DanielRavenNest Noe

The reason it does not have upload is LL is using a proprietary "convex hull" routine. Convex hull is the default physics shape that wraps the 3d model you are uploading with no dimples or holes. So someone needs to find or write a convex hull routine that is open source, or content creators will just have to use an official viewer for uploading.

Nany Kayo wrote: Yeah, but what is the difference between Maya and 3ds Max? It has been said that either is fine, but why are there two products from the same company at the same price? The product descriptions from Autodesk don't make the differences clear. The discussion here does not mention any differences between the two either. Because Autodesk is the Borg of 3D companies, and they buy up any successful software that comes along, which includes *both* 3ds Max and Maya. They don't just sell two 3D programs, it's more like 20. My impression is Maya is somewhat better at animation work, and 3ds Max is somewhat better at architectural rendering, but there is a lot of overlap in features. In some cases it may come down to available plugins and compatibility with other programs, or simply which toolbar layout you prefer.

They are both models of an approximately 1975-1980 Buick, which was a fairly popular car at the time. The possibilities include both modelers working from the same original car from reference photos, or possibly obtaining a source 3D model from the same place, and then doing different optimization and detailing on it. Both processes are prefectly legal, if the source model allowed commercial use.

Or at least make the "Fix my registration" a sticky topic.

Vray is a raytraced renderer that can take several minutes to do one image. It takes account of many physical features of an object, like index of refraction and displacement mapping. Second Life uses a much simpler rendering engine because it has to display as many images (frames) as your frames per second rate. So typically it has a few tens of *milliseconds* rather than minutes. That's why SL does not look as good. If you have a fast computer with a good graphics card, you can turn up the graphics settings in Preferences up all the way, or use Kirsten's Viewer, which has extra features for photography and movie making. That will make Second Life look better for *you*. It has no effect on people with slower computers using lower graphics settings.

Argus Collingwood wrote: Could it be that LOD settings are needed to be raised like with Sculpts? [/me is a mesh newbie] If I remember Drongle's math right, levels of detail for mesh are automatically switched from High to Med, Low, and Lowest at 4,6,16, and 25 times the object radius. This can be modified by the LOD distance multiplier, but that's the default settings. It is up to the mesh designer what model to use at each level. When I build with mesh, I test out how it looks on the beta grid (Aditi) at different view distances, and adjust the model at a given level if needed. You can use the same model at different LOD levels, or 4 different ones, that's your choice in the upload window. If you are lazy and only do the highest detail model, the uploader will do an *automatic* polygon reduction for the lower levels. That only really works well for simple shapes like spheres. For most anything else, custom LOD model files will look better and cost less prims. Note that the physics shape - what your avatar bumps into and sits on, is a 5th model you can upload. I have a nice bumpy mesh rock which is 900 triangles at high LOD, but the physics only has 6 triangles since that is enough for a realistic result as far as walking and sitting on it. Less complicated physics is less work for the server, and you get rewarded for that by a lower land prim cost for the object, as does efficient LOD files.

Smoke Carter wrote: I hope you will retain your enthusiasm for mesh when your products start getting returned (or people keep telling you that you are wearing a balloon) because you cannot see the thing right at spitting distance.....even though it is half an avatar high. You appear not to understand how "levels of detail" work for mesh objects. "Levels of Detail" are displayed at different distances to reduce work on graphics cards for distant objects. The high LOD is shown within 4 times the object radius (plus or minus your graphics settings), then it changes to the lower detail versions as you move away. Did you make LOD versions of the chair you uploaded? The uploader makes mechanical decimations if you only provide the high LOD version, which in many cases results in a bad lower LOD result, but it's up to the designer to do the other levels. You can apply the same model file to multiple levels of LOD and keep the viewing detail high at a wider range of distance. That will up the PE cost of the object, because more detail is more work for the server and user PC. If the model shows as a blob or prim at any distance, that is a different issue, one of the viewer not having the bits of code required to show a mesh. The solution is to get a viewer with the new feature included.

122k is too much to upload, and will cost a lot of prims if you upload it in sections. Smoothing groups will make an object look smooth curved without a lot of triangles. If you dont use it, you get hard edges at each triangle.

Grace Winnfield wrote:.. but why the hades did someone have to make blender so friggin complicated. I think i need to memorize or write down every little key stroke gad! Blender 3D was written by open source programmers for open source programmers, not for human beings, and certainly not for artists. <-- that is a joke. It's complicated because you can do a lot of things with it. There are more limited programs that only do one or two things, but then you can't do that third thing you need.

