hubs-foundation / hubs-avatar-pipelines Goto Github PK

The video based avatars do not seem to show the video camera on start. Not sure if this a limitation with hubs cloud vs moz vanilla hubs

Exporting to glTF from the provided AvatarBot_base_for_export.blend gives a model that does not scale the head when you speak in Mozilla Hubs. AvatarBot_Base.glb in the "Exported GLB models" folder has the same issue.

Exporting two objects from Blender makes Mozilla Hubs guess (I assume) at which model is the body and which is the head and does give a setup where the "head" scales. So it's not clear why the setup in this repo does not work.

I'm on Blender 2.82.

Why it fails to import in Blender 2.9?

Feedback from IEEEVR: the humanoid avatar animations are a bit hard to see who is speaking - We may want to consider if there is a scale we can apply that would change how the audio responsiveness is handled, but wanted to capture that feedback here as well.

This happens with some of the currently uploaded avatars and also the avatars I made myself. When I use them I see portions of the mesh visible as I move. I see some vertices wobbling or stuck at a point. This glitch is only visible to the user of the avatar other users cannot see it on my avatar.

https://imgur.com/a/t7U9PbO - example

Also I have noticed that if I use a selfie camera this glitch disappears, however when I remove the selfie camera it occurs again.

Any clues? Thank you!

I'm putting this here to track it, not because I think it's necessarily an issue with the .blend file, so we don't lose this. This was observed on 2.80 beta (b331515d2e47).

1. Download the https://github.com/j-conrad/hubs-avatar-pipelines/blob/master/Blender/AvatarBot/AvatarBot_base_for_export_altbody.blend file

2. Attempt to export as a .glb or .gltf file using the exporter out of box

3. Result:
   

This did not happen when I followed the flow of importing the alternate body .glb, making changes, and re-exporting. I'm not familiar enough with Blender to quite get a read on the issue.

Follow-up to #5 - it isn't always clear that there are additional steps that have to be taken to enable the audio driven behavior for avatars. This issue is to track work done (e.g. if we create a plugin) to make this easier

I can't open the .blend files in any these version of Blender:

2.93
2.92
2.9
2.83.13
2.8

message error:
File format is not supported in file 'path/AvatarBot_base_for_export.blend'

Just that. File open and Blender says "File format is not supported"

Any ideas?

I've tried opening several of the files in https://github.com/MozillaReality/hubs-avatar-pipelines/tree/master/Substance and they all seem to crash the latest version of Substance Painter.

Do they need to be updated or should I use an older version? I've filed a bug report (several actually) with Substance also.

Substance Painter version: 2020.1.2 (6.1.2) on Mac

i have made a humanoid avatar (custom made in blender ), everything work fine exept the fact that when i come too close to another player or the "selfie camera" option the model disappear insted of become faded like the default one ,any guess what am i doing wrong?it is an expected behaviour?

I can't seem to find out how to export a webcam based avatar from blender. I've tried importing and exporting both the provided astronaut GLB and the BLEND file, and while both exports show the avatar in hubs, neither allow avatar webcam sharing.

Is there a specific method of exporting the GLB that's required for it to include the relevant webcam properties? Thanks

Hi, the mode provided in the hubs-avatar folder cannot be opened with blender 2.82