search

cnr-isti-vclab / corto

Mesh compression library, designed for rendering and speed.
Other
196 stars 43 forks source link

Issue with UVs read back in Unity #28

Open lalalune opened 3 years ago

lalalune commented 3 years ago

Hello @ponchio ! Really love your work!

We are having some funny business with UVs. I am sure you are busy, but maybe if you have a second to offer some insight into where to look to fix this problem --

We are using Corto in our volumetric player package here: https://github.com/xrfoundation/universalvolumetric

So on the web, we have volumetric video working amazingly. You can see that here: http://wildcapture.co/volumetric.html

Screen Shot 2021-06-15 at 7 55 54 AM

But in the Unity build the UVs are really messed up:

Screen Shot 2021-06-15 at 5 35 49 AM Screen Shot 2021-06-15 at 5 35 46 AM

I would be happy to try to solve this myself, but if you have any pointers to what it may be / where to look --

One note, I compiled the Unity plugin for Mac. I will need to test with plugin builds on the other targets, but can't at the moment. Of course WASM seems to work everywhere.

Unity code is here: https://github.com/XRFoundation/UniversalVolumetric/tree/main/unity

Thanks for the help!

ponchio commented 3 years ago

I can't really debug or test unity now. but I strongly suspect the problem is with the normals, not the UV:

line 82 in https://github.com/cnr-isti-vclab/corto/blob/master/unity/CortoMeshLoader.cs seems a likely culprit: the normals are recomputed and fail on the seams.

I got a student to write this plugin, could you try to uncomment the line (I suspect there might be some issue with normal format or such hgence it was commented) and test it?

lalalune commented 3 years ago

Yes! Thank you, I will report back :)

On Tue, Jun 22, 2021 at 4:18 PM Federico Ponchio @.***> wrote:

I can't really debug or test unity now. but I strongly suspect the problem is with the normals, not the UV:

line 82 in https://github.com/cnr-isti-vclab/corto/blob/master/unity/CortoMeshLoader.cs seems a likely culprit: the normals are recomputed and fail on the seams.

I got a student to write this plugin, could you try to uncomment the line (I suspect there might be some issue with normal format or such hgence it was commented) and test it?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cnr-isti-vclab/corto/issues/28#issuecomment-866399106, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEOFEMAGQ5MJZ7QGPOPRSJTTUEK4JANCNFSM46XOL3XA .