What is ASPIRE?

ASPIRE is a a first of its kind, audiovisual speech corpus recorded in real noisy environment (such as cafe, restaurants) which can be used to support reliable evaluation of multi-modal Speech Filtering technologies. This dataset follows the same sentence format as the audiovisual Grid corpus.

Some samples of the data


The following people contributed to the planning, development, collection, and annotation of the INSPIRE Corpus: Mandar Gogate, Kia Dashtipour, Ahsan Adeel, Ricard Marxer, Jon Barker, and Amir Hussain.


This research was funded by the UK Engineering and Physical Sciences Research Council (EPSRC project AV-COGHEAR, EP/M026981/1)


The package including the videos, audios and the metadata is available for non-commercial, and academic research. You will need to sign a License agreement before getting access. Please fill the form available at https://forms.gle/bwLnkyJrkBHaaEwg7 for a copy of the agreement. The signer should get familiar with the latest EU General Data Protection Regulation (GDPR), and is responsible for controlling who in his/her group gets access to the data, and how the data will be stored and used properly according to GDPR. The signer will be fully responsible if any issue rises in relation with the GDPR about the usage of shared data in his/her group. Once approved, you will be supplied with a unique link, and password to download the package. Please cite the following article if you make use of the dataset.