evidence that the dominant carrier of emotional prosody in this class of models is localizable, by elimination, to the co-trained speaker embedding, where training-free centroid arithmetic remains effective even under cross-lingual transfer.

v2: expanded related work