Tutorial for making SL animations with Blender?

Is there a tutorial for making SL animations with Blender? I searched back a bit and found a perl bvh exporter and some munging to make SL happy, but I'm not sure where to go from here. Do I need to get a reference file with the skeleton to animate from somewhere?

I'm a GNU kinda guy, so I'd love to be able to use Blender for this.

To my knowledge, there is no tutorial that is specific to making anims in SL with Blender. I searched for some basic tutorials on the program and found some good ones at:

www.blender3d.org (user manual was the most helpful =/)

www.10secondclub.org/users/juicy (very basic and easy to understand, not a huge amount of info though)

I also found another site that had an e-book that was free to use, but I can't find the site now. I'll look again and add it if I find it.

As for the bone structure and joints and such, I have no idea. It's why I haven't tried using Blender yet. I think somewhere in the forum the bone structure is listed, but damned if I can find it.

I know, not much help, but ya gotta start somewhere!

I'm pretty well set for use of Blender - I already use it for various other things. Problem is, I'm not sure about how to go about doing anything useful for Second Life with it. There must be a reference skeleton to animate or something. Hopefully someone out there knows how!

In theory, anything that can export BVH files should work for SL animation. The thing is Poser's figure models are already rigged properly. You could certainly make one from scratch in Blender if you want, but Poser is so cheap by 3D standards you might as well make it simple and just get Poser (unless you are strapped for cash). Besides, it never hurts to have an extra 3D animation package on hand. Some are inherently better at certain tasks than others.

Anyway, if you're dead set on using Blender, I would suggest downlaoding the Poser free trial and taking a really good look at how the default skeleton is rigged. Make sure you replicate every joint, handle, attribute, etc. and you should be all set.

Just keep in mind you can't skimp on any detail. SL is expecting the exact data set that Poser would produce. Make absolutely certain you have an exact copy of every node. As long as you have all the same info, you should be fine.

Actually, the theory that anything that can export BVH files should work is not true. The reason for the Poser/SL connection is that the Poser model's bone structure and naming cloesly matches that of SL, though only if you use the old Poser 2 figures - you can't use Poser 4/5 figures and import them into SL (the minute you pose any bone that is not part of SL, the import won't recognize the file, which is annoying). I have not heard of anyone successfully using anything else except Poser, have you? I tried Character Studio, and had no luck. I can't imagine something like Blender (a horrible horrible program I think, btw) would work.

It's a philosophical/pragmatic thing for me - I try and ensure the only non Open Source / Free Software I need are games. If it's possible to do it with Blender, I'd love to know how. Otherwise, I'll pass when it comes to doing character animations unless I find the time to make it possible with Blender.

In other words "I'm broke and a cheapskate"

It's okay, Kral. I am too. I can't even afford the $109 for Poser Artist.

Anybody want to donate to a startving artist?

Okay.. a well-fed wants-to-be-artist?

No, I'm a well paid programmer that deals in Free Software (capitol F capitol S, not 'freeware') as much as possible, as it just plain makes sense to do so.

The most obvious example is in the embedded device / handheld market. Compare WinCE (and the Compact Framework), Palm, BREW, and the sparse set of applications and libraries that had to be written specifically for those devices to an 11+ CD set of Debian GNU/Linux compiled for any those embedded processors - the same software you'd be using on your desktop or server - all with modify + redistribute rights, no NDAs, every development tool imaginable, and you can talk directly to the people who wrote it.

Now tell me, which path would you choose?

As a programmer, I also like what Free Software gives me - the ability to make it do what _I_ want. You might have had that experience in Second Life. It's neat when people give you modify rights on their objects, isn't it? It sucks when they don't, right? Like when you want your outfit a slightly different shade, but you can't as you lack modify rights? Or you want to shoot rutabagas rather than watermelons?

See, you're on my side on this already.

It's better to work at getting all the functionality you want into Free Software, as it will go anywhere, last as long as people use it, and let you grow in the direction _you_ choose.

