[HELP REQUIRED][PAID] Real-time performance feasibility

Hi!

We’re building prototypes for two online viewing services that need real-time or near real-time delivery (5000-7000ms in one use case and 30000ms).

What concerns us is if it’s possible to feed the models in the scene with a continuous stream of *.FBX data which we’re rigging real-time to one or more models found in the scene.

If you’re a 3D generalist with Unity/UE4/WebGL and coding chops (I assume some torturing of JS to pull this one off), come talk to us.

Stay safe everybody!

Hey! Are you building realtime holographic / animated scan viewer or smth similar (HCAP, etc.)?

The engine itself doesn’t support .FBX import (it’s locked for in-editor use only via FBX -> JSON converter), but it’s possible to convert incoming FBX to GLTF via ThreeJS lib, and then import GLTF into PlayCanvas engine as a valid model asset. Or write custom FBX importer for your purposes, if that’s an option.

But that is for overriding the entire asset. If you want to provide a stream of FBX data that should be converted to some new vertices / triangles added to existing mesh - then it will require a significant custom effort I suppose. If it’s just stream of new animation data - using FBX -> GLTF pipeline should work.

Cheers!

2 Likes

Yeah, I’d recommend an FBX2glTF tool like this one. FBX isn’t a transmission format - its bloated and difficult to parse. Unlike glTF, which is fast to transmit and easy+fast to parse. And PlayCanvas has excellent 1st-class support for loading glTF scenes.

1 Like

Hello,
Greetings

I’d Glad to assist you.
PM Sent, Please Check.

Very Best Regards
Carter W

1 Like

Hello Will,

Thanks for the suggestion. Now I actually see we need to look at the fundamentals of our current stack - the FBX is clearly both on its way out and not suitable for streaming.

No problem, @bunraker - if you have any more questions around that, just ask.

1 Like

Hey!

Thanks for the reply. We don’t start in the XR domain, but of course, we build with it in mind - it’s just more further ahead on the roadmap. I’m acquainted at a glance with gLTF via my other work that is located in XR space and the Khronos group spec, but nothing more besides that.

I don’t believe we need further FBX manipulation, we just need a way how to feed the pipeline continuously the incoming data at runtime.

The offer is still on if you’re interested, shoot me a DM with your details and we’ll talk more.