#SL Mesh Deformer Debate Update

There was lots of debate on the Mesh Deformer and whether it should have a single default shape as the basis for deformation calculations or whether it should have multiple base shapes. That question is still undecided. Discussion has kinda died down, at least in my circles.

If you are unaware of the sides of the debate, see: The Great #SL Deformer Debate.

Open Source Meeting

Questions

One of the prime questions I asked in the article is how it will be funded. That is not likely to be a problem.

Another problem is what it would cost to add the capability. Emma Gilmour asked Qarl what it would cost, not that much. We are certain we can raise the needed money in a very short time. We have almost half of it covered now.

So, a couple of questions are down and the possibility becomes much more likely.

And the Lab…

The next challenge is where are we with the Lab and Avatar 2.0?

Avatar 2.0

This is a hard one to make a call on. In the previous article Max makes some good points in the comments.

I asked Oz Linden at the Open Source User Group meeting if there were any proposals from open source peeps pending on Avatar 2.0. There aren’t. Oz answered:

Avatar 2.0 … I have not seen a proposal…. I’m not sure how/where this got started again. Even if there were one, I’d bet a lot against it being given a green light. I think we’ll probably get some minor fixes to the current avatar done… storm-1800 at least. There are a couple of more things that are clearly not great that need fixing. I’d like to get them all done at once.

A couple of people have asked me about it [Avatar 2]. I have not been encouraging. The problem is that there are a _LOT_ of dependencies (some reasonable and expected, some less so) on the current avatar.

That brought up the question from me, “Does the weighting on the avatar have dependencies?” I’m relatively new to weight painting, but I can’t see it having dependencies. The painting is dependent on other things, the skin mesh and armature. But

Oz answered, “I’m the wrong person to ask. The trouble is that I’m not sufficiently expert on these things to make any real decisions. But, we had a good review meeting on it today, and have gotten some testing done. There are some things to investigate, and more testing to do.”

Alternate Base Shapes

I asked Oz if having multiple base shapes was a reasonable idea. The answer:

That’s an unresolved question. The real question at this point is whether or not it’s actually needed. So far, there are two reasons I’ve heard for doing it:

  1. Reduce texture distortion
  2. Compensate for odd avatar artifacts at extreme settings

Oz Linden: I don’t find #1 persuasive because:

  • Everyone lives with texture distortion now just fine
  • It wouldn’t improve texture distortion due to bone length changes anyway and besides, the tests I’ve done so far don’t seem to me to have much of a texture distortion problem.

#2 is more persuasive, but personally I’d rather see an effort to correct the underlying flaws in the base avatar mesh that cause the artifacts in the first place. I’m trying to figure out if that’s possible. But overall, I’m fairly impressed with where we are on that (Mesh Deformer]…. it basically works. I need to pull together the results of our [Linden] meeting (which was very shortly before this one) and send it to Qarl. So … we’re making progress.

The Scenario

While we have a segment of the user base that is convinced alternate and multiple base shapes are important and a solution to some problems, Oz Linden is not convinced. His thinking that users are living just fine with the current distortion is a hint at his perspective. This is one of those places where the programmers and the users see things differently.

Tolerating is not living just fine with a problem. I assume he thinks this because we are not chasing him with signs proclaiming our dissatisfaction with texture distortions.

I think the issue with bone distortion is for a limited user set. Oz thinks that will distort textures no matter what. Emma Gilmour has a Plurk about… more oddly sized avatars… and the problems they are having with mesh clothes. We think other base shapes can minimize what bone sizing could distort.

I also think that with the deformer in action we are going to see more users unhappy with distortion. I can’t know that so we are balancing opinions. So, which weights more, a cubic foot or a cubic kilometer of nothing?

Convincing

To change the Linden minds we need to do some work.

One, we need to get the funding in place. Since we can’t know whether the Lindens will accept this idea that makes this a candidate for a Crowd Funding project so money can be returned if need be. This shows a level of commitment. We should probably try for lots of small contributions rather than a few large ones.

Two, convince Oz that distortion is a problem. He has given us an opportunity to do just that and help with testing the Deformer. He suggested testing the Deformer saying:

I’d like to have a session with 20 or so people running the deformer and wearing mesh outfits that are deformable. Maybe we could do that at this meeting next week? Throw in a little avatar physics for good measure…. Ok… well, next week may not be best, but I’ll set it up.

When we know when that is going to be, we should get together those that are unhappy with distortion and have non-stereotype shapes show up. We can’t have hundreds trying to show up. But, we do need good examples of the problems to make our case. If we can’t, then we know we’re out in left field and the one-shapers are main stream.

Oz needs to be able to see the problems. He also needs a sense of resident dissatisfaction with the current avatar. May be we all need that sense to get things in proportion. I may think more people want multiple base shapes and are unhappy with things as they are.

Looking Forward

We need to make decision quickly because things are moving forward.

The Lindens will be deciding what can and cannot be changed on the avatar. I’m very strongly convinced that Avatar 2 is not a possibility. But, STORM-1800The vertex weights of the default character mesh could be better, is likely. Oz hopes to see some of the issues in that JIRA handled.