Isn't that exactly what I said?

Let me restate once again for those who only like to read the first sentence: In theory, anything that can export BVH files should work for SL animation, but the advantage of using Poser is that its figure models are already rigged properly for SL, and so if you use any other program you will have to make sure to rig the skeleton EXACTLY as Poser already has it set up.

There, all one sentence. You should have no trouble following it now.



As I stated before, download the free trial of Poser and take a really good look at how its characters are rigged. Replicate this exactly in Blender and you're good to go. If you don't know how to do that already, I suggest you either suck it up and join capitalist society (in other words get Poser) or else follow your own advice and stay away from animation. It takes a lot of knowledge of animation principles and of the specific software you are using to know how to rig a skeleton properly.


Sorry, but I couldn't resist pointing out the irony in that statement (no offense intended). Wasn't the whole point of this discussion the fact that Blender is not doing what you want?

Personally I think Poser is a completely silly and childish program. I'm much happier with grown up tools like Maya, Messiah, SoftImage, etc., but I lower myself to use Poser for SL because it already works. Why bother re-inventing the wheel?

Again, please take no offense, but Blender is at least as silly as Poser when compared to "real" packages. In 3D software, as with most things in life, you get what you pay for. If your only desire is to pull programs apart, add to them, modify them, etc. then by all means stick with your open source toys. However you have to be willing to also use industry-stnadard tools if you want to be able to work seamlessly with others. Case in point, the current gap between you and SL animations.

Sure, you could take the time to get any number of other programs to do the job if you were so inclined, open source or no open source. Hell, you could even write your own from scratch if you really wanted to, but why bother when a perfectly viable tool already exists?

Well said, Chosen. Now would someone PLEASE write a .bvh exporter for Max?! Pretty please? Poser isn't a bad app, but it's definitely more of a hobbiest toy than a production tool It's very good at what it was intended to do though... put character animation and rendering within reach of hobbiests at a very affordable price and with a minimal learning curve. High end animation apps are worth every penny they cost. Consider that 3ds max costs about $4000 and that a freelance animator in a good market charges $200 an hour or more... Know any other business where you can recoup your startup costs in a week? You do get what you pay for. Poser is a great value for the money if you're interested in creating art with it, learning the basics of pose to pose animating and function curve editing, or using it for the purpose it was originally created... as a foreshortening reference tool for traditional artists. It's a fun program to play with and, if it entertains you for 40 hours or more, that more than justifies the price of Poser Artist.

I'll have to remember to tell IBM it's not being part of capatalist society.. Also, my own advice was that I'd stay away from doing SL animations until I can make Blender do it (or encourage others to). That's a big difference from just staying away.



No, the point of this discussion was to see if anyone had a Blender tutorial for making SL animations, as I had heard it could be used with SL. I'm not interested in Poser. But to answer your question, if Free Software doesn't currently do what I want, then using proprietary software instead isn't going to help it do what I want. It just means one less person asking for, or working towards, that feature.

It might sound odd to avoid using proprietary software when Free Software could but can't currently do what you want, but that's what powers Free Software projects. E.g., why Ogg Theora took so long to happen - avifile relieved the pressure by letting folks use Microsoft's proprietary codecs. Less pressure, less development. It's never a good situation. The folks depending on avifile suddenly realized they were screwed when they wanted shiny OS/X boxes or cellphones that could play + stream their existing media, only to realize they had no way of making that happen. Only Microsoft can, and they obviously won't.

In the long run, I've found following this path is a big win for me. On many occasions, rather than settle for 80% of what I wanted as I would have had to with proprietary products, I've been able to add that last 20%, often trivially, to get something truly special. I've modified a lot of projects (Mozilla, PHP, etc.) to get exactly what I wanted and benefitted from the ease of integration and automation when dealing with open formats and tools (GIMP, Blender, etc.). I am able to use any hardware platform and still have the exact same tools I am familiar with, and have spent time learning. Also, I can cooperate with a larger group of people, as the formats I use they all will have access to regardless of platform or budget.



