Implementation:Speechbrain Speechbrain Prepare Voicebank Revb
| Knowledge Sources | |
|---|---|
| Domains | Dereverberation, Data_Preparation |
| Last Updated | 2026-02-09 00:00 GMT |
Overview
Concrete tool for preparing the Voicebank-SLR reverberant dataset for dereverberation tasks provided by the SpeechBrain library.
Description
This script prepares JSON manifest files for the Voicebank-SLR (VB-SLR) reverberant dataset, designed for unsupervised speech dereverberation experiments. It processes clean and reverberated speech pairs, splits speakers into train/valid/test sets, and creates JSON manifests at 16kHz sample rate. The script requires a manual download of the VB-SLR dataset from the CITI Sinica repository and pairs each reverberated utterance with its clean reference.
Usage
Use this script when preparing data for speech dereverberation experiments using the MetricGAN-U or similar unsupervised enhancement methods on the Voicebank-SLR reverberant corpus.
Code Reference
Source Location
- Repository: SpeechBrain
- File: recipes/Voicebank/dereverb/MetricGAN-U/voicebank_revb_prepare.py
Signature
def prepare_voicebank(
data_folder, save_folder, valid_speaker_count=2, skip_prep=False
):
Import
from voicebank_revb_prepare import prepare_voicebank
I/O Contract
Inputs
| Name | Type | Required | Description |
|---|---|---|---|
| data_folder | str | Yes | Path to the folder where the Voicebank-SLR reverberant dataset is stored |
| save_folder | str | Yes | The directory where to store the JSON files |
| valid_speaker_count | int | No | Number of validation speakers out of 28 in the train set (default: 2) |
| skip_prep | bool | No | If True, skips data preparation (default: False) |
Outputs
| Name | Type | Description |
|---|---|---|
| train_revb.json | JSON file | Training manifest with clean and reverberated audio paths |
| valid_revb.json | JSON file | Validation manifest |
| test_revb.json | JSON file | Test manifest |
Usage Examples
from voicebank_revb_prepare import prepare_voicebank
prepare_voicebank(
data_folder="/path/to/datasets/Voicebank-SLR",
save_folder="exp/Voicebank_revb_exp",
)