Vegro Solari

Knock yourself out. Might want to call yourself something other than "Realtors" though, because unfortunately that's a real world trademark. In sick and demented trademark law theory they (the association owning that trademark) could swoop into SL and demand you and your friend stop calling yourselves "Realtors", and LL would then say to you "you stop doing that please". In practice, though? They don't care, you don't care, LL doesn't care, and you go ahead do what you want, have fun with it! You could be "VirtualRealtors" or something and that way proudly fly the middle finger all the way.

Glad to help! I only get to look at the forums rarely, but in case of needing a bit of consulting about something, feel free to IM me in world. The mesh upload part (the SL Mesh uploader) is a tricky business in itself, but I eventually worked out a workflow for it too. The thing to realize is that while you are in ZBrush, just think about making a finished, lowpoly model with great textures that capture all the highpoly details you sculpted. This is going to be your "Highest" LOD. Which lower LODs you will need to work hard on and which can be left to automatic or skipped is going to depend a lot on what type of item you're building, its actual dimensions in-world and how far away will the viewer usually be. It's sort of a juggling act of its own that you have to get to grips with, that isn't ZBrush-specific. I'm staying away from trying to export LODs directly from ZBrush, because so often that ends up being counterproductive. The best I found is to focus on the "Highest", then depending on what's needed, you could reload that lowpoly (with textures and UVs) back into Zbrush and experiment with Decimation Master using the "Keep UVs" option turned on. This lets you generate lower poly versions for specific LODs (there are only 4 in total to worry about so it's not so bad) that you decided are important to your item. For a piece of hair, I'd personally use the same model for the Highest and the one below, and then either send in my own Decimation-Mastered versions for the lower poly LODs (halving the number of faces with each step is a pretty good rule of thumb), or even just be quick and use the SL automatic decimation for those lower LODS. Since it doesn't matter so much what the hair looks like when it's seen from 1/2 a sim away, it's an item that must look best at up close and medium distance.

For trimming off unwanted faces, you have a ton of options. My favourite is using ctrl-shift and drawing a polygon selection with one of the tools (box, cicle, lasso, curve etc) then ctrl-shift click to isolate it so the rest of the model is visible, while the part we want gone isn't. Go Geometry->Modify Topology->Delete Hidden. This gets rid of those points. If you are still in Dynamesh, you can just ctrl drag on canvas to redynamesh and it will close any holes neatly by itself. There is a Close Holes command in Modify Topology that you can use. You can also use masking, to protect the part of the model you want unchanged, then a clipping brush or clipping curve to smush those "bad" faces into shape. You can also use those new trimming brushes, which cut the mesh, then automatically resulting close holes in one step. Finally, you can pull up a Slice Curve brush, and slice off a part of the model as neatly as you want, which creates new polygroups there. Do a ctrl-shift click like before to isolate, Delete Hidden, and then close hole, (or redynamesh). This lets you get really nice smooth cuts because Slice Curve creates smooth new edges around where it cuts automatically.

There is definitely a way to move the UV seams to wherever you want. In ZBrush, you would do so with the UVMaster plugin. Watch a Pixologic video or any short tutorial about it and you'll be up and running. The idea there is that you can attract seams to areas which are "darkened" by ambient occlusion, and/or also freehand paint areas that you want to protect from seams or attract your seams to. These don't need to be precise at all, just a general ballpark "line" painted on will do fine, since the UVMaster process is still fully automatic. That's the way you can get your UV seams to go exactly where you want! Btw, sculpting hair and clothing was one of the things that set the masters and the amateurs apart in ancient Greece classic sculpture So don't give up on it, you can get amazing results the more you practice! For your final result lowpoly, you can use this step by step ZBrush workflow: -------------------------------------------------------------------------------------------- Sculpt your model using all your favourite Zbrush techniques, Dynamesh etc. You get something with millions of polys. Decimate to about 15-20k points (or whatever is enough to make sure you keep almost all your detail in the model) Run zRemesher on it Run zRemesher a second time on this, and see if you get an even better result! (often you do) Use UVMaster to automatically create UVs for the lowpoly model (this is where you decide where the seams will be) Subdivide the lowpoly model until you are at approximately 4 million polygons again Use polypainting features, Spotlight, etc. to colour it While on the highest subdivision, go Texture->Create from Polypaint. You will see a texture appear there in the box. Press "Clone". The texture pops over to the Texture palette. In the texture palette you can flip the texture vertically (needed for SL) with a button. And, it's here that you can export it as a PNG, to start working on it in Gimp or whatever you like (if desired). Go to the lowest subdivision where your lowpoly mesh is, and export that as a collada file for Secondlife. Done! Seems like a lot but you'll zip through it after only a little bit of practice. This workflow is designed to introduce UVs only at the last possible stage, because you don't want them around when you're sculpting your millions-of-polys mesh, they take up resources and slow you down. Another alternative way to work is to sculpt, polypaint (all without UVs), then create a lowpoly mesh as I described, and give it UVs. Save both the highpoly and lowpoly meshes. The highpoly will have polypaint information in it and no UVs. The lowpoly will have UVs but no texture yet. Use the free application xNormal to bake different maps including ambient occlusion and normal, and colour, to the lowpoly, from the highpoly polypainted one. Extra tip: for zRemesher, when "adaptive" is turned on (default), it's trying to maintain quality as hard as it can, which can be not what you want, because it then is allowed to increase your polycount. If you disable this or dial it back, you can force zRemesher to give you the best result it can for specific polycounts you're aiming at (like we do for Secondlife!).

