Raggar

Members
  • Content count

    113
  • Joined

  • Last visited

About Raggar

  • Rank
    Advanced Member
  • Birthday
  1. I'm not gonna make it, like, at all But I've learned a lot, and will continue making something I can hopefully show you guys sometime in the future.
  2. Yayh. Commit confirmed to fix the bug. Nice @Deltakosh. However, and I don't hope I'll be hijjacking the thread here, althought the issue has been resolved. I have a new issue, that I'm looking for a way to resolve, one way or the other. Here's an example of the issue: The player model has a mesh, a skeleton with bones and some animations. The mesh's arms are rotated accordingly to the mouse/camera orientation, and done using rotation of the bones. Now, if the skeleton runs an animation, the manual rotation is being overwritten by the keyframes from the animation. Not only that, but bones without any animations assigned to them, overwrite the manual rotation, so when running, you are not able to see the 'aim' of the model, only when no animation is being played. Is there any way around this? Like offsetting the bones' rotations and positions or maybe have multiple skeletons? As mentioned, even if the bones have absolutely no key frames assigned to them, their positions and rotations still overwrite the manual ones.
  3. I struck the same issue some days ago After updating Babylon, my own mesh is unable to play it's animations, as well as rotate bones using any of the setters. But I honestly haven't checked whether my skeleton is animating without the verts binding to said skeleton. I will do however, as it might help in narrowing down a spot where the issue might occur. It weird, if you try importing the "dude" model, he still has his mesh bound to the skeleton. EDIT: Just checked. The skeleton is indeed playing the animation I assign to it. The connection between the mesh and the skeleton is somehow 'lost?'. Maybe @adam knows which changes might have caused this.
  4. Thank you for all the help. I've marked this thread as Solved, as I see the main issue as being resolved. Although using your babylon.max.js file, my own .js files and even the overwritten prototypes functions from your PG examples still produce some weird issues, I'll figure this out, eventually. For now, I'm back at importing models, but I'll make some time to look it over again at a later date. When using the dude model, I can manually set the rotation, as long as i DON'T run any animations at the same time. But my own model won't either animate nor manually rotate after updating babylon. Same thing with the overwritten prototypes. The model is only a placeholder, and took less than an hour as I didn't refine skin weights or anything. Therefore I won't completely rule out that the problem might be fixed when finishing a proper model. Attached is the model I've been using for reference. e.babylon
  5. Oh. I did understand what you meant. I was just thinking out loud about how I was planning on achieving something similar
  6. If you use the 'Black and White' filter in Photoshop, you can adjust the strength of all colors, like the blue, orange, green etc. This took literally 5 seconds, and I'm sure there is something in GiMP that creates the same result if you prefer a free application.
  7. Very nice description. I thought about doing something similar with monsters in a future project. But using rays might just be bit too intensive on the server.
  8. Impressive! One question, though. How does the 'monsters' find their ways to your characters? They seem to follow you around, but it doesn't look like ray mechanics. Are you, like, saying player X attacked you, player X is at position Y, go to position Y?
  9. http://www.babylonjs-playground.com/#1HH4OJ#18
  10. It's weird. Same thing as my own compiled version. Getters seem to respond, but no change whatsoever with the setters. With the 2.5 version, everything works fine, although the cloning issue is present. But with the newest version, keyframe animations stopped working, and neither of the setters seem to work. .rotate, .setRotation, .setYawPitchRoll etc. I guess I'll have to investigate what's wrong. The code is the same, with no changes. Btw. Have you checked your PG? Rotating the bone manually seems to work independently of the original mesh, but the look control seems to have the same issue, or at least one resembling it.
  11. Rotating the bone manually seems to work, but the look control still seems to affect both the imported mesh as well as the clone. I hope my compilation somehow messed it up. Can you upload your version?
  12. mesh.skeleton.bones[boneIndex].setRotation(new BABYLON.Vector3(x, y, z), BABYLON.Space.LOCAL); The X, Y and Z values are based on the rotation of the bone in local space, from when the mesh was first cloned. Then altered of course. But the keyframe animations stopped working, too. So I'll do some hunting in the next couple of hours.
  13. Compiling the master branch somehow broke the bone rotation completely. Switched over from 2.5, so I don't think it's due to API changes. I'll check the simpler example from your PR when I get home. It was my fist time compiling from source, but everything else seem to work just fine.
  14. I just tested the multi-load solution, and can confirm that it is working, although your PG already did just that. I'll let this be a temporary fix.
  15. Yeah. I thought of 2 solutions. 1) Split the mesh into multiple meshes, and rotate the specific parts instead of their corresponding bones. 2) Loading the model everytime it is required. But then, what if I need 20 player objects? That's 20 x importing of meshes and skeletons/bones, and might give some drops in performance every time a new object is required. Maybe IndexedDB could help, by at least cutting the bandwidth? I tried #1. Didn't quite like it. I'll give #2 a go, and see whether or not it is practical. Can you verify that this is indeed a bug? I've been through my code many times, doing various optimizations and bug hunts, and I'm certain that I've done it right.