Drongle McMahon

Select the car, check "Edit Linked", xlick on the seat, then click "Unlink". That will unlink the seat. Now select the rest of the car THEN shift select the seat, then click "Link". Voila!

"The normal map and AO are applied to the mesh inworld." The normal map(bumpiness) s applied inworld, but if you want to use baked AO, I think you have to combine it with the diffuse texture (brightness contrast adjust AO and multiply diffuse by AO) before uploading*. This is because, so far, there's only one diffuse texture per face in SL. If they ever allow compositing of multiple diffuse textures, which has been requested, this will mean we can use separate AO and potentially save a lot of texture data download (tiled small diffuse plus small AO). *unless you know some way of using the specular map to mimic ao?

... or tell us where they are, so that we can go and see what you mean?

Ah yes. Now I can see the numbers - so it is the GLOD library failing to handle the complexity. In that case, the solution is to make your own LOD meshes. This is very flat. So for the lowes applicable LOD I would use a 6 or 8 sided circle, duplicated with inverted normals so it has two sides, with an image of the high LOD as the texture. You have then to add a hidden triangle at the higher LODs to put a new material with the low LOD texture on. Which LOD to use this at depends on the final size. If it's less than about 8m across, it's the lowest LOD, then the low LOD as well up to about 15m across, and the medium too if it's >60m.

Look mat the individual weights by clicking the "More Info" link in the edit dialog. My guess would be that you have a high physics weight. That depends on how you set up the physics shape. By default, it's the convex hull of the low LOD mesh. If you also used default LODs, the low LOD of an item like this still has a complex hull with many vertices. If you selected a LOD and didn't "Analyze" (convert to a set of hulls), that would be even worse. If it is the physics weight, then you need to provide a nice simple mesh for that - maybe a six-sided cylinder (Analyzed physics weight 0.52) and Analyze. Otherwise, if this is very big, so that only the high LOD counts, that will give you a high download weight. ETA: also, the automatic LOD algorithm might have a hard time with a convoluted thing like this. What are the triangle counts in the LOD steps when you upload?

Should LL decide to use any of these , I want 0.5% of the net income" Sorry. I think you need to read section 7.4 of the ToS. :smileysad:

I went to Aditi and played around with two identical 2x2x2 cubes exactly superimposed. I don't really umderstand. but here is what I observed. When you shift drag, the ome that moves is the old one, and the duplicate is left behind. After superimposimg them, I could onl;y see the blue cube. Then when I linked them, the yellow cube became the root, and I could only see it. If I used edit linked and either moved or scales the selected yellow cube, then the faces that remained superimposed (in the same plane) immediately exhibited z fighting. Whem the edit was restored, the z fighting ceased. Now I started making duplicates with shift drag. The new copy (the one left behind when you drag) always showed z fighting as soon as the old one (the one moving) started to move. If I then moved the camera around, and/or zoomed, the z fighting evenyually stopped, sometimes with the yellow cube, but more often with the blue cube visible. All new cubes showed z fighting whether they were made by shift dragging blue, yellow or z-fighting cubes. All three kinds re-rezzed after copying to directory also always showed z-fighting, although occasionally it was only transient. From what I said before, from my point if view the unexpected behaviour is when the z fighting stops. I assume this happens when by chance the camera angle etc. cause one cube to completely occlude the other, causeing the rendering system to mark the latter as occluded, so mthat it never tries to render it again. This is consistent with what happens if the blue is made transparent. Now the z fighting surfaces are yellow and grey (yellow seen through transparent blue). The behave the same way, except that the grey cubes start z fighting again at some camera angle and/or zoom. Thta is presumably because the yellow cube cannot be excluded by the transparent cube, so that it is never marked as such. Advice remains the same. Move the one of the surfaces so that they don't z-fight.

Now you are confusing me. Your first post seemed to imply that you had two objects exactly superimposed. Now It sounds as if you have one object with two copies of the geometry superimposed. If it's all one object, that's rather different. It should still be possible to get the two layers of geometry separated by enough to prevent the effects, but the separation has to be enough to avoid rounding errors when the geometry coordinates are converted to 16 bits on upload, and when their distances from the camera (z) positions are calculated inside the rendering engine. I would still try the solution out, as long as one of the layers is supposed to be always on top. If not, I can't suggest anything. If this is the explanation, I have to suppose that the changes you see when you drag or rez the object might be dur to differences in the rounding errors for different object and camera locations. If that is the case, then at least the drag remedy would be unreliable, sometimes turining the z fighting on and sometimes off. How reproducible is it? I suppose there is a slight possibility that the two layers are made of (presumably identical) quads but are differently triangulated by the uploader (which only ever uploads triangles). In that case, the calculations would be different for the two layers and that might have an effect on z in some camera/position combinations. If that's possible, then you might get improvement by triangulating them before export to collada, and checking that the chosen diagonals are the same on each layer.

