- Rating
- 1
- Confidence
- 4
- Informativeness
- 1
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
This paper proposes a training-free method for controlling pronunciation in frozen Chinese TTS models through input embeddings. It uses phonological features and donor-character embeddings to construct embeddings for target pronunciations. Experiments on held-out phonological combinations, rare-word correction, and human evaluation show promising results, although the benefits over simpler regression are less consistent.
Critical strengths
-
The phonological withholding experiments and donor-only controls provide convincing evidence that the proposed embedding transformation can steer pronunciation beyond simple homophone lookup.
-
The paper evaluates real pronunciation errors and includes human listening studies. On 193 verified errors, composition repairs 77%, compared with 65% for regression alone.
Critical weaknesses
-
Format is wrong. Lack of novelty.
-
The advantage over regression remains unclear. The two methods are not independently optimized, and regression even outperforms composition in one longer-context setting (98% vs. 95%). The improvement is also statistically inconclusive in the no-common-homophone human evaluation (36/45 vs. 32/45).
-
Most sentence-level preservation experiments use homophone averaging rather than the proposed donor-based composition. Given that composition breaks 26% of previously correct words in the complete-withholding setting, its impact on surrounding speech and overall naturalness needs more evaluation.
Questions for authors
-
Does composition still outperform regression when both methods are independently tuned with the same budget?
-
Why are five composition outputs missing from Table 10? What would the results look like on all 50 words, including the regression fallback?
-
Could the authors compare composition, regression, and the original model in matched natural contexts, reporting pronunciation accuracy, collateral errors, and naturalness?