"open source toys"? Oh, that can't be good. So WinCE, Palm, BREW, etc. are better than Linux? Or do you just mean GUI applications? Like, umm, FireFox/Gecko compared to IE? Or Ethereal? Or like Blender, which was originally proprietary (so then, it used to not be a toy, but then became one)? Or like a large portion of OS/X? Surely Industrial Light and Magic has produced a toy with OpenEXR, and id Software can't seem to figure out that OpenGL is a toy. Good thing we have you to set us straight.

Maybe proprietary software works better for you. Free Software has traditionally been light on the media side, but its strengths are my own, so even short-term, it works better for me. Neither are toys in all areas. I may not be able to produce Poser's format of BVH file today with Blender, but then you can't write a Windows application that runs on more than one architecture. However, you can improve Blender, but you can't improve Windows. I like having the rights to do as I please and have benefitted from them in the past, and as such, I would rather help see Free Software move into more areas, then abandon it for software where I have no rights.

Eek! Calm down a bit, man. Take your finger off the Defend Against Any Perceived Attack Button for a second. I wasn't trying to ofend you in any way. My appologies if that turned out to be the case anyway. That was not my intention. Let me clear the air on a few things.

First of all, when I used the phrase "open source toys", I was referring to the manner in which you are using some of them, not the open source world as a whole. There are plenty of good open source programs and plenty of silly ones, jsut as there are plenty of good proprietary programs and plenty of silly ones.

That having been said, Blender IS a toy as far as I'm concerned. It is weak. It is limited. It is not currently in a state that would cause most people to want to use it for any reason other than price. Maybe one day that will change through the efforts of people like you, who enjoy sepending their time imroving it, but until then, it is what it is.

That in no way means that ALL open source material fits that description. I love Mozilla. I think Gimp is a great alternative for those who do not wish to spend the money on Photoshop (although I personally love PS too much not to use it). There are countless other examples.

It's also worht mentioning also, by the way, that not all proprietary programs keep the user completely in the dark either. Maya for example can be reconfigured any way you want. If you don't like its interface, it allows you to create a new one any way you want. If there are features it doesn't have that you'd prefer it did have, you are free to add them. You can design plug-ins, add or change functionality of existing compnents, practically anything is possible. It's an expensive program for sure, but that doesn't make it secretive, opressive, or evil. As I said earlier, it is great example of getting what you pay for.

The bottom line is it doesn't matter who made a tool or whether or not they decide they'd rather be paid for providing it. If it works, it's worth using. If it doesn't it's not. That's all I was trying to say.


Look, when I said "join capitalist society", I was referring to the fact that YOU had stated that you are unwilling to use anything program that costs money, aside from the occasional game or two. I fail to see what IBM has to do with the discussion.

The free tool you are using doesn't currently do a job you want it to do. A tool that requires you to spend money in order to use it does. Getting the latter would mean making an exception in your no-proprietaries policy, which I whimsically referred to as joining capitalist society.

Since it seems that any degree of figurative speach on my part gets viewed by you as some sort of attack, I will now refrain from the use of any further metaphors. Yes, I am aware that you receive a pay check; I know that you buy things in stores, that you pay taxes, that you contribute to our society's economic status. You are of course a fully fledged member of our capitalist nation. So, relax. I am not currently, nor did I ever, attempt to accuse you of anything else. As I said, my statement was simply an analogy.

Now, as for the animation part, I thought it was pretty obvious that I was referring to SL animations and not animation in general. Your question was specific for SL purposes and so was my answer. I didn't think repeating those two little letters in every single sentence was necessary.

Clearly if you are proficient with Blender (which I assume you must be to some degree or you wouldn't have asked your original question in the first place), then you already are animating. Great! Why in the world would I tell you to stop what you are already doing?


