Implementation:Speechbrain Speechbrain Prepare TIMIT

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

Overview

Concrete tool for preparing the TIMIT dataset (LDC93S1) for phoneme recognition and ASR tasks provided by the SpeechBrain library.

Description

This script prepares JSON manifest files for the TIMIT dataset, a standard benchmark for phoneme recognition. It processes the TIMIT directory structure to extract audio paths and phoneme-level transcriptions, supporting three phoneme set sizes (60, 48, or 39 phonemes). The script handles both upper-case and lower-case versions of the TIMIT corpus and produces train, validation, and test JSON manifests at a 16kHz sample rate.

Usage

Use this script when preparing data for phoneme recognition or ASR experiments on the TIMIT corpus. It must be executed before running any TIMIT-based training recipe in SpeechBrain.

Code Reference

Source Location

• Repository: SpeechBrain
• File: recipes/TIMIT/timit_prepare.py

Signature

def prepare_timit(
    data_folder,
    save_json_train,
    save_json_valid,
    save_json_test,
    phn_set=39,
    uppercase=False,
    skip_prep=False,
):

Import

from timit_prepare import prepare_timit

I/O Contract

Inputs

Name	Type	Required	Description
data_folder	str	Yes	Path to the folder where the original TIMIT dataset is stored
save_json_train	str	Yes	The path where the training JSON file will be stored
save_json_valid	str	Yes	The path where the validation JSON file will be stored
save_json_test	str	Yes	The path where the test JSON file will be stored
phn_set	int	No	The phoneme set to use: 60, 48, or 39 phonemes (default: 39)
uppercase	bool	No	Set to True when TIMIT dataset is in upper-case version (default: False)
skip_prep	bool	No	If True, skips data preparation (default: False)

Outputs

Name	Type	Description
train.json	JSON file	Training set manifest with audio paths and phoneme transcriptions
valid.json	JSON file	Validation set manifest
test.json	JSON file	Test set manifest (core test set)

Usage Examples

from timit_prepare import prepare_timit

prepare_timit(
    data_folder="/path/to/TIMIT",
    save_json_train="output/train.json",
    save_json_valid="output/valid.json",
    save_json_test="output/test.json",
    phn_set=39,
)

Related Pages

• Principle:Speechbrain_Speechbrain_Dataset_Specific_Data_Preparation