---
license: cc-by-4.0
license_name: cc
license_link: LICENSE
---

# MSDWild

The MSDWild dataset is designed for testing multi-modal analysis in the following tasks:

- Multi-modal Speaker Diarization
- Multi-modal Speaker Localization
- Audio-visual Lip Sychronization

For further details, please visit the [MSDWILD GitHub repository](https://github.com/X-LANCE/MSDWILD).

A sample from the dataset can be viewed on the [visualization section of the repository](https://github.com/X-LANCE/MSDWILD/tree/master/visualization).

**Important Notes:**

- The database is intended **solely for research purposes**.
- Responding to community feedback, we have uploaded a `video.zip` file to our repository due to the unavailability of some videos online. This action is aimed at improving the replicability of our research within the community. These videos are provided for research use only and any other use is strictly prohibited. All usage must comply with our licensing agreement.
- These materials may be removed at any time upon request from the original video owner.