Kwakkelde Kwak

I don't know the first thing about Blender, but the only thing I can think of is a combination of size, the way your model is rotated (pivot..as in does it rotate around its centre, the grid centre etc.), near clipping and unit setup. Any of these can cause clipping.

The thing is I don't use a phone line I understand how the up and down will vary (slightly) if you take a long enough route, but it's the speedtests that simply give numbers I cannot and do not trust, at all, see the post above your latest. Especially the ping is just incorrect.

Of course, but the ping times just make no sense, more than the up and down. If I run a tracert to SL I get a ping of 110 or so, in the ctrl-shift-1 menu I get 150-170. That makes perfect sense. What doesn't make sense is that when I run a speedtest, I always get a ping of over 200, not just to SL, but even close to home. It's so bad that even my ISP which is pretty much around the corner, returns a ping of 205. That is when I have SL open with a 5000 mile ping of 150 ms according to the ctrl-shift-1 thing (and also when SL is not open). How can the ping to my ISP take longer than it does to the US? When I do a direct test to my ISP I get the same results as the tracer gives me, a ping of 1 ms.

A megaprim is a prim bigger than 64 meters. You can find them here for example. Just type in a size like 1024x1024x40 and the website will find the closest match. On the marketplace you can look for the landscapes by using "sim surround" as search keyword.

SeraphInquisitor wrote: So the workflow I should be going for is.. Model > Unwrap UV > Make different copies and re-scale/re-fit them to the other sizes > skin > export. Right? Pretty much, I'm not all that experienced with rigging to be honest, but I suspect you can re-use the weights. That would mean you can model (around the most neutral avatar size) -> unwrap -> skin -> refit to other sizes (rescaling is a tricky choice of words here, although you can rescale at vertex level) -> export If you can't directly do this I am pretty sure you can make a copy of the weight map in the skin tools. The vertex amount and order won't change as long as you just move them around, so it should be a copy paste kind of situation. Another question. I tried making a test dummy avatar (boxes and cylinders), unwrapped and skinned, export to .DAE, and it worked as an avatar. Then I tried making an attachment, just another box, rigged to random bone (Left elbow in this case), followed the same exact steps, and when uploaded on the viewer, the skin weight option is greyed out. What changed? It could be missing bone names in the dae file. A complete avatar uses all 26 bones in 3ds max, but a box tied to one bone does not. There are multiple ways to add the bones, all explained in the wiki I posted earlier. I haven't had any luck with the "bonefixer script", so you might want to try the other options.

I'm pretty sure those settings are local, so it will only look brown and flat to you. Others visiting the sim will see the standard water.

I'm not a big fan of them, but since you're on an island, you could use a megaprim. There are entire landscapes available on the marketplace, or you could just use a hollow cylinder twice as large as your sim and texture the inside with a cityscape. When people move their camera up and away it will be very obvious, but when standing on the ground it gives a very believable effect.

Going by your question you're especially interested in rigged mesh, so first you need a rigged avatar with skeleton. Standard sized armatures you can download here. There aren't any .max files included, so you will have to import the fbx files. Because the mesh is rigged to the bones and not the avatar skin, one size won't fit all avatars, at least for now. The same is the case for these five sizes, but you'd certainly reach a bigger group with these. For smooth deformations, I'd copy the skin weights from the avatar to your build, then paint in the details. In general that should be it I suppose, you could spice up the avatar with a free skin from this site.

Give it a few weeks, as I remember the lower end 600 cards had compatibility issues when they came out, I'm sure it will be fixed, either by LL or NVidia. A shame though.

Nalates Urriah wrote: Make sure multi-threading is on. Older advice around SL is that the viewer can't use multi-threading. That advice is way out of date. Could you explain how to enable it then? This is my typical taskmanager view when running SL: I have four cores with two threads each, so it looks like SL is using only one thread of only one core to me.

You don't get any extra options for editing your avatar when you upgrade to premium. You'll be able to contact live support and buy mainland plots, that's about it. I think your hair is an attachment. The sliders you are using are for the default hair on the avatar, it's part of the default avatar and it's something pretty much nobody uses. If the hair you are wearing is editable, you can change the color and texture by right clicking it, then selecting edit. In the pop up menu you will get some options. Attachments are objects, they can never be changed by changing the sliders you mention.

If you see things loading, I suspect something in the SL internals, not your own computer, it's something I haven't experienced myself, but heard of recently. Let's hope it's a bug that will be fixed rather than the inevitable result of some permanent changes. Like you and Jean say you're at the ceiling of SL performance, but you might want to try a couple of things anyway: One thing SL seems to like is a very fast CPU. Unfortunately SL doesn't have a clue when it comes to multi-threading. I always see one thread loaded, the other 7 idling or doing something else. If you want to upgrade something in the near future, you might want to overclock or replace the CPU, as high end as it may be. One with( fewer but) faster cores could make a difference. You can clean and increase your cache. You could check the cache hit rate in the Ctrl-Shift-1 menu to see if the reloading items are actually read from cache at all. You seem to have plenty of RAM for SL, 2 free GB should be plenty, so you could set up a RAMdrive for your cache. I used to have one instead of caching on the SSD. In theory that should be a lot faster, but I didn't notice any difference to be honest. I really do not think it's your video memory, but there are programs out there that monitor your graphics card. I use Asus GPU Tweak which will show how much video memory is in use. I have never seen my 2GB maxed out, which isn't a surprise considering the maximum 512 MB used for textures. Never seen the gpu load maxed out because of SL either btw, unless I turned off VSynch, then on a very empty sim or in a skybox the fps would shoot up to over 200, maxing out the gpu. It would be nice to hear if the 780 makes a difference...