This could not be further from the truth. I have said three times now that if you want a Poser-like skeletal system in Blender so that you can use Blender for SL animations, the way to make one is to take a good look at how Poser rigs their character models and replicate it exactly in Blender. Now isn't that an example of using a proprietary program to help an open source one?

You asked for a step by step tutorial. I can gurantee you if one ever is written, it will be writen by someone who has done exactly what I'm saying. The ONLY way to get Blender to do what you want it to do is to build a human model that is rigged exactly as a Poser figure is rigged, plain and simple. I don't know how I can make it more clear.

If you are truly such a champion of the open source universe, why in the name of all that is decent and holy, would yuo be so opposed to this idea? Just downlaod the free trial of Poser, which costs you absolutely nothing. You then will have 30 days in which to pull apart its skeletal systems and learn how they are set up. From there it shouldn't take you more than a few minutes to build a comparable skeleton in Blender. If you document your efforts, you will have created a tutorial for all other Blender users to follow.

It seems to me that by your own definition, that would be perhaps the greatest contribution to SL animation you could make. From that point on anyone would be able to follow your teachings and get Blender to do SL animations. You'd be able to feel you had provided a viable alternative for those unable or unwilling to buy Poser. Done deal.

So the only question is are you up to that challenge or do you just want to hope somoene else will do it?

As for me, I've got a tool that already works and I'll stick with that.

I'm not using Blender as a toy. For an application I work on, I use it and some custom and modified Python scripts to automatically translate a NURBS surface model to several different levels of detail of a plain vertex model, then export them each in x3d interchange format. The result is that I just have to store the .blend file in source control, not each LOD level, and whenever anyone modifies it and tries to rebuild the app, the build goo (GNU autotools) kicks in and generates everything needed. Same deal with GIMP, its scheme binding ("script-fu", and its XCF format. It makes it significantly easier for artists to get their work done, as they aren't worrying about doing all the conversions by hand to make the app work right, or any hacks required for certain platforms (Windows can't do alpha transparent PNG, so I automatically toss out any layers with alpha when making icons for it, like the drop-shadow layers used on Linux) and I can change what I want in the build without requiring the artists to re-do all the conversion work.

Blender and GIMP are extremely powrful programs. You'd probably have to be an incredibly talented artist to need anything more than they provide. The only complaint I usually hear about Blender is the UI has a steep learning curve, which it does.



So tell me, what's missing from Blender that you need? If you're so sure it's a toy, weak, and limited, you must have a large answer to this question, right? Without running off to the webpage to try and cook an answer after the fact, full of obscure features you've never used in other tools?

Is it that it can't create a Poser-format bvh? Well, yeah, it's not Poser. Can Poser create a Blender-format x3d? Probably not. There's not much interest in creating bvh files, since you wouldn't normally want to create bvh files (your motion capture program would, then you'd import it where needed, and save in a richer format).



I don't enjoy improving things, I enjoy the result. If it's not going to save me time or enable me in some way, I don't do it. I do look long-term on this, though. E.g., if I use proprietary software, there's a good chance I'll run into a need to modify it in the future, which I won't be able to do. That will waste me time, working around the lack of a feature, or a bug. As a programmer, not an artist, I'm more sensative to this, as people expect Windows builds, and Windows has a huge pile o' bugs that wastes me massive swaths of time. I also do embedded development, where people expect WinCE builds as well, which wastes me gargantuan amounts of time, as WinCE is trash.

The embedded market is moving towards Linux for this reason, btw.. So much productivity is wasted steering around bugs and misfeatures in embedded proprietary software, that it's the natural choice. You'd be amazed at how much time is spent doing workarounds to avoid issues in proprietary software, and how many features are dropped from design documents because of them.



