Jump to content

Maniac Choche

Resident
 View Profile  See their activity

  • Posts

    39

  • Joined

  • Last visited

 Content Type 

Posts posted by Maniac Choche

    3D max triangulated Physics shape problem

    in Mesh

    Posted

    Did you clicked "Analyze" and set it to Prim ? Mine are blue. I will quote Drongle here:

    “I don’t know the actual numbers, but the color indicates the magnitude of the physics weight, that is the strain it puts on the physics engine. Blue is nice and simple, going from green to yellow to orange to red as the weight becomes excessive.

    It shows you that the default convex hull is smaller than it should be!”

     

    Wearing ~8Mil triangles on a single attachment point = not a problem...

    in Mesh

    Posted

    So I was able to increase my teapot linkset to 9.3Mil tris. (yay). I've set my RenderVolumeLODFactor to 0.00 and was able to attach it 20 times ( just for test ) on the same attachment point - right hand. The resulted tris count should be 186Mil. When I made my RenderVolumeLODFactor back to 4 and tried to zoom , I crashed on 20Mil tris. So I can walk around with this  "bomb" with my LOD factor set to 0, and anyone who is stupid enough to zoom at me is down.. 

    This will be great for jumping out of a cake and crashing all guests :)

    • Like 1

    Wearing ~8Mil triangles on a single attachment point = not a problem...

    in Mesh

    Posted

    Haha yea, I've made the test on an empty sim :P

    Still, I see no reason for designers to make optimized meshes for attachments. I always thought lindens are trying to make new users to stay, but how to explain to a newbie that he/she is lagged like hell because the freebie he/she is wearing is 2Mil tris...

    Attachment limits need to be placed. This will force designers to make better and more optimized meshes.

    [Feature Request] Skinning with Dual Quaternions

    in Mesh

    Posted

    Actually there is no need for support inside your 3d software of choice. It's cool to have one to be able to see the result in real-time, but the method only interprets the data that is already inside the COLLADA in a different way ( current is linear ). That why the implementation should be easy ( *guessing here* ), because the changes will be only inside the SL Viewer's vertex shaders. Because of the more complex calculations this can slow down the rendering a bit, but in comparison to the other available methods ( log-matrix and spherical blending ) this one is really fast. The current practical fix for "candy wrapping" problem is to add extra bones ( twist bones ) to the skeleton. This can't be achieved in SL at this point. So dual quaternions can be a cheap way to avoid those problems and it can be available only to users on Ultra graphic settings, for example. 

    How are we supposed to market Mesh prim count?

    in Mesh

    Posted

    Hello,

    One other choice is to provide demos of your products. Demo textures, no mod with resize script. Mesh objects are much more difficult to steal and even if the thief is successful he/she needs to have "payment info on file" in order to reupload the model and to pay for it. So providing demos will be much more safer. 

     

    Best regards

    Help with bug test?

    in Mesh

    Posted

    Okay my uploading tests (same steps like in your second comment on SH-2045):

    Five attempts:

    1. 3.196

    2. 3.196

    3. 3.167

    4. 3.595

    5. 3.869

    There is definitely something wrong ..

    Best regards

    PE = tier = real dollars

    in Mesh

    Posted

    Well no matter what is merged, the streamed information for a 1024 vertex mesh is several times bigger than a 1024 vertex sculpty. With a sculpty you get only vertex positions for the highest LOD all rest is calculated by the viewer. That's why sculpties was invented to ease the streaming cost not the FPS. It's logical that LL cares more about their internet costs than our FPS. 

    Best regards

     

     

    PE = tier = real dollars

    in Mesh

    Posted

    Gaia Clary wrote:

    But how does this match up with the fact that i can make meshes which use much less resources in terms of triangles, LOD, etc...  compared to equivalent builds with regular prims/sculpted prims.

    Maybe the problem is related in the amount of streamed information? For example:

    - regular prims have fixed normals , fixed UVs , all those are not streamed

    - all information related to a sculpty object is hold in a very compressed image file

    I've made a very rude test with a sculpty:

    In it's native JPEG2000 format the size is ~4k. Same sculpty in uncompressed .dae is 1.05Mb. Since i have no clue what type of compression LL will use internally I've used RAR just to see how much i can compress it. I came up with something like 57kb and best possible compression settings. The differences are quite big , but the amount of information is more too. And let's not forget custom created LODs and Physics all those have UVs , normals , vertex positions etc. It's not just one DAE.

     

    Best regards

     p.s .dae file have tons of tags that are probably ignored and not streamed

×
×
  • Create New...