Sign in to view source links and access this dataset
Description
Synthetic vocal burst audio samples generated from a taxonomy of 202 vocal burst types across multiple text-to-audio and TTS models. The dataset was created by the organization 'laion' and was last updated on June 2, 2026. Each sample is a short (3–10 second) non-speech vocalization such as laughs, cries, gasps, sighs, and growls.
Use Cases
Training or benchmarking text-to-audio models based on the described taxonomy of 202 vocal burst types.
Researching emotional expression and non-verbal communication based on synthetic laughs, cries, and gasps.
Developing or evaluating models for generating vocal bursts with specific attributes like gender and age group.
Creating synthetic datasets for audio classification tasks involving non-speech human sounds.
Strengths
Includes a taxonomy of 202 distinct vocal burst types.
Samples are generated from multiple text-to-audio and TTS models, including DramaBox which contributed 2000 samples.
Audio samples have a high sample rate of 44.1 kHz.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count is unknown, which may limit suitability assessment.
Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
Source
huggingface
Collection Method
Synthetically generated from text prompts using multiple text-to-audio and TTS models.
Freshness
Last updated 2026-06-02 06:03:24; freshness should be verified.
License is unknown; terms of use must be verified before application.