Changing plugins and the UI isn't the value of modification. The value is in making the impossible possible, which is often extremely simple if you have modify rights. I'll give a couple SecondLife examples:

If you're going to send a string in a CSV list, you need to base64 encode it to remove any comma characters. E.g., like in "Hello, world!". However, llStringToBase64() and its companion takes over 250ms when called on even just 1 character! That must be a bug, however, what can I do? I can't fix it, so I have to work around it. For me, that meant abandoning a fancy linked message-passing framework I was writing, as I simply can't take that much delay (imagine having to decode 31 link messages from a broadcast - it'd be unresponsive for seconds).

When vehicles and similar cross sim boundries, they often lose their attributes. It's probably a trivial bug to fix, but we can't, so everyone has to waste time writing complicated workarounds where possible, or simply do without when not.

llSetPrimitiveParams() fails on attached objects, even if you were just setting attributes that are legally settable on attachments (texture stuff for example). The result is you have to call the texture functions individually and for _each_ face, which is ungodly slow. Again, this is a trivial bugfix (probably less than 10 lines in the code), but since we can't, we have to write complex workarounds. For me, this meant gutting xyText so that I use the 1-prim version modified to work with three scripts per prim, doing part of the work in parallel. This took a lot of work, and I can't use the 6-char-per-prim version due to it being insanely slow with such a scheme, so I have to make do with less visible text and a larger font.

llList2CSV() discards type information. If you want to pass lists with type info, you can use a replacement script (see the wiki) but it turns my 31ms calls into 1 second calls! Totally unusable at that speed, so I have to write a very complex workaround. I'm still not finished.

And those are just a couple SecondLife examples. In the real world, programming with/against proprietary software, especially if you get stuck doing middleware, will make your hair drop out faster than shampooing with Nair. It's all about navigating the bug and missing feature minefield to come up with some least-common-denominator path that is bug-free, which is often 90% of the effort. This is why I like Free Software, and Free Software tools. I don't need to find that least-common-denominator path, I can just blaze my own. Less effort, better result. As such, keeping proprietary software (especially tools and libraries) away saves me time long-term, and lets me get the exact result I wanted, rather than having to gut the idea, or abandon it.



Btw, Free Software doesn't mean you don't pay for it. See: RedHat Advanced Server, SuSE, GNUPro, etc., and the folks that sell exceptions like Qt ($5k per developer), MySQL (which people like the guy who wrote this forum haven't realized they owe money to, yet), Cygwin, etc.. Free Software means freedom, not 'freeware'.

I consider not looking at the rights your tools give you a short-term worldview, as you'll eventually run into limitations you can't overcome. You've surely already felt the burn of something in SecondLife you wanted to do, should be able to do, but couldn't. The more you play SecondLife, the more things like this you find, the more you conform to the bugs, rather than overcome them. And that's just SecondLife. What recourse do you have but to cry to the Lindens to fix things, like implementing outbound XMLRPC? None.



I'm quite aware of this, I don't see why you think I'm not. I had heard from a player that it already could be used with SL, which is why I posted here.

The way I'd probably do it is by importing a .bvh produced by Poser to get the bone names, saving that as a template x3d file (possibly add some skin to it) and writing some goo to extract the bones from a saved x3d file and write them into a .bvh. It wouldn't be worth writing a real .bvh exporter, as no one else would use it. I might take a look at what would be required to do this next weekend.



I'm not, why do you think I am? I thought I was pretty clear that I'd not be using Poser to do animations, I'd do it with Blender, make it doable with Blender, or just not do it.



A lot of people will say things like "If you don't like it, change it yourself" as a blow-off. That's not really how Free Software works. Users that can't program can still ask for a feature, encourage others to create a feature, document how the feature should work, make good bug reports, etc.. It all adds to the development process.

In this case, though, the only folks that would care are SecondLife players, as there's no other reason I can think of to export only the skeleton to a Poser-format .bvh, so it'll have to come from one of us.

I think you hit the nail on the head right there. I am NOT a programer. I AM an artist. I think we're just having a hard time understanding eachother.

