Skip to main content

Unprocessed Data

  1. Usage of raw data is not recommended unless you know what you're doing, or can be used to obtain component types such as binaural audio or 3rd person video data
  2. There are some missing raw components. Please create an issue on GitHub if you are wanting this data.

All data has been processed to some extent. The data available here is simply what each member of the consortium has provided in their "raw" form (as raw as you can get). Most video components have been compressed. Top level metadata specifies the commands used for compression and other potential issues with them.

You can download the data with the CLI specifying the string --datasets components/<type>, where <type>is one of the following values:

<type>DescriptionSizeQuantity (# components)
videosVideo components20TB24246 (8 missing)
imuIMU114.8GB6396 (498 missing)
gazeGaze (CSV)0.5GB32
burned_in_gazeVideos with a 2D point overlayed on them specifying the Gaze175GB38
binaural_audioBinaural audio assocaiated to videos20GB29
3rd_person_video3rd Person Videos155.6GB37

For some components there are some deficiencies to them. So please read the appropriate section in the document before using it.