Implementation:Speechbrain Speechbrain Prepare ESC50 Root

Knowledge Sources	SpeechBrain
Domains	Audio_Classification, Data_Preparation
Last Updated	2026-02-09 00:00 GMT

Overview

Concrete tool for preparing the ESC-50 dataset for environmental sound classification provided by the SpeechBrain library.

Description

This script creates JSON data manifest files for the ESC-50 (Environmental Sound Classification) dataset. It automatically downloads the dataset if not found locally, converts metadata into SpeechBrain-compatible JSON format, and splits data by the predefined 5-fold cross-validation scheme. The ESC-50 dataset contains 2,000 environmental audio recordings across 50 classes.

Usage

Use this when preparing the ESC-50 dataset for environmental sound classification training with SpeechBrain recipes.

Code Reference

Source Location

• Repository: SpeechBrain
• File: recipes/ESC50/esc50_prepare.py

Signature

def prepare_esc50(
    data_folder,
    audio_data_folder,
    save_json_train,
    save_json_valid,
    save_json_test,
    train_fold_nums=[1, 2, 3],
    valid_fold_nums=[4],
    test_fold_nums=[5],
    skip_manifest_creation=False,
):

Import

from recipes.ESC50.esc50_prepare import prepare_esc50

I/O Contract

Inputs

Name	Type	Required	Description
data_folder	str	Yes	Path to the folder where the ESC50 dataset (including metadata) is stored
audio_data_folder	str	Yes	Path to the folder where the ESC50 audio files are stored
save_json_train	str	Yes	Path where the train data specification file will be saved
save_json_valid	str	Yes	Path where the validation data specification file will be saved
save_json_test	str	Yes	Path where the test data specification file will be saved
train_fold_nums	list	No	Pre-defined folds to use for training (default: [1, 2, 3])
valid_fold_nums	list	No	Pre-defined folds to use for validation (default: [4])
test_fold_nums	list	No	Pre-defined folds to use for test (default: [5])
skip_manifest_creation	bool	No	Whether to skip the manifest creation step (default: False)

Outputs

Name	Type	Description
train.json	JSON	Training split manifest with audio paths and labels
valid.json	JSON	Validation split manifest with audio paths and labels
test.json	JSON	Test split manifest with audio paths and labels

Usage Examples

from recipes.ESC50.esc50_prepare import prepare_esc50

prepare_esc50(
    data_folder="/path/to/ESC-50",
    audio_data_folder="/path/to/ESC-50/audio",
    save_json_train="train.json",
    save_json_valid="valid.json",
    save_json_test="test.json",
    train_fold_nums=[1, 2, 3],
    valid_fold_nums=[4],
    test_fold_nums=[5],
)

Related Pages

• Principle:Speechbrain_Speechbrain_Dataset_Specific_Data_Preparation