Or if the rendervolumeLoD isn't manually set over 2.0, use the object detail slider in preferences to avoid the debug settings. It might not go as low as 0.0 (not sure), but will definitely give you a workable LoD range/camera distance.

Mara Inkpen wrote: 1) SL is not a game not only conceptually ( its a world), but also technically: its engine is not directx-based, like almost all modern games, but opengl-based (like professional graphical software) - thats, for example, why we cant play SL in 3D glasses, while we can do it with most of the games (http://en.wikipedia.org/wiki/Comparison_of_OpenGL_and_Direct3D) and maybe (maybe?) quadro cards can work with openGL better Even though SL uses OpenGL rather than DirectX, it's still more of a game than anything else as far as architecture goes. I can't be sure since I have no Quadro to test with, but I think the car comparison would work again. Quadro being the semi, now fitted with tires which perform very well on tarmac. On the other side a race car, with tires designed for sand. Now put them both on a track, I think you get the idea. Quadros excel in certain areas, areas common in professional work, where precision and reliability are the main factors. If you do something "3d workish" in SL, such as editing an object so the wireframe shows, especially a high poly one, the Quadro would probably do better. When just walking around, or flying in your case, you'd probably need a GTX680 hardware equivalent Quadro, so that would be a K5000, which costs well over $2000.

In addition to all this, if you want to examine other people's work, there's another "trick". In the develop menu, under "show info", check the "show render info". Then select the object you want to examine and zoom in. In the right lower corner of your screen you will see the number of triangles for the object (kTris, or 1000 triangles). Now zoom out to see how the number goes down when the object switches to simplified models. There are up to four levels (of detail), depending on the size of the object. This will give you a good indication of how the builder uploaded the model.

yw, glad it works now!

Did you download the 2013.2 or 2013.3 version of the fbx plugin? V2 always works for me. I think the 2013.1 version had the same issues as all 2012 versions. There were also people having issues with spaces in material names, so try to avoid those. Also put a material on the object. I don't think DAE files include any ID information, just material information. So if you don't have any assigned (or not all of them) you could end up with missing selectable faces.

I was just referring to the single triangle approach you see around SL. I didn't mean there was no right way of doing things btw you could really save on geometry on the higher LoD, stop thinking in loops (above the bars). You could bring it down by 88 triangles pretty much without losing shape. Nice little bench though.

JG Hotshot wrote: Just out of curiosity, what does the "ti" mean? From what I read it means Titanium, which is just a name. In reality it means it's an upgraded version with more shaders (chips) enabled. And you're welcome!

XxiMarshmalloWixX wrote: i have see many sofas, mesh hands, necklaces with many small parts on them and more that when i drop them says 1 objects selected - land impact 1 At this point it's a guessing game. Could you provide some more information? Preferrably a wireframe model of what you are trying to upload, your upload settings and the resulting weights.

A tutorial won't do a lot of good, unless it exactly describes the object you want to build. If setting all lower LoD models to "triangle limit" zero (or uploading a single triangle for the lower LoDs) doesn't set your LI to 1, either the model is too big, the model is too complex, or it's not the download weight causing the higher LI. Did you check the three weights in edit mode? Download, physics and server? btw, you say "1 prim only", do you really mean "1 prim only" or do you mean "a LI of 1"? There's a big difference.

It's very simple. For mesh items you upload five models. One for each level of detail and one for the physics shape. The highly detailed objects with a very low landimpact have a highly detailed model for the highest LoD and very simple (could be as little as a single triangle) for the others. On an item the size of furniture, the medium and low Lod probably have the biggest influence on the landimpact. Depending on the size, a detailed model could be uploaded for the medium LoD too. The downside is the models will look horrible from a distance. This might not be noticable on indoor items which you never see from a distance, builders can and should exploit this situation to keep the load on systems down, side effect is a low LI. (You could view it the other way around too.)

A 550 Ti sounds like a good alternative. It's benchmarked slightly higher than the HD 6770, costs about the same and supports a PCIE 2 bus which is probably what's on the motherboard. The 6-series costs a lot more and if it runs on the slower bus, the performance difference might be marginal. The power consumption of the 550 is in the range of the Radeon 6770, with a 400W power supply as minimum.

One other thing that crossed my mind, is everything in that picture on one UV map? If you have multiple materials (multi-sub), 3ds Max doesn't understand the multiple ID's. In that case you should either combine everything into one material, or split the object into multiple objects, each with their unique material. I never understood why a program like 3ds Max doesn't bake multi-subs, but that's just the way it is. When you have your occlusion maps baked, you can combine the seperate pieces into one object again. The suggested resetting of scale and such is "Reset XForm" in 3ds Max. It's in the utilities tab on the far right, the little hammer. After that, turn the object into an editable poly again and check the orientation of the normals.

Two things I can think of... One, bad UV mapping, which would be easy to spot so I don't think that will be it... Two, maybe inverted normals? I always turn on backface culling in the object or layer properties, that way you can see if the normals are facing the correct way. There are more ways to check the normals, but this is what i use.