You might also like to look at http://wiki.secondlife.com/wiki/Limits There is a section on Mesh, and the following note that imposes a tighter limit on triangle count (21844/material) because of a long-standing bug. However, while it's difficult to give advice with any precision without knowing the type and size of the objects, you should almost certainly be aiming at triangle counts much smaller. Fine detail that isn't required for the silhouette can be simulated by using a normal map baked from a high poly mesh to a low poly mesh that has geometry at a lower level of detail. Most of the main 3D modelling programs can do that, and there are specialised programs as well. I would suggest using the decimator in your software to make a who;e lot of models with decreasing poly counts. Then try them out on the Adit test grid (or the main grid if you 3D printing company is profitable enough!). See what they look like and see what their land impact is, remembering that you can rez about 234 LI units per 1024 sqm (32x32m). When you get an idea of what is feasible, then start baking detail into normal maps if you need it. There is also information on the download weight part of LI. and its dependence on size, here, although it is quite complicated and probably not necessary for your purposes. Also, for import into SL, you need Collada (dae) files, which you didn't mention in your post. So you may need another program to convert from obj or whatever.

I don't think that would help. All the system clothes for each body segment use the same UV map. So mthat only needs one for the old system clothes. You could combine that with the a different one for new clothes, if they provide the facility for alternate UV maps and a means (GUI and/or script) of specifying which to use. However, that would not help with compositing the different layers or responding to the clothing slider settings, as far as I can see. Anyway, I know almost nothing about SL clothes. I haven't changed mine for five years pr so. So I won't say any more.

Yes. I suppose you could use mesh layers, but I wouldn't count that as using the default avatar. I's really a highly modified one. Do they work perfectly with all the sliders? Are they perfectly rigged for all animations etc?

"... to port an existing SL Avatar including its shape and textures ..." Did you consider the baking of legacy outfit layers? I would guess that is unlikely to be replicated in the new world in a compatible way. I think the old avatars would be of limited use if they can't wear system clothes (might be wrong, maybe everyone alse wears entirely mesh and needs on;y one skin layer?). I guess you could get around that if you could get access to all the layer textures and composit them with photoshop/gimp or a purpose built utility. Wouldn't get the same flexibility, but might be useable, I suppose.

"Of course sales are down, and merchants have stopped building new items, every one is to busy reading the forums." :smileylol:

Perhaps I need to distinguish more clearly between adding the modifier and applying it. When you add the modifier (to the object's modifier stack), everything you have done up till then gets mirrored. So does every modification you make, whichever of the two mirrored things you make the chanfes to. Adding the modifier is not the same as applying it. You apply it when you click the "Apply" button in the panel in the stack (it can also be applied temporarily during collada export). When you do that, it's effects are carried out, instead of just being displayed as a preview, and the modifier disappears from the stack. From then on. changes in either part will not be mirrored. BVefore applying it, you can't separate the mirrored part into a separate object, because it isn't really there - it's just being previewed. After applying the modifier, it is really there and therefore you can separate it. Because "Apply" is used by the relevant button, the term "Apply" is used to indicate this final application of the effects of the modifier. So adding it to the stack without clicking "Apply" is not applying it. I hope that's clearer. I guess it's a pity they chose "Apply" for the button, as intuitively it could equally reasonably by used for adding the modifier to the stack. "Finalise" or something might have been more clear.

"So you do not need to apply the mirror mod before you have finished your mesh" I'm not sure what you mean there. You can apply it whenever you like. Before you apply it, changes made to one of the mirrored parts will always happen in both (which is the whole point of using the modifier). After applying it. they will not. Before applying it, you cannot separate the two parts ito separate objects. After applying it you can. Whether you apply it before or after finishing it is entirely up to you. If you want the shoes to be exact mirror images, then you cannot make any more changes aster you apply the modifier. If you want some differences that are not mirrored, then you have to apply the modifier before making those changes.