You're welcome, Peterson111!

One way you can take this same basic idea a step further is by applying it to cyllinders and tubes. Given the exactly correct ratio, you'll be able to get really sharp texturing on SL cyllinder prims, without any of that usual nasty stretching visible. The quickest way to do this is to just rez up a cyllinder and set its height to 1. The "unwrapped width" of the cyllinder of height 1 is then going to be super easy to calculate, it's 2 times 3.1415, or 6.283 secondlife units. Now you can just use the above tip, and start out with a texture in your graphics program that has a pixel ratio of 6.283 to 1, and get exact reproduction of what you draw, wrapped round the prim cyllinder. Just for example, let's say we wanted to use a fixed height of 256 pixels for our cyllinder texture, and now need to find the proper width so that it'll conform to that ratio we just discovered. That's going to be easy to find, by just multiplying 256 pixels by 6.283, which gives you 1608 pixels. Now you can start a blank texture sized 1608x256, and when it's done, treat it like any other custom-aspect ratio texture and follow the ideas/steps from the previous post to upload it correctly. Happy texturing! Handy resolutions table to use for crisp and distortion-free cyllinders textures: 1608x256 3217x512 6434x1024 (but of course you can use -any- resolution you want by doing a quick calculation as already shown)

You can easily get much nicer quality in uploading images to SL if you take a moment understand this tip. You know how whenever you upload an image to SL, it gets force-resized to 512x512, or 1024x512 , or other "degrees of 2" pixel size? The force-resize that gets used is not very good quality-wise, and even worse, it ruins all your image proportions, so stuff looks stretched out and wrong when it's on a prim in SL. Also, those numbers mean that the only aspect ratio you get to use with optimum quality is the clunky 2:1, or just 1:1. That's not a happy situation for art and photography, where 3:2 and other custom ratios are important! There is a simple way around the limitation. Follow me in this tiny example, where let's say we have a piece of art that is 1200x800 pixels and we already know it won't fit into any of the predefined ratios, but you still want it to look good in SL. What's the fastest way to get it done? 1. First, use your graphics program to resize the art to a resolution you think will be enough for it (512x512, 1024x512 etc, 512x256 - it depends on your art's size in horizontal and vertical, we just use whatever's closest) , this lets us bypass the crappy-quality force-resize on the Linden server. 2. Save a copy of this resized version and add the original resolution to the filename. E.g.: "Joyful Maenads (1200x800)" 3. Now we just upload this resized version into SL, and here's the main trick. Rez up a cube, but instead of hand-sizing it for the art, we're going to set up the actual exact pixel ratio we wrote in the filename. Default cube's dimensions all come as 0.5, so we go into its editor window and type in: X: 0. 1200 Y: 0. 0800 Z: any desired thickness As you can see, you can just put "0. XXXX" where xxxx is your original pixel resolution for width and height. This gives you a small cube, but with the exact dimensions you need to correctly display your artwork with the custom aspect ratio. Then, just use the normal grey sizing handles on the prim to size it up or down as you wish, all the while automatically preserving the correct ratio. That's it! You will get a nice looking frameable result with the exact proportions from your actual art piece, while also being in complete control of whether you want to add a bit of sharpening or other adjustment after you resize (in step 1) and bypassing the bad automatic Linden resize.

I did it with Phoenix 1.5.2.1185 and it did export my shape as an OBJ file. You can get this particular version still even now, in the old versions part of the Phoenix site. Once you get it, you need to enable the advanced or developer mode to find the export menus for the avatar.

