my only caveat is that people will try to upload anything, and care none for optimisation. How to get around that, good question.

To be specific, I would like to have a tool that would create a sculpt image from the exterior surface of a union of a linked set of overlapping prims. This would warp to follow straight edges and adjust to surface curvatures. with the proviso that the sharp edges could be adjustably rounded to elimate the Phong shading artifacts.

I saw this idea mentioned long ago on the sculpt wiki discussion.

If you decided to go with import mesh.... for the love of god, please set a polygon limit.

(TREEESSSSSS)

Yes, I'd like to see an inworld tree and plant modeller, like the one that's been built for 3d Coat, for instance.

http://3dbrush.kriska.hvosting.net/forum/index.php?showtopic=1003

see, just one dude did that!

yea, I want my cake and eat it too... no surprise

Could do it in OpenSim, though.

I like the 3d Coat idea... its using sprites/billboards as the leaves rather than geometry, that face the camera. (similar to particle systems) Megasavings on geometry.

What ever happened to Ventrella's projects?

The comments in the other entry are mostly about using professional tools - and that creates a barrier the regular user has to creating content that they can feel competitive with. My other objection to this is that this is one more thing that moves users back out of SL. In some ways, even though I sell on it, I think SL Exchange has harmed SL itself by bringing people out of SecondLife itself, back to their web browser.

I remember using POVray, and enjoying the rich array of primitives available there.

And it will export COLLADA, so no worries.

http://www.collada.org/mediawiki/index.php/XSI_6_Mod_Tool

Cant think of anything better to animate with. Forget about Maya.

However, I don't think that SL is ready for general mesh importing. The system already has a lot of load on it, and this would add more. Plus, this will put more strains on the client, and SL will get an even worse reputation as being "laggy" because of slow framerates.

I was actually somewhat against (and not much for) sculpties when they were announced. They're great and people are doing a lot of cool things with them, but they also showcase the unwillingness of most SL content creators to consider the load impact of their content on viewers. All it takes is one person with a monstrous quad-core nvidia whatever spewing out a ton of sculpties that "work fine for me!" and you end up with a laggy sim. Most of SL is already pretty laggy because people don't use any restraint with the features we already have. Think the old-school phenomenon of "torus hair" lagging up an entire venue, or the current way people stick hundreds of flex-prims on a single avatar, all of which end up rendering at 1 FPS or less because there are so many on-screen. Entire builds are made with sculpties, so when you teleport in, you're surrounded by a bunch of grotesque spheres and ovoids for a few minutes until the sculpt textures load in.

It would be great if LL managed to do mesh import in a sane way, and I might come around to it. But you have to consider that people WILL overuse it. Entire builds will be made with it, and you have to consider the worst-case scenario. Be careful how many ways you give users to shoot themselves in the foot, because they shoot the rest of us too.

http://www.realityprime.com/articles/volumes-of-reading

I gotta think we can do better than what LL does right now, but doing it without raising the bar on the system requirements for the SL viewer is an important consideration also.

> Pro tools? XSI Mod tool is like, free. You can learn it and use it to model for any game for exactly zero money.
> Its a free version of XSI Essentials, with only a few limits (polygon limit and a max 512 render). Which is
> enough for SL. It is FREE mental ray. Srsly.

> And it will export COLLADA, so no worries.

> http://www.collada.org/mediawiki/index.php/XSI_6_Mod_Tool

> Cant think of anything better to animate with. Forget about Maya.

Or you could use Blender, or I'm sure there's some tool out there to convert POVray models to whatever your favorite format is. You're still getting pulled outside of the SL environment.

Heck, I'd like to see some tools for working on textures within the game - nothing on the scale of PhotoShop or CorelDraw, but something would be nice.

BTW, "Pro" doesn't have anything to do with cost - it has to do with the tools those who do this kind of work for pay use.

The only way I'd really want mesh modelling to be put into SL would be if you could do it in-game, grabbing a vertex and stretching it, or splitting a face down the middle to generate two new faces - things like that.

Prim-based modeling always results in redundant geometry and requires more polygons/vertices than necessary. Take a look at other computer games and ask yourself why they look so much better AND run at higher frame rates than Second Life. It's because of two things: Mesh-based modeling and normal-mapping.

The big advantage of mesh-based modeling is that only visible surfaces need to be included in the model. This reduces the load on the graphics engine. In the long term, mesh import will also open the door to custom avatar meshes. These will be much more efficient than standard meshes with dozens of attachments.

