AR cloud does not share information about objects.
I looked at LeapBrush.
Leap Brush is too hard to follow, there is no documentation about it.
I'm getting the idea to use photons for share information about objects.
Can you suggest that method to me!
Hi @thangpro, One guide we just published focuses on using Photon Fusion: Building Multi-User ML2 applications with Photon Fusion.
Please let me know if this helps unblock you.
Hey @thangpro , can vouch for the guide Kevin provided and would recommend using Photon over the simple Go server that comes with leap brush. We may consider updating the leap brush sample to use Photon as that has been common feedback from developers.
With that said, what size difference are you seeing with the anchor location?
anchor location deviation about 3 to 5 cm with ar cloud, I will check Photon Fusion I don't know if it's very accurate
Oh that seems higher than I would have expected if it is a high quality map, would you be willing to share the map file with us so we can analyze the quality? Additional data such as AR Cloud and OS version is helpful too
Also can you confirm from the screenshot if that is an outdoor location or indoors / outdoor adjacent?
When I use MagicLeapPhotonFusionExample, I see there are 2 options for the location.
-Marker Tracking
-Magic Leap Anchors
I think about Marker Tracking.I won't need to use ar cloud.
With ar cloud I will have to scan many maps, but with Marker Tracking there is no need.
I think it depends on your use case, AR Cloud with a high quality scan will yield accurate results that aren't tied to the current headpose session. If any of your participants lose headpose you would need to realign with the marker where anchors will realign automatically once the headpose session is restored. Also note that running marker tracking passively will add additional Compute and thermal concerns to your app experience.
If you want an ephemeral space and a highly accurate digital twin is not a value add then you can get something working quickly using Marker Tracking to align the origins with the above downsides mentioned.
We have heard similar feedback from customers and will be soon previewing a feature that effectively maps automatically which would allow you to effectively share a lower fidelity map under the hood with a single anchor and bypass some of the Spaces / Mapping flows you referenced.
With all that said it sounds like Automatic Mapping is a feature you want in the near future but for now I would refine your application logic with a Marker as a proxy and if you find the added precision and failure handling of headpose loss is important to you then to use AR Cloud scanning instead.
I'm looking at working outdoors with low granularity, map sharing sharing skewed anchor points. , don't have an api yet, and have to scan the space many times. Looks like marker tracking would be more convenient. I am planning to use gps to increase the accuracy of the header session, how can I do this
