# VQ-Bench Dataset

The VQ-Bench dataset is hosted on Zenodo for long-term preservation and to provide a citable DOI.

## Download

Visit the Zenodo repository to download the dataset:

**[Download VQ-Bench from Zenodo](https://doi.org/10.5281/zenodo.17662700)**

The dataset is provided as a single ZIP file (approximately 2.5 GB).

## Dataset Structure

After extracting, you will find the following structure:

```
VQ-Bench/
├── modified_modal/
├── modified_breathy/
├── modified_creaky/
└── modified_end_creak/
```

Each directory contains audio files organized by speaker and scenario.