Of course prim-based modeling in SL will not go away. Of course there will be reasonable limits on the number of polygons per custom mesh. But unlike sculpted prims, the number of polygons in a custom mesh is not static and can be much LOWER for simple irregular shapes. Furthermore a custom mesh can have any topology, not just cylinder, plane, sphere and torus. All these things will bring the number of polygons DOWN and rendering speed UP.

Clothing artists arent crying because they have to use a gasp, graphics program, many times even, bigger gasp, a 3d program! to make good clothing textures. Fashion designers arent crying because they have to use Lightwave or Zbrush to develop properly seamless baked clothing textures or sculpted items to accent their clothing designs. They were using outside programs already and =always have=. Probably why so many jumped right into sculpties so well.

Animation artists have been using mocap and programs outside SL such as Poser for quite a long time. In fact this is one of the things that should have been INWORLD for a long time. People should be able to interact in the environment more naturally. Yet that work by Ventrella was all put aside. I'd like to see that taken back up again.

Only "builders" scream they need to "remain in the environment". Only "builders" want to be subsidized.

No, I am realistic. There are things that math can do better than mesh, but mesh can do better than prims on a host of other things. This is why I voted for both. I want the CHOICE to pick what works best for the project at hand.

I don't have anything particularly against meshes if they are something that would work properly in SL - it sounds rather like a lot of work on the back-end, but I am no expert - but there should definitely be an inworld tool to manipulate them as well.

Also, I would agree with Lex Neva that it had better be absolutely bulletproof in terms of lag. Casual and widespread use of sculpts is starting to become a problem now.

(As well as this I could name you half a dozen things that I would put far before either of the options given.)

But as Ordinal says, the implementation needs to be bulletproof. That's why I could see a number of serious limitations on uploaded meshes. Among such potential limitations, I'd suggest the following be considered:

• high upload costs*; staged to polycounts
• automatic asset IP tagging/tracking* (with a searchable catalog so IP-owning 3D modelers can more easily flag violations)
• phased-in ability to "Copy"; starting off with "No Copy", proceeding through limited duplication and eventually ... when bandwidth is less of a concern ... to unlimited.
• phased-in rezzing limitation; first non-Mainland and then the Mainland

If we can't implement mesh uploads in a highly controlled manner I'd rather LL focus on providing more capable in-world modeling tools.

*Note: I've always wished an "asset removal/refund" system was in place for items which haven't been duplicated and could be wiped from the system.

The other option seems mostly self-serving...people with expensive tools/relevant job experience/infinite time to learn meshes wanting to corner the ingame market as it were, with LL's help. Their skills already put them over most creators, but that apparently is not good enough and they must CRUSH the non-pro builders in the market.

Rather than a mesh import system geared to giving pros the market, give us some simple things that were in VRML 10 years ago for christ sake'....say allow overlapping a prim and joining them into one 'object mesh' ingame OR overlap two prims and subtract one to make a custom shape. or let us build something out of prims and have SL turn it into a mesh object. The pro would STILL have an advantage over the average builder, but not crushingly so.

Some of the artiste' types in the other JIRA should realize once they have driven all the inferior creators who make 'junk' (as one pro poster referred referred to prim builders) who use the money they make off THEIR content to buy OTHERS conten or to pay for landt, who's going to be left to buy their ubersnazzy mesh content? For my part, LL should stay true to 'your world, your imagination' on building, rather than 'your world, content by pros only' and keep as much of the tools as possible INGAME and accessable to all of varying skill levels.

I see it over Fashcon which is an excellent barometer of what people are making. Yet nobody is screaming that good designers are all using Photoshop of whatever flavour. Which is a pro tool used widely by non-pros. Nobody has been arguing to GIMP them.

Arguments for hobbling SL will only help tremendously whatever world competitors that come along to crush it (the response for mesh import is obviously overwhelming), and helps those who run Opensim grids tremendously as content makers will eventually need to go elsewhere to places that are willing to innovate. Don't fear mesh import, fear what will happen to SL if it doesn't keep up with its own competition just to placate some people who don't want to learn.

And no, I'm not a pro from some pro school professing the wonders of Maya when they can barely operate the program and just want to import meshes out of lazyness (which I admit are a significant amount of people who ask me questions, and I ignore them by now), I am self taught in 3d from a 2d illustration background, and a perfect example of what people can do if they are persistent enough. And I use midrange 3d software which is easily affordable by the enthusiast rather than a pro.

I would prefer if people looked at this from not a monetary but a purely performance perspective. As the average user and the average landowner is not a businessman but uses SL for their own pleasure, they dont care nearly as much about the plight of builders business as they do bad performance.