"Prim" is short for primitive object. Originally these were simple shapes, cube, sphere, cylinder etc., which could be modified by all sorts of cutting, hollowing, twisting, tapering and so on (often called "torturing"). Everything was made from combinations of these shapes. They work by using a set of numbers (parameters) that control the various distortions, to generate a "mesh" of points (vertices) which defone the shape. Later, a special type of Prim, sculpted prim or sculpty, was introduced where the positions of the vertices were defined by the coulours of the pixles in an image called the dculpt map. These were much more versatile, but had severe limitations. So, finally, "mesh" was introduced, which use much more versatile industry-standard ways to define the vertices. They also provide ways of defining the collision properties of the object (physics shape), and they can be "rigged" so that when they are worn as clothes they move along with movements of the avatar's bones. That is the attraction of mesh clothing, compared to traditional clothing made of prims or just textures. "Fitted mesh" even moves with some of the adjustments that you use to customize the avatar's shape. Technically, sculpties and mesh are implemented as, and are still often referred to as, special types of prim, although cleraly the word "primitive" isn't really appropriate any more. The apperance of all kinds of prims is finally determined by application of textures. These used to be just simple images, but can now be supplemented by two additional kinds of image, normal and specular maps, that control the bumpiness and shininess of the textured surfaces. The combined images are called "materials", but seeing their effects depends on turning on the advanced lighting model in graphics preferences. That is quite demanding on the capabilities of your graphics card.

Ah yes. I think that's called z-fighting. It's horrible. Different from the alpha-glitch. I didn;t think it depended on alpha channels at all. I don't know what you can do about it other than moving the layers slightly apart.

Can you explain exactly what you mean by "conflicts"?

Not sure if this is what the problem is, but... The mirror modifier doesn't really create the mirrored geometry until you nexplicitly apply it by clicking the "Apply" button in the modifier panel. When it's applied, the modifier panel disappears. Until then, the 3D view just displays it as it will appear when it is applied. If you click the icon that makes it active in edit mode, then it even seems thast you can edit the mirrored version, but you can't, as it is just editing the original. You wil see all changes made to either also occur on the other. Since there is no real mirrored geometry, you can't separate it (e,g, by using P). You have to apply the modifier first. After that, the mirrored geometry is actually generate, so that you can select it (all its vertices, In Edit mode) and separate it by using P (or Mesh->Vertices->Separate:Selection from the menu - the menus are a very good way to discover functionality and usually tell you the keyboard shortcuts if you want to learn them).

I agree. Since the diffuse isn't mapped properly, it has to be missing or incorrect UV map. If the collada contains multiple UV maps, SL will use only one of them (the last in the file in my experiments, but I can't say that's always the case). Vendetta, you need to tell us which software was used if you want instructions for removing the wrong UV map.

"Also Intercontinental boating should be made possible.! If there are oceans, there could be ocean currents too, and hurricanes, and a Bermuda Triangle. And on land, let's hope there will be proper terrain instead of just heightmaps, so we can have caves, tunnels and mines.

In case you haven't already, you have to take the IP questionaire on Aditi as well as on Agni. For some reason, they don't work on both grids at the same time.

It might help if you put an image of your normal map (reduced if nevessary) her. Meanwhile, with the default settings for evrything else, the effects of normal maps are often hard to see, especially if there is a didtracting diffuse tecture. S if you want to exaggerate the effects, st the diffuse texture to blank, with whatever color you want; set the specular map to blank and then diak up the environmental slider to >150. Then switch between your normal map and the blank normal map, or look at a duplicate mesh with blank normal map. If you don't want the environmental shading, then dial up the glossiness slider to b120 and put some simple local lights next to your prim. Mover the lihjt(s) around to see the specular highlihjts move, and compare with what you see with the blank normal map. Learning how to get good effects in both daylight and local lighting can require a lot of experimentation. Oh, ad by the way - you won't see any of these effects unless you have enabled Advanced Lighting in the graphics preferences. PS. If your normal map has an alpha channel (specular exponent map), you will need to make sure it's completely white or has some bright areas in order to see the specular highlights. A missing alpha channel behaves like a completely white one.

Thanks, Arton. You saved me from having to write something similar. I think most of the discussion here is failing to take account of how different the new world may be, not just for mesh and materials, but for everything else. No doubt the blender files etc that people have will be useful starting points for generating content suitablr for the new world, but I would be very surprised if they did not require very substantial modifications to fit in with the new technology. We won't know until we have much more technical detail, especially of what the engine is and how the system will deal with resource allocation/conservation. Right now, those things are probably still in the discussion/research phase,even inside LL, in which case it will be a long time before anyone really knows what the content transfer possibilities may be. ETA: Oh. I did write somethging similar after all! Never mind.

"We will get the ability to upload custom skeletons (rigged meshes)" Gaia, can you say where that comes from? In particular, do myou have any idea whether it applies only to avatars, or to non-avatar mesh? I am thinking about some ideas discussed a long time ago if we could access bones in scripts etc.