We can spend the next 10 weeks going back and forth on this like we have been or we can just agree here and now that our communication styles are differrent, our needs are different, our use of available tools is different, our sensativities to various areas of this discussion are different. I'm getting the feeling that each of us is drastically misinterpreting what the other has said.

So, I propose that we just drop the debate and agree that each of us is making a valued contribution. Were it not for people like you, people like me would have none of these fancy tools to use, and were it not for people like me, people like you would have no one to use the tools you've put together. It's a rare bird indeed who can understand both, and all evidence thus far suggests that neither of us is such a bird.

I can see now why you didn't like my answer to your original question of how to get Blender set up for SL animations. It seems that perhaps as a programmer you were looking for some sort of translation solution between the two applications, where my first instinct as an artist was to just say "look at the thing and rebuild it". As I said our communication styles are different, as are our very thought processes. Both are equally valuable.



Okay, since you asked, here's my answer. I'll admit it had been a while since I last looked at Blender and I am impressed at how far it has come. However it is missing some VERY important things.

If I'm wrong, plese correct me here, but as far as I can tell, Blender still has no tools for creating dynamic fields. It has no volume deformers. Is has no dynamic curve solvers. It has no paint effects. It cannot do real fur or hair. It cannot do real cloth. It cannot do real muscles. The list goes on and on and on, but these are some of the most commonly used things I could think of,

Obviously, I understand that these kinds of features constitute a good amount of what you pay for when you buy a program like Maya, Messiah, or SoftImage, but they are some of the things that seperate proffessional 3D modeling packages from what I referred to as "toys".

I applaud Blender for its ability to put basic 3D modeling and animation tools in the hands of anyone who wants them, but as I said it is very weak. I can see how someone who has never used something like Maya might think Blender is "extremely powerful", but I can assure you it is not in the same league.





Sounds like an interesting project. I hate to tell you this, but (assuming I'm understanding you properly), Maya can already do this. Since your goal is to avoid such things as Maya though, then I guess it's good that you are adding this functionality to what you are using. It's very useful.

Kral,

You do know that SL is closed-source and proprietary right?

Azelda

Azelda, Kral did mention he uses proprietary games. Secondlife is technicallly classed as a game of sorts.

As someone who does both a bit of progarmming and art, I can see where Kral and Chosen are coming from. They're both right in their own ways.

At any rate, Kral, I would love to see some sort of automated Blender support for exporting (secondlife compatible) bvh files.

I'm going to miss many of those terms as I don't use Blender as a renderer, I use it as a modeller to make models for other software to render. The rendering I'm interested in takes place outside of Blender. It sounds like we're coming from two different backgrounds, modelling and rendering. I know scripts for fur and cloth exist (e.g, http://www.blender3d.org/cms/Misc_improvements.355.0.html) but I've never used them. I'm not sure what you mean by dynamic regions - if it's software-automated regions that are updated per-frame like ocean simulators, then yes, I've seen such scripts for Blender, but not used them. For muscle-ish volume deformation in Blender, this explains it better than I could: http://www.cgtalk.com/archive/index.php/t-50471.html. It's probably now (8/2004) possible to drive a lot of the model's behavior during rendering using python scripted physics objects and constraints from the game logic system, but it's another thing I haven't looked at (although it is one area I'm interested in for other things - making keyframes that act proper wrt gravity without all the fidgety work), as I don't even look at the rendering capabilities of 3D packages, since I'm the one that will get stuck doing the rendering.

Personally I think the 3D Max default renderer sucks. Sure it's fast, but I'd rather use a ray tracer any day. Fork out another $500 for a raytracer plug-in for max. Let's put it this way. 3D Max wouldn't be very good if it weren't for it's 3rd party plug-in support. Which there are tons of for anything you can think of to make life so much easier.

Honestly, I respect people who do their work strictly in something like POV-Ray code over 3D max. Sometimes 'reinventing the wheel' is more gratifying than using something someone else made to do your work for you.