Meshes, if implemented, need to have limits on polygon sizes. I want meshes to exactly create those very small size meshes that do not have a quad based topology (using triangles) - this is very difficult with sculpties and extremely wasteful from a performance perspective. Hobbling me doesnt stop me from making hard edged objects with few faces, it just takes me longer as I am forced to use a topology which doesnt work well for the task, and prims are even more wasteful resourcewise to create such objects which should be a cinch to make in SL.

I really am having a hard time make this choice I welcome and fear mesh all at the same time.

As I said before, the REAL issue for nice appearance in SL, as far as I know, is network bandwidth, not vertex-count. And, Qarl just added the upload/display support for the new x by y sculpty textures. Before I'd support new types of meshes, I'd like 1) to see what can be done with the new sculpties, 2) to see what the original programmer left out of the primitive set and options that SL currently supports, and 3) see what the bandwidth cost of supporting new meshes will be like. Oh, and better caching of textures would be a good thing to implement first, as well. Not to mention better animation support for sculpties, and stitching support for scultpies of various sizes and shapes.

My guess is that for the kind of virtual world Second LIfe is (arbitrary content used by arbitrary individuals in arbitrary simulators), ad hoc geometry like scultpies and primitives will ALWAYS look and behave better until everyone using SL has a direct T1 connection all the way from the sim to their home computer.

And no, in this case, the fact that YOUR mileage may vary doesn't matter one bit: it's the average case that matters, not the extremely fortunately few who currently have such connections.

Performance in the viewer is a combination of vertices viewable on the screen (directly impacting frame rate) and bandwidth considerations.

Small meshes will be highly useful for NON quad topologies, something that is not very feasable with sculpties and prims. (not impossible, just very prim AND polygon intensive if you use a prim to draw each triangle, or highly annoying and not very high quality to do with sculpts in comparison to using the actual meshes.) I have done them ALL, and so I can compare. When you have done them all and can compare, then you are in a position to talk about it.

I do a lot of this kind of modelling, so I am very MUCH in a position to comment about it from a point of experience. I'd like to be able to use my own gemstone designs in SL. They are VERY small meshes. I can't very easily. This annoys me to no end.

I was the one who posted the edge stitching and sculpty skinning animation JIRAs. I still want these improvements to sculpties to enter SL. Sculpties are highly optimised for NURBS and subdivision surface modelling - they should continue to be improved. And let us use something else for hard edged low poly objects with non-quad topologies. (using triangles instead of quads) Sculpts were NOT designed for this kind of modelling - its trying to fit a square peg into a round hole to keep on bothering.

Use a hammer when you need a hammer, and a drill when you need a drill. Let's stop using a hammer when we need a drill.

I read Avi's article, and again, he makes the same point many of us make when supporting low poly mesh. Prims are not automatically better from a performance perspective as they are wasteful with polygons. This problem impacts frame rates in the viewer.

"The downside, of course, is that these pseudo-solid volumes are not as efficient to render as arbitrary hand-crafted polygons, since there are lots of overlapping and hidden surfaces (which take time to render, even if you don't see them) and require lots of tiny state changes, including simple changes in position and orientation, each of which takes a small-but-cumulative amount of time for any 3D hardware system to set up and work through."

http://www.realityprime.com/articles/how-sl-primitives-really-work

This is the program I use to make gemstones - I have used it for years. (way before SL)

http://www.gemcad.com/

Note its not a "3d modeller" per se. It is a lapidary simulator, for cutting gemstones. Most users of this program are not 3d modellers... they are jewelry designers. Most of whom are not very computer savvy. (I've spoken to many when showing my 3d renderings made with this program)

Here's some of my work from way back in 2004, when SL was just a fledgling.

http://www.flickr.com/photos/hypatiacallisto/2812253153/

The center diamond is a replica of the Hope diamond. Want to know how many faces it has? A whopping 96 polygons. Tack on a few more when triangulating the table facets, so if brought to SL, would maybe be... oh, 110 ish or so.

That is ALL.

A sculpty is MUCH larger than this! The OBJ for just my Hope is 25.5 KB. Tiny. A sculpt texture is both much larger in texture and polygon size, as it has to define the placement of 1024 vertices, not the 88 vertices which is all this model comprises! Prims inworld are also much LARGER than this!! It's very hard to get a sculpt to do this kind of nonstandard geometry accurately, and with gemstones, ACCURACY of angles is PARAMOUNT when you are doing replicas of actual stones for a jewelry design.

As altering the end / profile cuts would trigger a mesh update, it's also the ideal way to implement sculptie animations as film strip sculptie maps could be changed to the next frame just by updating the cut positions.