---
pretty_name: SoundBible Sound Effects
size_categories:
- 1K<n<10K
task_categories:
- audio-classification
- text-to-audio
- automatic-speech-recognition
annotations_creators:
- found
language:
- en
multilinguality:
- monolingual
source_datasets:
- original
configs:
- config_name: default
  data_files:
  - split: train
    path: files.zip
  default: true
tags:
- audio
license: other
---

# Dataset Card for SoundBible Sound Effects

### Dataset Summary

This dataset contains 2,120 sound effect files from [soundbible.com](https://soundbible.com). It includes a diverse collection of audio clips covering various categories such as action sounds, nature sounds, mechanical sounds, human sounds, and more. The dataset provides both audio files in WAV format and comprehensive metadata including titles, descriptions, tags, licensing information, and author attribution.

### Languages

The dataset is monolingual:
- English (en): Sound titles, descriptions, and tags are in English.

## Dataset Structure

### Data Fields

Each sound effect in the dataset is accompanied by metadata containing the following fields:

- `id`: Unique identifier for the sound effect
- `title`: Descriptive title of the sound effect
- `description`: Detailed description of the sound, including context and potential use cases
- `tags`: Array of strings representing descriptive terms, categories, and related concepts associated with the sound
- `license`: License type under which the sound is distributed
- `author`: Name of the person who created or contributed the sound effect
- `download_url`: Original URL where the sound file was hosted

### Data Splits

The dataset contains a single split:

| Split | Number of Examples |
| :---- | ---------------: |
| `train` | 2,120          |

Total dataset size: 2,120 entries

### Data Files

- Audio files are stored in WAV format within `files.zip`
- Metadata is provided in JSONL format (`soundbible.jsonl.zst`)