Implementation:Speechbrain Speechbrain Prepare Fisher Callhome

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

Overview

Concrete tool for preparing the Fisher-Callhome-Spanish dataset for speech translation provided by the SpeechBrain library.

Description

This script prepares JSON manifest files for the Fisher-Callhome-Spanish dataset, which contains Spanish conversational telephone speech with English translations. It processes LDC-distributed speech and transcription data, resamples audio to 16 kHz, segments utterances by channel and timestamp, and normalizes transcripts using Moses punctuation normalization and tokenization. The output is suitable for training Spanish-to-English speech translation systems.

Usage

Use this when preparing the Fisher-Callhome-Spanish dataset for speech translation training with SpeechBrain recipes.

Code Reference

Source Location

• Repository: SpeechBrain
• File: recipes/Fisher-Callhome-Spanish/fisher_callhome_prepare.py

Signature

def prepare_fisher_callhome_spanish(
    data_folder: str, save_folder: str, device: str = "cpu"
):

Import

from recipes.Fisher_Callhome_Spanish.fisher_callhome_prepare import prepare_fisher_callhome_spanish

I/O Contract

Inputs

Name	Type	Required	Description
data_folder	str	Yes	Path to the folder where the Fisher-Callhome-Spanish dataset is stored
save_folder	str	Yes	Path where train/valid/test specification files will be saved
device	str	No	Device for computation, e.g. "cpu" or "cuda" (default: "cpu")

Outputs

Name	Type	Description
train.json	JSON	Training split manifest with audio paths, Spanish transcripts, and English translations
valid.json	JSON	Validation split manifest
test.json	JSON	Test split manifest

Usage Examples

from recipes.Fisher_Callhome_Spanish.fisher_callhome_prepare import prepare_fisher_callhome_spanish

prepare_fisher_callhome_spanish(
    data_folder="/path/to/fisher-callhome",
    save_folder="/path/to/output",
)

Related Pages

• Principle:Speechbrain_Speechbrain_Dataset_Specific_Data_Preparation