Mesh Deformer 2013-23

There really isn’t any new news. Oz Linden and Lab are involved in other projects. Oz says he is continuing to pester management for people to work on the avatar and Deformer. But, people, users, are poking around trying to find out what is happening with the Deformer. That is leading to more drama and fussing in the JIRA.

Mesh Clothes in Non-mesh Viewer

Mesh Clothes in Non-mesh Viewer

I Won’t Buy Mesh

Yeah, yeah, whatever… I understand the sentiment. I don’t by mesh clothes. I do like some mesh hair and have bought some. But, posting in the JIRA that you won’t buy mesh is of no help to the development process. Duh! Nor do the Lindens care that much about what we do or don’t buy. There is plenty of buying going on from the Linden viewpoint.

I follow some fashion blogs. There is a lot of mesh clothing out there. People seem to be buying it. I have no way to know if the hold outs not buying mesh outnumber those that are buying mesh. Whatever the case, I doubt the Lab is concerned one way or the other. 

Deformer Delay

For some time there has been a concern about the initialization time for the Deformer. One person recently testing could not see any significant hit from using the Deformer.

Some have suggested a pile-on test to see what happens when 30 residents use the Deformer Project Viewer and wear Deformer enabled clothes. That would be an interesting test.

There is some concern that those with single core CPU’s would see a mini freeze as the Deformer initializes. Those with multi-core CPU’s would not see the freeze as the Deformer initializes in its own thread which can run in another core, parallel processing.

Others are claiming the Deformer takes 10 minutes to initialize… I haven’t seen that. But, I run a Quad Core2.

So, even after Karl, the Deformer code author, tells them the initialization wait is shorter for a recalc than it is for calc, save, and download process, people are asking for a calc, save, and download process. Sheeesh!

Henri Beauchamp chimed in that the calculation does not take 3 to 5 minutes. There is some LoD handling that creates the delay. Supposedly that can be fixed. Henri sees it as a relatively simple fix.

Let’s Tell Everyone about the Project Viewer

Oh really! Let’s see we have pushed the Deformer Project Viewer on Plurk, Facebook, SL Forum, Metareality, numerous blogs, Designing Worlds, YouTube, Flickr and I’ve even IM’d dozens of those making clothes… tweet your heart out. Still people don’t know where to find the latest version of the Project Viewer. (Try the left column here)

Even those now posting in the JIRA thread are uninformed. If they would even read the thread they would know more than they obviously don’t. So, getting the ‘word out’ is NOT a neglected effort.

There is an Agenda

OMG! Some think there is an agenda to block the Mesh Deformer and prevent it coming to the grid. OK, that is the height of ignorance or serious transference.

Over on the right click the tag: Deformer. Read through dozens of articles explaining how we got here and why the Deformer stalled out.

Release and Update

Some think that releasing the Deformer as is and then updating as needed would be a good idea. It is reasonable. But, that does not take into consideration the Linden aversion to breaking content. Several revisions of the Deformer have broken all the previous Deformer enabled clothes.

In fact that is one of the contributing factors to the present development stall. Since more is being asked of the Deformer, we cannot KNOW what else needs to change. If the feature is released, that pretty much locks it in and could prevent the changes we may need.

So, releasing before we know what is needed is not a good choice.

No Problem Just Turn it Off

A Pre-265192 Mesh Failing

A Pre-265192 Mesh Failing

There are a few people that think we can have a feature to turn mesh rendering and/or the Mesh Deformer on and off. Well, yes that code could be written. But, think about what an avatar wearing mesh jeans is going to look like with either no-mesh render or the Deformer disabled.

If the mesh does not render, we will have either a nude avatar or an invisible one because of the alpha layers.

If we turn the Deformer off, the jeans will take the shape and size as made by the designer. This is much like what we have now with the Standard Size clothes. If you have a large avatar and try to wear skinny jeans, the waistline is going to be funny looking.

The image (left) shows a skirt that fits, if the deformer is used. This is what it looked like in a new viewer that would not deform the older model. This is similar to what you would see if the Deformer were released and you turned it off.

We currently have no problem with mesh and avatar sizing because we all see the same thing. So, those wearing something can see when it looks odd. With an on/off switch we won’t know what other people are seeing just what we see.

Of course this would create a violation of the Third Party Viewer Policy for the shared experience.

Alpha Layers

Some are complaining that too much reliance on alpha layers is necessary.

They obviously don’t know that in most games clothed avatars have no skin under the clothes. Those polygons are removed to reduce render load. Nor do they know how hard it is to get to spate meshes to deform in such a way that they work well together.

Weight a long tight fitting skirt. Then you’ll start to see the problems. Alpha layers are going to be a necessity Deformer or not.

Summary

More people are coming into Second Life. They know nothing about the history of the Deformer or the complexities of its design or the complexities of what it takes to make clothes work well. They want more features from it. Karl has commented about some of the preposterous requests.

Whatever, the Deformer JIRA remains an example of why the JIRA was locked down. Of course the new people are unaware that ever happened. Oz wouldn’t even comment on the JIRA Change last time I asked. So, I guess that change is here to stay.

The Deformer is stalled until Oz can get people for it. For now those people are probably working on Materials and SSA.

7 thoughts on “Mesh Deformer 2013-23

  1. As I’ve said before. I don’t think a fully optimized deformer is possible without reworking the avatar. Avatar2, while long overdue is, at best, a year away. Perhaps the problems getting the deformer working will prompt the needed work on the avatar.

    • The Deformer code will be the same whether it is used with Avatar 1 or 2. The results would be better with 2 than 1.

      I don’t see problems with the Deformer. It is about as optimized as is possible now. It does work well. The problem is what people want it to do and the demand for Avatar 2 with better topology layout and weighting.

  2. Well, one problem with the need for a corresponding alpha layer for mesh clothes is that you can’t use the alpha layer on an avatar that is itself all mesh. I have a boy mesh-shape that I sometimes wear, one of the two that contour themselves as close as they can to the legacy shape. It of course needs an alpha layer itself to be worn, though. But the only clothes I can wear with it are the single set of mesh clothes that came in the box with it and were made specially to go with it. And even then, if you look closely, there’s a tiny bit of poke-through at the armpits of the shirt. I can’t wear most of the other mesh clothes I have (not that I have a lot) with that shape because, of course, they have and need an alpha layer… which applies itself to the legacy skin, not the mesh body.

    • I think it is pretty much only Second Life where people try to layer mesh clothes and expect them to work. Designers in most games remove the mesh under the outer clothing and avoid the poke through problems.

      Using alpha layers with user mesh just increases the texture use. Actually the same is true with system mesh (like the avatar).In user mesh the efficient solution is to modify the texture to have transparent sections. No mod stuff blocks that. I’m not sure adding alpha layers to user mesh would solve the no-mod problem.

  3. People still use single core CPUs in second life? Then expect good performance? While it is bad design to assume everyone has cutting edge workstations with water cooled cpus and multiple graphics cards, by the same token doing something because *someone* out there has an old machine that they want to work is an unreasonable demand.

    Yes, I have had SL system specs outpace one of my systems. I dealed and used my new machine (my old one was a backup, I could live without it but was nice to have it around).

    • Any business is going to try to provide service to as many people as possible. Those with older hardware already feel the Lab is ignoring them. Developers want more advances. The Lab is making judgement calls. So, far their choices seem rational… even if I’m not always happy with them.

Leave a Reply

Your email address will not be published. Required fields are marked *