saindd

Does anyone know if the T3D integration with Polyvox was finished? Bill Vee was working on this, but apparently abandoned the project, without releasing any examples or source. This would be a very nice addition to the codebase. :D

https://bitbucket.org/volumesoffun/polyvox/

Has anyone implemented IK & FK inside T3D? I've been searching the old forums and the TDN, but i wasn't able to find any docs, samples or anything. It would be really useful for snapping the players hands to the weapons or items he's using. Kinda like Final IK in Unity.

I've always been curious about how the people behind this kit managed to achieve ledge grabbing. Is it raycast-based, like a ladder?

Although the realtime lighting in T3D looks OK, it really shows some age compared to other open-source rendering engines like Ogre and Panda. This could be "easily" fixed by introducing lightmapping, instead of refactoring the engine into supporting a modern approach like PBR and realtime GI (as seen in Tesseract).

Right now, i'm lightmapping my scenes in Blender. This has the disadvantage that i have to build all the scenes in Blender and export everything as a single mesh. And i can't use terrains. Is there any plans of adding a lightmapping solution to T3D? Maybe using an open-source offline rendering solution like:

https://github.com/tunabrain/tungsten

http://www.mitsuba-renderer.org/

I remade the weapon skeleton and animations as well, both for arms and weapon models. The originals also had terrible issues.

Just to give you an idea of how buggy it is: i bought the art packs from GarageGames and used the source MAX files to export everything back to the engine again, with no changes at all. Bugs everywhere: wrong animations, bone issues, etc. What a mess.

Don't use the original skeleton, or the built-in animations, it's completely buggy. Take some time to build your own. Much less headaches.

http://www.garagegames.com/community/resources/view/21450

https://www.garagegames.com/community/forums/viewthread/124513

That's nice, thanks! In order to display the transparent version of the mesh, should i just apply a different material and, if the user clicks, apply another? Also, how can i efficiently do this without killing memory? I intend to spawn the TSStatic at the raycastPos and remove it soon after the mouse moves and just not remove it once the player clicks.

Is there a mesh creation API for Torque 3D, that allows me to create meshes in code and adapt them procedurally according to parameters? I'm planning to create something like this:

I've been working with World Machine for a while and i'm producing some quite useful masks for my terrains: grass, bush and tree masks, as well rock masks. Those masks are really useful because not only they are precise, they speed up the workflow.

Is it possible to import a mask into the Forest tool or the object placement tool and place whatever object i configure to be randomly placed following the mask data (blacker = less items)? If not, how would you guys approach this? With a shader? Right now, i'm using the mask as a diffuse terrain texture and manually placing stuff. Which is AWFUL.

I've tested on Unity 4, btw, which doesn't have DX9. I'm pretty sure this is something dumb with limits or something like that in the terrain class itself. I'll mess around with it.

I did try on UE4 and Unity, i didn't had those issues. :(

Max > DAE Exporter > Save into art/blah/blah > Open Shape Editor > Open collision tab

Try to change the fit type. Also, which log should i send to help developers debug this?

ps.: I'm on Win x64.

Just use any mesh without a "bounds" object inside the DAE file. And try to create the bound inside the Spae Editor. It crashes.

Can anyone try to reproduce this problem? It's quite easy to do. Hopefully it will get the attention from someone maintaining that part of the code.

Yeah, the problem only happens in Torque. I have tried Unity and Unreal and there are no "strange triangles" in the terrain. I'm almost sure this is a resolution problem. Maybe the terrain computing a wrong LOD for the current camera?

Problem is, when there's no box, the physics get fucked up and the player is positioned incorrectly in the world.

Nope? :(

Is there a way of having some visual feedback when setting the bounding box?

How do you determine them in the datablocks? Also, wouldn't that be bad, due to differently sized player models?

That's what i'm referring to. Look:

http://torque-3d.readthedocs.org/en/latest/artist/playercharacter.html#bounding-box

Without them, the engine is not able to create the collision properly and the player ends up falling through terrain and so on. However, i'm not creating one with the character and wanted to generate that inside the engine. The docs say it's fine, but the engine crashes.

How do you create player collision, then?

When rigging a character, should the bounding box be rigged to the skeleton, to match the player body throughout animations?

When trying to generate a bounding box for a shape inside the shape editor, the editor crashes. Select fit type "box" then fit target the model itself. Boom. It also crashes if you change the fit type to anything other than box.

How can we properly report engine crashes? Should we record a video with the steps? Or just a console.log and a few breadcrumbs on how to reproduce it?