Yes. I think it would be a very good optimization feature. Devices with bad internet struggle to load in many different meshid. It should be more performant loading one unique meshID for identical imported mesh rather than 500 different meshid for mesh that look the same with each other. I think you guys already did it with the TextureID for mesh that share the same texture.
Some mesh also have the same base mesh that is slightly scaled in different direction to get new shapes, etc. If you can identify this as well, it would be very nice.
Very interesting, I’ll look into that. If so that really solves the major conundrum.
My worry would be that when exporting whatever instancing sceme you have in blender or maya gets “squashed” and copies of the instances are duplicated.
If you export a file with instance as FBX and re-import it in the same editor (say blender), are the instance relationships conserved?
Pretty sure, they are linked if they are the same base mesh using the same textureID unless you set the object relationship (material, transformation) to single.
We are aware of the 500 issue, it seems to only happen on the first import of a studio session. We aren’t totally sure why its happening but are investigating.
As we said, the simplification is very crude. We have 2 algorithms, a “precise” one and a “sloppy” one. The precise one is obviously much better, but it cannot guarantee that we will reach 20k triangles. If we aggressively run the precise algo but it cannot get below 20k; we then run the sloppy one which guarantees we get below the limit, bit it is, uh well… sloppy
I’m not sure if anyone’s asked this yet, but are you planning to add multiple materials/textures per mesh? It’d be a really helpful addition, that way we don’t have to break up meshes into multiple parts.
I’m loving everything about the new importer, everyone involved did a great job! (yes I am a bit late to this, I know).
This is something we are brainstorming. Currently there are engine limitations. The engine itself does not permit more than one materials per part/mesh-part.
The only possible short-term solution, which doesn’t seem to be the solution most people want, would be to split the mesh into multiple meshes at import-time, one for each material, but this brings with it a slew of new issues and inconsistencies.
Right now the team is still quite small, and since there is no clear “great” solution for this problem without engine changes, we are focusing on other higher impact improvements. But it is something we are thinking about.
I like to think of this as a feature! Though I must admit, not being able to load the normal, roughness and metallic maps has made it so I ignore all materials imported by the importer.
We have a few updates planned for the next few months, while we work on those, the feature will remain in beta.
Also, Studio now has too many different import products, releasing this one out of beta too early would just add yet another importer which makes studio more confusing and harder to use / learn.
The goal is to unify the different importers into one user-experience that is the best of them all. Until we feel like we can achieve that, this feature will remain in beta.
However this importer creates meshes and models that are just a “real” and legit as all the others, the beta moniker is being for this transition period as we do the work needed to consolidate the various import experiences without removing any tools or workflows that our devs are currently benefiting from.
