Sign in to view source links and access this dataset
Description
ExpressiveSpeech is a high-quality, expressive, and bilingual (Chinese-English) speech dataset created by FreedomIntelligence. It was curated from five open-source emotional dialogue datasets: Expresso, NCSSD, M3ED, MultiDialog, and IEMOCAP. The dataset was last updated on 2025-10-24.
Use Cases
Train expressive text-to-speech models based on the high-quality, emotional speech recordings.
Develop cross-lingual voice synthesis systems based on the bilingual (Chinese-English) nature of the data.
Benchmark emotional speech generation models based on the curated selection from established emotional dialogue datasets.
Fine-tune speech models for consistent vocal expressiveness in dialogue systems.
Strengths
Dataset is described as 'high-quality' and 'expressive'.
Source data is drawn from five established open-source emotional dialogue datasets.
Dataset is bilingual, containing both Chinese and English speech.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count is unknown, which may limit suitability assessment.
The description references a 'rigorous processing and selection pipeline' but specifics are not provided in the input.
Provenance
Source
FreedomIntelligence
Collection Method
Curated from five open-source emotional dialogue datasets: Expresso, NCSSD, M3ED, MultiDialog, and IEMOCAP.
Freshness
Last updated 2025-10-24 10:18:55.
License is unknown; users must verify terms of use before downloading.