This is all coming from an artistic standpoint. Now if you're doing something professionally that has a due date set upon it, then yeah I'd use something like Max of course.

Ah, sorry for not explaining. Most of the things I was talking about are probably terms that only 3D artists are aware of, not everyday knowledge, and it's possible the names may vary in some packages. None of them are rendering tools, however. All are modeling tools, and a few also act as animation tools as well. Let me give some definitions:

Dynamic Fields - These can best be described as simulated energy, applied to an area in 3D space, to part or all of piece of geometry, or to particles. These can include gravity, turbulence, chaos, air simulations, friction, drag, etc.

Dynamic Curve Solvers - This allows curves to behave in a dynamic state, effected by gravity or any other dynamic fields. This an EXTREMELY powerful animation and modeling tool. For example, use a few curves as a guides to shape the flow of hair follicles, apply a few fields, such as gravity and wind, set the curves resistance to the field accordingly, and you now have incredibly realistic, lifelike hair that behaves exactly like the real thing. Another example would be to loft a surface between a couple of parallel curves, apply some dynamics, and in 2 minutes or so you've made a completely articulated flag that is able to droop with gravity, flap in the wind, fold, wrap around another object, etc.

Volume Deformers - these are objects that can be placed inside surface models that can govern the behavior of the surface with respect to the volume of the deformer. They are usually spherical, and are often applied inside character models to help ensure that a character maintains its volume as it moves. Without them in place, it's easy for something like an arm to behave more like a rubber hose than an arm when it bends (obviously, using a deformer is not the ONLY way to avoid this problem, but there are advantages and disadvantages to every technique). Deformers can also be used to let one object influence the shape of another when the two are in contact. For example, model a hand, turn the hand into a deformer, and push it into a plane. You now have a hand print in your plane, similar to the hand prints children like to make on newly constructed sidewalks.

Paint Effects - These can be 3 dimensional or 2 dimensional in nature. A good example of paint effects in action would be Smiegel's body hair in Lord of the Rings. (I can't say for certain that these were paint effects since I didn't work on that project, but it's most logical that they would have been.) There are thousands of hairs on him from head to toe, just as there are on any of us. Modeling each of them as geometry would obviously take close to forever. Applying them as a texture would not look realistic. The solution is to model one hair or a small clump of hairs, turn this into a 3D paint effect, and then "paint" the hair onto the body with a virtual brush. Settings can be applied for a certain degree of randomness for each hair so that they do not look like they all came from the same model. Control curves can be used so the hair will flow properly and so that it will respond to dynamics. ..and the rendering speed of paint effects is phenomenal. Their only drawback is that they are not real geometry and therefore cannot be exported to things like video games, which only understand polygons, but for film, they are wonderful. (They can of course be converted to polygons if you really want to be able to export them, but the poly count would be so high in most cases it's not worth it.)

"Real" Fur and Hair - Take a look at Monsters Inc. Notice every time Sully moves, his fur articulates, responds to gravity, responds to the air it is passing through, responds to objects it collides with. There are millions upon millions of individual strands of fur, far too many to be individually modeled, and each one behaves properly. This is because "real fur" was applied on top of Sully's skin. The static particle model in your link, by comparison, looks incredibly primitive. The hairs move with the creature alright, but they remain rigid like porcupine quills, not like flexible hairs. There is obviously a good deal of wind in the scene, as is evidenced by the snow blowing around, but the fur does not react to it at all. This fur appears to be made of steel, not made of hairs. The artist gets away with it to some degree here only because the hairs are so short. However, it would be very problematic to model a something like a lion or a horse with a mane using this technique. The lion would end up throttling himself with all his rigid spikes that should have been his hairs, and the horse would end up with a 1980's punk rooster Mohawk instead of a flowing mane.