When we look forward, we think once the Deformer is in place, it won’t change and that it will be hard to get the Lab to reopen discussion on it. I don’t believe that. I think the Lab resists changing the avatar, but I think the Deformer is like any other Lab project. Projects move from development to maintenance. In maintenance, lots of stuff changes. Since the development work is being done outside the Lab, I can’t see it being too much of a problem to change the Deformer later. But, getting it the way we want now is desirable.

Summary

Let those you know in the fashion community about this article and let them know Oz is looking for more male clothes to test with the deformer. While he got some clothes to test with, he did not get many males clothes examples.

To stay organized, follow Emma, Gianna, and myself on Plurk. …and of course follow my blog… 🙂

11 thoughts on “#SL Mesh Deformer Debate Update

  1. My feeling was that distortion is going to be a problem – Oz doesn’t deny it, he just thinks it’s not a priority because it already exists. Perhaps that’s the difference between a Product person and an Engineer person. 🙂
    Another reason i think sizing is so useful is a fuzzy one. It gives creators another level to make even betterer content. I can’t say entirely why, but it’s the kind of feature which clever creators take and run with.

    I’m glad to see those that are involved leading this project. Oz is a good Linden, Gianna is thoughtful, and Nal & Emma rock. I think it might turn out ok – which is me cheerleading. Go Team!

  2. It seems to me it should be fairly simple to solicit some “extreme” avatars.
    Female: Breasts/Hips/Butt big/small. Tall and short. (16 avatars).
    Male: Mostly big & beefy and slim and scholarly, perhaps short and tall. (8 avatars).

    But you are right, we have an opportunity to improve the situation with texture distortion not live with the same problems we always have. I thought improvement was what mesh was all about!

    • Improvement is what mesh is about.

      You did notice that Oz admitted he doesn’t know enough about mesh modeling and avatar design to make the decisions. He is the public face of the Lab’s for whoever is going to make the decision. I think it is important for the community to understand his perspective, if we are to communicate well with him.

      As a programmer, manager, and allocater of the Lab’s resources his perspective is different and his priorities are different. In a large part the Lindens experience is users do not know what they want. The viewer side bar is a good example. Everyone hated it. They got rid of it and everyone hated that it was gone. The Lindens tend to be a bit fatalistic and users lack significant credibility. So, being convincing is not a simple task.

  3. Today I sent an alt out to try some newer mesh clothes.
    The creators have gotten better and “standard sizes” do help.
    If / when some standard variations (hopefully) make their way into the deformer I have a suggestion. Do not call them “sizes”, especially in the deformer they will be better called “proportions”. A busty, blocky built short av will probably be best fit with a Large; whereas a tall, slim flat-chested av might be a small.
    How things are named seems like a small thing but it can help or hinder understanding for the users.

  4. Pingback: Qarl's Mesh Deformer -- latest news - Page 9 - SLUniverse Forums

  5. Pingback: Do #SL Mesh Clothes Fit

  6. Pingback: Do Your Mesh Clothes Fit? - Zeus Gay Club » Zeus Gay Club

  7. This fix for the issue is not for any1 to pic what the deformer deforms to. The total and complete fix would be for the user to choose what the deformer deforms to.

    Every1 wants to keep arguing about making a mesh outfits fit this limited range of human avatars, but if it was look at in the broadest terms, then every1 would be happy. The Default is just what it is, a defualt standard to create from. If an avatar mesh could be labeled an avatar mesh, giving it the ability to be the default, then the mesh clothing would fit any avatar. Of course there would extreme cases, but you are going to have extremes no matter which way you go. This solution would give mesh clothing the widest range of avatar that it would fit. It would also eliminate the need for LL to ever make an Avatar 2.0. This solution alone would negate the whole Avatar 2.0 debate. Texture stretching would be lessen by the custom avatars, but the real fix for texture stretching is for creator to make custom textures for extreme bodies. There is just no way around that.

    • I agree that other size avatars and clothes designed for them would have less stretch and compression.

      The problem with an infinite number of user created base shapes is how the viewer handles them. I’m not sure how you get your custom base shape to me so my viewer can start calculation on it. I’ll have to ask Karl if that is a problem. I think it would be. So, a finite set of shapes is probably mandatory.

  8. I’m not a creator. I don’t understand how one “fits” mesh clothing at this point. So my question to creators is this:

    In your 3D program can you create a garment, in mesh, say a standard size large and texture it, so it looks perfect on that size, and THEN refit it again to the default avatar shape without changing the textures, just, idk, squishing them? So, when it is worn by an avatar in world and deformed by the deformer, it looks perfect on the size it was originally designed for?

    I’ve seen examples of clothes deformed for what some consider extreme shapes (basically a curvy girl like myself) and the mesh “breaks”. But I dont’ know if that’s an issue with the way the mesh itself was created or if it’s a problem with the deformer that needs to be “fixed”. Or are several default shapes the only way so the garment never has to deform too far?

Leave a Reply

Your email address will not be published.