A. Mesh items can be attachments like prim or sculpt objects, or linksets of more than one type. For example, I made a dress with a mesh top and a flexi prim skirt linked together. It acts like any other clothing item. B. It depends if the item is "rigged" to animate. If not, a mesh object can be resized when worn like any prim attachment. If it *is* rigged to animate, the size is fixed, because the animation depends on the position of the avatar bones. If you resized it, the animations would mess up because the parts of the mesh would no longer be in the right position relative to the bones. When rigged, however, a mesh object will change size with the underlying bones. So if you make the avatar taller, a mesh dress will stretch along with it. C. Mesh is either static or animated, depending if it's rigged to the bones. Position relative to the bones is fixed by the rigging, which the underlying avatar *skin* can change using appearance sliders. If the skin pokes out in places it is not supposed to, an alpha layer can hide that. Alternatives are supplying a shape for the avatar to wear (the old style SL shape as an inventory item), or supplying the item in different sizes. There is a JIRA request for a "parametric deformer" that will make a worn mesh follow changes in the avatar skin. That would make mesh more adaptable to the wide variety of SL body shapes. D. You can wear as many mesh objects as you have free attachment points, like any other attachments. E. Mesh is uploaded from one or more Collada ".dae" type 3D files exported from a 3D program. In that way it is similar to uploading a texture made in an image editor program. You make item first, then import it. Once uploaded, it appears as an inventory object, which can then be rezzed on the ground, or worn as an attachment. It does not use the sculpted texture box, sculpts are a different way to import 3D shapes, with a different asset type. If you want to create wedding and maternity items as meshes, you will need to learn some 3D modeling program to make the item shapes, then rig them to the avatar bones so they can animate when worn.

According to my Aditi Account Page I am age verified. And in the viewer it says I can access adult areas, but it will not let me TP there. Note that the Gibson default location is also rated Adult and I was directed there. Tried a second time after re-doing my adult verification, and it let me TP, then immediately crashed the viewer (Mesh Beta 3.0.3 - 239596).

(0) Linden Lab has never been good about documenting stuff (sorry guys, you haven't) Now to your questions: (1) If you want to animate the avatar body, it needs to use the same bone names and structure as the standard avatar. Search for Wiz Daxter SLAV on the web, or on the SL marketplace. He has a plugin for several programs with the rig already set up. I'm not sure what changes to the rig from the T-pose will work. (2) Existing animations will work if the new mesh avatar is correctly weighted to the bones. I made a full body mesh avatar and it works fine with pretty much any animation. Getting the vertexes weighted just right at the joints took a while, cause humans notice any defects in movement of other humans. (3) Use the Autodesk FBX 2011.3 plugin for your version of Maya. That is the one the Linden staff used when programming the mesh feature, so that works best. Export using Autdesk Collada, and at least in 3ds Max that I use, the only thing that needed changing in the export settings was to choose Z-up for the axis. Make sure your model feet are located at the origin (0,0,0) and is human scaled, facing +X, head pointing +Z.

Use the Autodesk FBX 2011.3 plugin (which contains the collada choice), and Export > Autodesk Collada (.dae). Make one Multi/SubObject material in the Material Editor, with 1 to 8 sub-materials. Apply that to your model. Make the object "Editably Poly", and in the Command Panel, select the Polygon editing level (red square). Below that is a place to assign the material ID. Select polygons of the model and assign a number to them. Each ID number will become a "side" in SL. Also in that section of the rollout you can assign smoothing groups. Any polygons in the same smoothing group will be "smooth shaded". Where they are in different groups, they will show a hard edge. Next use either UVW map, or Unwrap UVW to make a mapping from texture to polygons of the model. UVW map works for simple objects, Unwrap for more complex shapes. Without a UV map, you can only choose the color in the SL edit window. What the map does is define what part of a texture will be used for each polygon of your model.

Mylar wrote: That said; You don't have to learn to work with mesh. The way I see (and I reserve the right to be wrong) one of two things will happen. Either LL will add the ability for mesh avatars to wear existing clothing layers (something that has been hinted at for the future), or some industrious mesh designer is going to do a line of avatars that become popular and include their own clothing template that you can use. (which I've already seen work being done in this area) Regardless of what happens, there will always be a need for good photoshoppers. I second what Mylar says. I'm more of a sculptor than a painter as an artist - I can do the 3D shape, but not paint things like human skins as well as others can. I would *love* for good artists to make skins and clothing textures, while I work on variations on the body shapes (not everyone wants the same shape) and make clothing models (the 3d part). This by the way is how the Poser/Daz3D market works. One person makes the clothing item with templates, and then another comes along and does alternate textures for it. I already have an avatar body out with skin/clothing templates out, and I expect other's to come out soon, if not already. The templates alone are here: Templates download and the marketplace item with the body and some sample clothes are here: Female Mesh Avatar My avatar has a 2nd layer besides the skin layer, so you can add makeup / tattoos / inner clothes separate from the skin. I think that's a necessary feature so people can customize.

That looks great! For textures, what I use is Multi/SubObject material, with up to 8 submaterials (that is the most SL will accept per object), and apply to the whole object. Each material ID shows up as 1 "side" in the SL edit window. For funiture that is mostly seen at close range, you can keep more detail on the high and medium LOD, and be very aggressive on the lowest LOD. For objects this size lowest LOD counts the most for prim cost. What I did for plants is make a rendered image of the object in Max with camera set on X and Y axis, and put that image on a flat plane for the lowest LOD. That looks so-so, but the lowest LOD is used when very far away, and you can barely see it.

