Fluffy Sharkfin

Since texture animation works on all texture channels you can make some very nice running water effects using a seamless fractal noise generator to create normal/specular maps & alpha channels, etc and tweaking the UV mapping on your mesh to control the speed/direction of the flow of the water.

All very valid points and I'd agree that the onus is on LL to fix the issues with the lack of incentivization for creating more efficient content (and conversely the lack of penalization for those creating poorly optimized content). That being said the fact that SL content creators are, as you put it, so "good at peddling snake oil" and are often looked upon by their customers (and SL residents in general) as experts in content creation, I believe they do share some responsibility when it comes to the information (and sometimes disinformation) that they disseminate. A large part of the process of improving the optimization of content uploaded to SL is going to involve providing residents with much better information on the best (and worst) practices involved in content creation, and the difficulty in doing that will depend a lot on how many differing opinions users are exposed to. If LL are saying one thing and various content creators are saying something completely different, it's just going to lead to confusion and the propagation of more bad practices.

The fact that the additional tails only became visible after being taken into inventory then re-rezzed is curious, and lends credence to the theory that it's using texture animation somehow (since texture animation is an object property that persists after the script that initiated it is removed, but can be broken when taking the then unscripted object back to inventory and re-rezzing it). Sorry to ask a really obvious question but you do have "Select Faces" selected in the build window and you've selected the tail face and not the face that corresponds to the body rather than the tail right? It's just the texture in the preview window doesn't look compatible with any type of texture animation and it's hard to imagine what method they could possibly be using to do alpha switching on the tail given what you're describing and the screenshot you posted.

I'm not sure I follow your math here Qie, or maybe we're thinking of slightly different methods. To illustrate here's an example tail alongside the uv layout and corresponding texture. As you can see each of the four tails use approximately 1/4 of the texture space (the other 3/4 being transparent). Adding a texture animation with the PING_PONG flag set would cause the tail to appear to move from side to side as that opaque section of the texture lines up with the uv mapping for each segment. Of course the limitation of this method is that you have to use the same texture and uv island layout for all versions of the tail and only modify the vertex positions of the mesh rather than having individual textures and uv layout for each version of the tail.

Some additional information on how exactly you "broke it" may help to shed light on the method being used to animate it.

It's hard to say for certain without inspecting the object in question but the most likely explanation for what you describe is: The body is one material and has a regular texture on it. The tail(s) are all sharing the same material/texture, but have different uv mapping, with the UVs of each tail being located in a different quadrant/section of the texture map. The tail texture appears in only one section of the image, the rest of the texture is transparent, and the tail "animation" is achieved by changing the offsets of the texture on the tail(s) to change which one of them is currently visible (most likely with the use of llSetTextureAnim). It's a crude but reasonably effective way of simulating animation, but before you use this particular method I'd suggest taking a look at animesh. It allows you to create far more realistic and efficient animated content than the various forms of "alpha swapping" that are currently in use.

The basic principle for creating LOD models is pretty much interchangeable between most traditional 3d apps so assuming you're familiar with the toolset in 3dsmax then your best bet is just to find one of the numerous tutorials in this sub-forum on how to make LOD models for SL mesh and follow along, ignoring any specific information regarding the blender toolset (or whatever app the tutorial is based on) and substituting the appropriate 3dsmax tools. Here's a tutorial created by Penny Patton that covers the basics and should help you get started Optimization Tutorial: Creating Your Own LOD Models! I'd also recommend searching the forums for posts related to 3dsmax just to see if there are any specific limitations that you need to be aware of or recommended workflows you should follow when preparing your meshes for upload.

The person that posted that image has already said that he did so because that is his personal preference style-wise. He then provided further examples of less stylized and more realistic character models (all of which you seem to want to conveniently ignore so you can harp on about the first image and how much you dislike the aesthetic). As far as animation goes, if you don't like that particular animated gif I suggest you follow the link I posted and read some of the articles, even better do a google search and learn a little about what IS possible rather than just declaring something is impossible simply because you haven't seen anyone do it properly in SL yet. The ceiling for quality of content is NOT defined by "the best stuff you've seen in SL", and if you try to argue from the position that it is you aren't going to win the hearts or minds of anyone that's taken the time to research what CAN be done.

Sorry, I'm not going to explain the meaning of the word "example" to you, go google it!

No. What some people are saying (and others are vehemently disagreeing with) is that you don't have to use hundreds of thousands of triangles to create a good looking character mesh that can be animated realistically. So far the only mention of using stolen content has been in your own posts. ETA. As a matter of fact if your aim were to create a mesh avatar that animates realistically in SL the very last thing you should do is try to use a character mesh from another game. The topology would be designed to work with a completely different rig which would make skin weights an absolute nightmare to get right and would still look terrible compared to a mesh that has had its topology designed to work with the SL avatar rig!

Hi, welcome to the thread! Since you seem unfamiliar with the topic being discussed let me fill you in. It's a discussion about what content creators COULD do to improve their content without having a detrimental effect on the visual quality of the content they're creating. If you're looking for threads that deal with the subject of the morality of creating original content vs stealing content from other games/platforms you'll find there are plenty of those too. Alternatively you can attempt to derail this thread and discuss the subject here, in which case I'll probably end up agreeing with you a lot because I'm very much opposed to ripped content in SL myself.

It's meant as an example, rather than literally that specific mesh!

The problem is that you seem to be arguing that having an avatar that looks like this... and moves like this... is somehow going to adversely effect your SL QoL. If you can look at these two images in comparison with most SL content and honestly say that you think SL content creators have no room for improvement when it comes to optimization and rigging/animation, then I guess we really are living in two completely different virtual worlds!

The point is there's a lot of poorly made content in SL. Those that know better keep trying to educate those that don't as to what makes "good content" in the hopes that once they're better informed they'll start making smarter purchasing decisions, thereby forcing creators to start creating responsibly. Unfortunately a lot of the people that don't know any better seem to think that because they spend a lot of money on "bad content" their opinions on what constitutes "good content" should carry as much weight as the opinions of those that do know better, which will never happen because those people... know better!

You don't need a high polycount in order to achieve realistic movement when animating a character model, what you do need is a good underlying topology and an understanding of how to use skin weights to simulate realistic muscle movement. The model in the images below has a lower polycount than most mesh avatar meshes you'll find in SL and no high quality skin texture yet when animated it manages to convey more realism than any high poly SL mesh body I've ever seen. (Source: http://wiki.polycount.com/wiki/ShoulderTopology) No amount of triangles will compensate for a poorly created animation or pose, the only solution for that is to stop buying animations from animators that don't understand the concept of joint constraints.