Dr. McMahon in the flesh! Thank you once again for your brill work unravelling these weird meshuploaded mysteries, sir! Honourary forum doctorate should be the least we could do for you.

This is one of those things where in theory you'd think about it one way, but practice may demand something different entirely. If what you are making are generic game models and you enjoy spending your life on perfecting manual optimization for these, all the power to you. But if instead you're making SecondLife products, then the fact that you can eliminate the subconsciously jarring LOD popping effect from your prop or clothing piece (and some of them are so sized & positioned as to make them very vulnerable to it, oscillating between High and Med LODs for almost all of their onscreen life) is going to be the overriding consideration in your product awesomeness equation, not the miniscule poly efficiency issues.

Work done by the Zbrush mat baker plugin is going to be based on projecting Polypaint, which means that your model needs to have a beefy pixel-polygon ratio (lots of polys, think 2million +, 4million if you want 2048pixel textures). Could that be part of the problem you're running into? If polypaint doesn't have enough polys, it's going to wash out into "pixellation". Getting more polys as usual is easy enough, tap Ctrl-D until you're in the ballpark desired. If still stuck, you could also look into using ZApplink projection baking, which will capture several views from the model and let you actually manually bake the shading and material layers in Gimp/Photoshop. This manual approach, of course a bit more tedius, but allows you extra fine control because you can actually grab the layers Zbrush sends out (when in Photoshop) to a temporary psd document, and that way be able to composite any number of material passes together, with the added power of photoshop blendmodes and filters to help you tweak the look to perfection.

Basically it used to be all called "prims", but it was changed and it is all actually based on Land-Impact now. But for traditon's sake and ease, lots of people just continue calling it "prims". Even though what they really mean is always Land-Impact.

For clothes, you can usually get away with only optimizing the High and Medium LODs. The other two can be ignored / staffed with a basic box or whatever. In many cases you might actually gain a better visual look for the clothing if you set the High and Medium LODs to be the same exact mesh (that's at the cost of some LI of course, but since for clothes the LI is a lot less crucial/noticeable to user, it's a useful shortcut to keep in mind.)

One optimization trick I learned (major thanks go out to some of the scientific pioneers of this art, here in the forum, who discovered and explained it) is that when you're uploading the model, your landimpact is going to vary heavily depending on the in-world scale of the object - everyone knows this - but usually not everyone realizes that each of the different LOD levels you can use is going to have a vastly different "weight" in the land impact calculation, again depending on the size of the object! So that means, sometimes the "Low" LOD is the most important, and sometimes it may not even matter at all, and so on, all depending on final object size in-world. It's good to get your head around that, because that means, depending on your object's intended size, you have to really optimize one specific LOD level for that size of object. That way all your work that Fraw has shown you, is going to have the biggest effect on reducing the LI and you can pretty often even get away with slacking and being pretty quick with all the other LODs. Because maybe in your case, they just don't have that much of an impact. Example: you can discover that, say, for objects less than 1 meter in size, the dominatingly important LOD to optimize is the Lowest one. Somebody's published all this, with graphs and usage ideas in another post in this very forum, just search for "optimization", and have courage, because it's really not as hard as it looks to use this and get awesome land impact reduction for your creations the smart way. LindenLab also published info on some of the simple formulas used in SL to see at what distance what LOD model is going to be used. They are all really simple, just based on the size of your mesh object ("object radius"). For example, the lowest LOD model will appear, according to those 4 formulas, when it's viewed at Radius / 0.03 meters. This is not as big for a sign, just because you probably want your sign nicely visible from all distances (meaning, you need all the LODs), but still, kinda keeping these things in mind lets you decide even better what sort of LOD models you want to use for your object, besides the primary high detail one. Good luck with it!

Hi! There's a semi-hidden feature called "Local Texture". You find this under your texture select, there's going to be a radio-button to call up Local Texture. With this, you get to select any PNG file on your computer, and you can preview what that's going to look like directly in SL without uploading it first. Overwriting the PNG file by saving it from Gimp or Photoshop, makes it instantly update in your SL window. That's real handy if you work with textures lots, you see a live preview. But only you can see it, everyone else will get a grey prim. There's no actual way around the upload cost of 10L per texture, it's kind of like the Linden Lab tax. Good luck and have fun with your business! PS: Also, if you want to make some Linden on the quick doing a few cool textures for me, send an IM. Original message poster only, please, 'cause I can't accomodate everyone who might be reading this post years later and this is just to help out.