CEAEval-D: Mandarin Speech Expressive Appropriateness in Rich Contexts

CEAEval-D is a Mandarin speech dataset annotated for context-rich expressive appropriateness. It was released by TianRW in association with an ACL paper and is hosted on Hugging Face. The dataset was last updated on May 10, 2026.

Use Cases

• Train models to evaluate expressive appropriateness based on long-range narrative context.
• Benchmark speech synthesis systems on context-aware expressive performance.
• Research the relationship between narrative context and perceived speech appropriateness.
• Develop evaluation metrics for expressive speech beyond isolated emotion or naturalness.

Strengths

• Focuses on context-rich expressive appropriateness, a novel annotation target for Mandarin speech.
• Associated with a peer-reviewed ACL paper, suggesting academic validation.
• Last updated on May 10, 2026, indicating recent maintenance.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

TianRW

Collection Method

Likely contains human-annotated speech utterances grounded in narrative contexts.

Freshness

Last updated 2026-05-10 11:57:43