MED-TTS: Multilingual Emotional Speech Dataset with Segment-Level Annotations

MED-TTS focuses on dynamic emotion flow within utterances through segment-level annotations, unlike conventional datasets with single static labels. The dataset contains Chinese and English text samples annotated with utterance-level emotion and was created by Chanson-0803. It was last updated on 2026-05-17.

Use Cases

• Train expressive speech synthesis models based on segment-level emotion annotations.
• Research dynamic emotion transitions in speech based on the annotated emotion flow.
• Benchmark multimodal generation systems based on the multilingual emotional text samples.
• Develop emotion-aware dialogue systems based on the annotated emotional speech data.

Strengths

• Annotates dynamic emotion flow at the segment level, a distinctive feature mentioned in the description.
• Contains multilingual samples in both Chinese and English.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Freshness should be verified as the last update date is in the future (2026-05-17).

Provenance

Source

huggingface

Freshness

Last updated 2026-05-17 14:23:17