As a Max user, my guess would be you need to apply "reset transforms", collapse the stack, and make sure all your LOD and physics models all face the same axis (use world coordinates). Easiest way to check is import all the LODs into one file and make sure they align. But keep them separated for upload. Finished model should be editable poly. Once the shape is worked out for all LODs, then you can go back and do UV mapping for the highest LOD. Assuming you use the same material IDs in the lower LODs, you don't *have* to UV map them (you can if you want). If you don't expect an avatar or moving object to pass through the interior of your cabinet, you can simplify the physics shape. You dont need full boxes, just some sides to stop avatars from walking through it. The upper two thirds on the left where you have 4 triangles can be merged down to 2 triangles. Cutting excess triangles will cut your physics cost and also make less work for the server.

Collada isn't a file format, it's a set of XML tags to put in a text file containing the model data. You can open a ".dae" file with any text editor. Unfortunately there are so many 3d programs and so many formats, that not every .dae output is readable by another program, and the Collada standard is evolving. SL can see v1.4 files, but 1.4.1 or 1.5 give it problems (they define new tags it doesnt know about). So program incompatibility is not unique to SL, its rampant in the 3D field. There are even problems between different versions of the same program, or plugins that work for one version and not another. The most interchangeable formats right now are .obj and .dae, followed probably by .3ds (the older 3ds Max format, which is now .max). Hopefully in a few years things will settle down to a standard - that's the intent of the Collada project - but its an ongoing task. The best advice I have is if you find a method that works (like how to get a file into SL), write it down so you can repeat it, and share it with other people.

Aditi is a separate grid with a separate account, so you have to take the quiz again. Visit: Beta Account Page ( https://secondlife.aditi.lindenlab.com/ )

It needs a UV map for the object to texture in SL, Generally its better to import textures separately, less cost when you have to redo something, and easier to see what does wrong when the model is separate from the texture.

What I need to do in Max 2012, and so likely to Maya 2012, is export to FBX format file, then use the Autodesk FBX standalone converter version 2011.3 to convert to Collada. The Autodesk 2012 exporter was sending a newer version of the Collada file than SL could read. Also on Max I need to use one Multi/Sub object material for the whole object, with up to 8 material IDs as sub-materials. You may need to use whatever is the equivalent in Maya

(1) A mesh is a connected set of triangles which can have a texture applied to them. Everything in SL has always been a mesh. For example, the avatar body is about 8700 triangles, and a box prim has 108. What is new is the ability to upload arbitrary mesh shapes with custom texture mapping and physics shape. (2) Prims have a fixed geometry (how the triangles are connected to each other), and built in controls in the edit window to deform them in various ways. Sculpts have a fixed number of triangles defined by the sculpt map texture, and a single UV map. Mesh has a designer selected number of triangles, and up to 8 custom UV maps to apply textures to. Mesh also can have a custom physics shape, which controls how avatars and objects bump into it, and be attached to avatar bones, so it will bend and move with animations. Thats how the standard avatar animates - the skin follows the bone motions. Regular prims and sculpts can be attached, but not bend. Linksets that contain mesh or objects using the new physics types will have a variable prim cost that roughly accounts for how much load and lag they produce. Unlike standard prims, size matters! so do a lot of other factors. Variable prim cost is probably a bigger change for average users than mesh. A prim has always been 1 prim cost before.

No, only official viewer 3.0 and Kirstens S21(9) can see mesh right now. Other viewers will be adding it soon

I may have said it before, but Thank You to all the Linden staff who worked hard on this. We now have 50% more tools to build with (prims, sculpts, AND mesh) and infinitely more possibilities

What you are asking for is translucency done as depth, not just surface texture, plus internal surfaces for the ice defects/inclusions. We could get an approximation in SL, but to do it correctly you need a raytrace algorithm that follows the light as it goes through a depth of material, and that is too slow for a real time game. Raytrace is used in commercial graphics, but it can take several minutes to do one image. Second life renders muliple images (frames) per second. The ashtray is also showing "Index of refraction", the bending of light that items like glass and water do. SL doesn't have that feature in it's rendering engine.

Josh Susanto wrote: And if the lag reported is not coming from the mesh, from where is it coming? Releasing mesh at a time when so much else in the system is so obviously borked seems like a pretty bad idea... but only until you consider whether this actually stands to reduce the total number of problems people will attribute to mesh. What you and many other people don't understand is that Second Life has always been 100% mesh. The base avatar, the terrain, the sky, and prims are mesh objects. What is new is we can now add our own custom mesh objects. Well designed mesh objects have less complexity than other one. For exampe the base box prim has 108 triangles, cause it needs to allow for all the distortions you can apply to it. A mesh box can be 12 triangles, so a lot less work for the graphics card. A wall only needs 4 triangles, for inside and outside, if its part of a larger building and the edges dont show. That is massively more efficient than your standard prim house. On the other hand, badly made meshes can be very laggy, just like using lots of twised prim toruses are. At least with mesh there is an incentive to be efficient through prim cost.