AR Scavenger help

Submitted by lawrencej on Wed, 09/14/2022 - 02:46
Forums: 
Content creation

I've tried and tried but cant get AR Scavenger to show a 3D object on a target.

I've used .glb and .gltf for the 3D objects, even used the simplest of 3D objects but still nothing is working.

Has anyone had success with AR Scavenger that could provide some help?

otacke's picture

otacke

Wed, 09/14/2022 - 17:14

Hi!

Could you please specify what exactly you're referring to? What is "nothing" for instance? It's a little hard to understand what you're having trouble with.

If this is about not being able to upload .glb and .gltf files, then https://www.olivertacke.de/labs/2020/08/02/arrrrrr-ar/ should help. This has been asked numerous times by the way (https://h5p.org/node/1257381 or https://h5p.org/node/1291917 or https://h5p.org/node/1253972) ... 

Best,
Oliver

lawrencej

Thu, 09/15/2022 - 06:07

Hi Oliver

Thanks for getting back to me, no I'm not having trouble uploading the file formats.

When I test on any mobile device the targets aren't bringing up the 3D object at all.

The instructions also to use a 16 x 16pixel area for the target seem wrong. I did this and the resolution was terrible compared to the example in H5P.org

Any suggestions would be awesome thank you.

otacke's picture

otacke

Thu, 09/15/2022 - 11:10

Hey!

Have you checked that the size/offset of the 3D model are fine? It could be popping up, but be rendered off screen.

If you assume a bug, then you should provide information as suggested at the top of the bug report forum at https://h5p.org/forum/9 - in particular #7 if you have a chance to get hold of the debug output - adding the content that you use or a link to a page that does would also be smart. Otherwise it's virtually impossible to help you, as many people are using this content type without the trouble that you describe.

Internally, the marker library used for the content type uses a 16x16 resolution for the markers. Images that are larger are scaled down to that size. You can use whatever you like, but ultimately, the scaled down markers will be used for comparison, and just because you have different details in two large images should not lead you to believe that the internally used representations can be distinguished from one another. The marker should be large (in relation to the border that's added) or detection will fail, but it should be based on a 16 by 16 grid.

Best,

Oliver 

Best,

Oliver