"Real" Cloth - Simulated cloth that behaves like the real thing is one of the hardest things to make. Even from a traditional artist's standpoint, just being able to draw or paint realistic cloth is one of the signs of a master. Giants like Michelangelo and Leonardo literally spent YEARS learning to draw and paint cloth. (I thought my most recent drapery studies in charcoal were pretty good until I took another look at DaVinci's.) For the 3D artist, the problem is compounded. Not only does it have to look right, it has to behave right. The team that developed Maya's cloth system spent several years getting it to work, originally with the fashion design industry in mind. Sure there are facsimiles of cloth that anyone can make, but nothing is in the same league as the realism achieved by a genuine dedicated cloth system.

"Real" Muscles - The hallmark of Messiah software is its ability to create realistic musculature in character models. Project Messiah Group took the traditional modeling tools of skin and bone, and added what has always been missing, muscles. These muscle expand and contract, tighten and loosen, flex and relax, bend and distort as the skeleton moves. They cause the skin to deform along with them, and give the impression that the model has real anatomical structure (because it does) and is truly alive. Musculature is such a tremendous leap forward in character animation I'll never animate a character without it. I'm not aware of any other package that has it other than Messiah (and thankfully Messiah will plug itself into lots of other packages). In regard to your link, yes the effect can be simulated using blend shapes (the traditional method), but this is limiting, relatively time consuming, and requires some skill to set up properly. It does not have the functionality or flexibility that Messiah's dedicated muscle system provides.


Okay, I think that covers everything. I'm not sure what you meant by the phrase, "dynamic regions". I never used that term. Perhaps you were talking about dynamics fields, which i already covered.

In regard to what you said about driving a model's behavior during rendering, I suppose there's no reason why that wouldn't be possible, but as an artist I have a hard time understanding why anyone would want to do this. I want to control what my models do BEFORE rendering. The last thing I'd want is to spend a few dozen hours rendering out a movie, only to discover after the fact that my Buggs Bunny had made a wrong turn at Albuquerque, and didn't end up at Pismo Beach.

I've seen lot's of talk about a fix to get blender working fos SL animations. you just need to work out a perfect skeleton they say, or import a skeleton that was exported from poser. i would really like to know why no one has yet posted a template and a fix for this?

why are animations being reserved for the upper class. equality now, the people shall be heard! the masses must learn to move themselves! animation is the way.

Will one of you geniuses please create a nice template that people could use for blender? wouldn't that do wonders for the game? am i just totally of base here? i've seen the questions posed many times and yet to find an a answer that wasn't specualtive.

can blender be made to work? can someone post a workable fix for the masses? a template or skeleton that could be used? the poor are begging... it's almost xmas....

i want my AV to sneeze and pick his nose. i want him to scratch his ass and pick the rocks out of his shoes. i want him to slink, to crawl....

ody

IIRC, it was Beatfox Xevious who posted the Blender skeleton *and* patches to the python script to make sure that Blender exported the BVH files correctly. That said the skeleton by itself is not quite enough. This is something I am working on in my copious free time, but I've also been asked to do a *heap* of other things in-world so it's fallen to pretty low priority.

- C

awesome. i can't wait to hear your solution. i bet i'm not the only one.

If Kral Playfair had taken the time spent arguing Open Source In General, and instead used it to make a skeleton and tutorial, it would have been done back then. I see this a lot in open source projects. Tens of thousands of words are spent talking about the things they say they could do if they decided to try. It's a wonder anything gets done.

It sucks a whole lot less than Maya's scanline renderer, but both of them are capable of stunning results once you know them well. Max now comes with Mental Ray out of the box, so you don't have to fork out money for a third party renderer unless you want to. I personally use Vray Advanced because of its excellent speed.

I've searched and been quite unable to find a BVH exporter for Blender, though I'm told it's in the CVS, and the BVH importer seems to work fine, but I never did find an export script. Could Kral (or anyone else) post a link to the BVH exporter? I've done a fair bit of playing with the BVH format (trying to get it to work with the program I use, Cinema 4D) and may be able to offer some insight to someone who is more familiar with python than myself.

Thanks