A bunch of news has trickled out on the Mesh Deformer Project. Check out STORM-1716 – Mesh Deformer for tayloring [SIC] mesh clothing. The JIRA item was filed by Qarl Fizz. He explains the issue very well and goes into some detail about the direction the project is heading. Clothes makers will want to read the item to see what to expect.
Since it is in the JIRA we can expect the more advanced content creators to be involved in the discussion. Also, show some interest. Drop by and click WATCH.
Qarl poses and answers questions. Probably the most important concept in STORM-1716 is the idea of basing all mesh clothes on the Ruth/Roth avatar rather than a custom avatar, as in one based on your shape or different size avatars… think small, medium, and large, I suppose.
The idea is that the ‘deformer’ needs a beginning place. From a consistant starting place our shape settings can be used to deform the mesh clothes. The shape settings deform our avatar’s mesh from the standard Ruth shape. By using the same base the deformer can make close fitting clothes that actually fit. This means a base template for all avatars is needed. The obvious one is the base model Linden Lab built into the viewer.
What is happening is the ‘deformer’ uses the shape values to figure out how much to deform the clothes. The alternative is to calculate what the shape of the deformed avatar is and then calculate how much to deform the mesh, much more complex and likely SLOW.
The advantage to the last method would be that custom avatars could be used. I could then put a sport coat made for me on a dragon and have it fit. But, that is not to be. The deformer that Qarl will apparently be building is for humanoid avatars only because it makes some assumptions for speed and simplicity.
The result of the chosen process is: if I were to make clothes based on a shape from my avatar’s exported shape, the clothes deformed by the ‘deformer’ would not fit correctly.
In clothes we have layering. I can wear an undershirt, shirt, and jacket on the same avatar undershirt layer and it works. They stack on top of one another within the avatar’s undershirt layer. I can then add more layers of clothes to the avatar’s shirt layer and so on. This allows me to have something like 15 layers of clothes. The neat aspect of this is that as a user we have control of how we stack the layers. We can add a jacket to an undershirt layer and wear something over it.
Think about mesh. I could put on a mesh bra then a blouse over the bra and then a jacket over the blouse. It could be handy to stack mesh clothes like we do layers. But, that is not to be… and yet there is a way that mesh layering is apparently going to happen.
The deformer is not going to provide layering to the user. However, mesh clothing is going to be deformed by the ‘deformer’ based on the idea of an offset. The content creator gets to set the ‘offset’.
So, when creating a bra I could have a very small offset for a close fit, a larger offset for the blouse, and an even larger offset for the jacket. I would then have layered mesh clothes. But the content creator has to design for that. I can’t just wear things and have them layer automatically. They have to be made specially for that.
If content creators elect to use a series of standard offsets, clothes will be interchangeable. I can use a jacket from one creator and a blouse from another creator and they will look right. It won’t be perfect. A bulky sweater is not going to fit inside a jacket. It will poke through. But, the direction of development is looking pretty nice.
If you have tried mesh clothes, you already know they cannot be adjusted as prim clothes are. That is not going to change with the addition of the ‘deformer’.
The deformations are to be calculated on the client side. This means no additional server lag. I’m hoping there is something in the graphics cards that helps with this calculation. If so, we likely won’t take much of a performance hit.
OK… this is a total guy thing… women realize it is like having bait, otherwise its… never mind. Whatever, it is under consideration. Qarl is looking at costs and the Lab is really interested. But, making the ‘deformer’ work with Enhanced Avatar Physics is not part of the current project. Expect to hear at some point what it may cost to add this feature in the future.
If you don’t make mesh clothes then this is over-geeking to the max. You can skip it.
To make clothes move with the avatar vertex painting or weighting is used to provide the computer information on which parts of a blouse move with the arm, torso, and etc. You can Google for tutorials on Weight or Vertex Painting if you’re curious.
Suffice to say, it is a tedious task. The tools for it are buit into Blender, 3DS, and Maya. However, there is interest in getting something to simplify the process. Something along this line is being considered as another separate project.
This is great information. It gives developers a direction to head. It also gives us a question to ask creators, ‘Are your avatars based on the standard avatar model?’ If they are, they probably won’t fit now, but will fit when the ‘deformer’ is released.
It does not solve the current dilemma of whether to purchase mesh clothes or not. If they are clothes based on a custom avatar shape (S, M, L, XL), as I think most are, they won’t fit when worn with the ‘deformer’ active in the future. Because of that some are asking for an on/off control for users. I could then wear mesh clothing that fits now and turn off the ‘deformer’. I could wear new clothing designed for the ‘deformer’ and turn it on.
I suspect we will get such a switch. Otherwise, I’ll be very unhappy with some of the mesh clothes I might buy pre-deformer.
We still don’t know what the final version of the deformer will look like or what options we may get. It is clearer, but it is not yet time to toss your wardrobe and start a shopping frenzy.
I do however love the mesh hair I got at Vintage Fair. More about that later.