Why is it so difficult getting Mesh Deformer to work?

[Myra Wildmist]

Sorry, I actually found a quick explanation to this question, somewhere, but now I can't find it, again.

I've been reading up on the Deformer project, but I'm still not clear on why exactly it's so hard to get it to work. Would someone mind explaining the basic issue, please?

I read something about the attachment points in our original skeletons being hard to pinpoint, I think, but I can't find that, now.

Thanks

 

[Frawmusl]

Well I mean just think about it even from a non-technical point. When animators make movies they spend hours trying to get the clothes to fit the movements right even though they designed the entire character. When you have hundreds of thousands of different characters all wanting to wear the same outfit, but rather than weight the clothes to each individual one you are trying to use a system that does something instantly, where as it takes hours to do on its own? At least thats how I view the deormer, Its a really powerful tool and I can understand why its taking so long, that can't possibly be easy to write. The amount of math involved and testing to get the deformer to work with thousands of skeleton shapes, along with the combination of high poly, low poly etc weighted mesh..

[Monti Messmer]

Hello, actually it is only one person really working on it and sadly it´s not on the Lindens interest list.

It would need more testing and maybe implementing in more viewers, like things are done with other stuff as the new texture model. Nothing works perfect from start, we all know, but if you work against walls ...

Monti

[Sassy Romano]

When you trawl through the deformer project JIRA, the general consensus:-

"IT WORKS AS IS! Linden Lab, why are you not implementing this, why do you make no further comment in the JIRA?"

There is no incentive for LL, there is no driver for it and in my view, implementing materials without this will lead to not just two but (if it ever gets implemented), three classes of mesh.

There will be "original mesh", undeformed, diffuse map only.

There will be "new mesh", undeformed, diffuse, specular and normal map

There will be "newer mesh", deformer, diffuse, spec and normal map

So once again, the burden is on the creator to either shrug off updates or revise older products as they wish and for the consumer, there is no simple way to determine which of the "mesh" versions they are buying because it's just known as "mesh".

Is that an issue? don't know, lets see...

[Coby Foden]

Soon we will need mesh versioning, like in software...

Mesh v1
Mesh v2
Mesh v3
Mesh v4
etc

... and a wiki page which explains what are the features in each mesh version.  :smileysurprised:  :smileytongue:

[Myra Wildmist]

Thank you everyone for the info. Very helpful.

I found this article, too. It's a bit detailed, but it does a good job of explaining the current status of deformer and the complications.

http://blog.nalates.net/2013/01/12/mesh-deformer-update-2013-2/

I just thought I'd edit this post to include a link to all of Nalates deformer info. It's below.

http://blog.nalates.net/tag/deformer/

Nalate is very informed about the subject and she has some very helpful posts on her blog. ty, nalates.

[Myra Wildmist]

I'm starting to find deformer mesh in-world. This is a deformer mesh dress I found at a store called San Toki. The owner appears to be very active in the deformer project.

If you decide to buy anything here, be aware there aren't any demos. Kinda a pet peeve of mine, but I got the dress to try it for science and all that, and fortunately it fit just fine. There are also regular mesh sizes included.

 

[Sassy Romano]

---

Myra Wildmist wrote:

I'm starting to find deformer mesh in-world. This is a deformer mesh dress I found at a store called San Toki. The owner appears to be very active in the deformer project.

If you decide to buy anything here, be aware there aren't any demos. Kinda a pet peeve of mine, but I got the dress to try it for science and all that, and fortunately it fit just fine. There are also regular mesh sizes included.

 

---

So here's the problem.  If you're using that with the standard LL viewer, then you are NOT using a dress made for *the* mesh deformer since that is not in the standard LL viewer!

Because LL are dragging their feet, some creators are trending towards using the collision bones for deforming but that's no good really because it won't deform to different breast sizes which is probably one of the biggest problem areas.

Multiple different deformer methods are doing nobody any good either.

[Myra Wildmist]

---

Sassy Romano wrote:

---

Myra Wildmist wrote:

I'm starting to find deformer mesh in-world. This is a deformer mesh dress I found at a store called San Toki. The owner appears to be very active in the deformer project.

If you decide to buy anything here, be aware there aren't any demos. Kinda a pet peeve of mine, but I got the dress to try it for science and all that, and fortunately it fit just fine. There are also regular mesh sizes included.

 

---

So here's the problem.  If you're using that with the standard LL viewer, then you are NOT using a dress made for *the* mesh deformer since that is not in the standard LL viewer!

Because LL are dragging their feet, some creators are trending towards using the collision bones for deforming but that's no good really because it won't deform to different breast sizes which is probably one of the biggest problem areas.

Multiple different deformer methods are doing nobody any good either.

---

If I understand that, properly, Sassy, you're saying there's no standard for deformer. Is that the problem you're highlighting?

I was using the latest FS version, when I took that picture, today. They just updated FS yesterday, I think. The dress looks fine.

I'm finally getting the issue with the breasts. This dress seems to fit fine, so while there's still work to be done to get deformer to work for everyone, it does seem that it's working to some degree.

Also, since I tracked this dress down by tracing the creator of some of the LL test deformer clothing available at Hippo Hollow, I'd bet it would work with the LL Deformer test viewer.

All that aside, though, I totally agree: Having a bunch of different deformer methods in the market place is going to muddy the water.

[Sassy Romano]

THE deformer is the one referred to and discussed by LL here https://jira.secondlife.com/browse/STORM-1716

This is not implemented in either LL's current viewer, nor in Firestorm.  In fact unless i'm mistaken, other viewers shouldn't have it be Oz Linden says that if LL's viewer doesn't have it then nor should others (the great culling of TPV opportunity a while ago).

It's most likely therefore that the "deformable" mesh that you've got there is just using collision volume bones and as I said, they don't work for breast slider for starters so is for me, the biggest instant fail.

Some creators are trying to get a jump on this by calling their deformable mesh by names such as "liquid mesh", "flexi mesh", etc. but it doesn't matter what kind of goey/wobbly/squishy mesh it's called, LL's deformer remains in a dark corner of a dusty room, pushed there by LL.

Firestorms recent update doesn't include it, nor does it include the new materials features for texturing.

All pretty frustrating really.

[Nalates Urriah]

The Deformer project is stalled. STORM-1716 is full of posts of people wanting it to do more than it was ever intended. Some of the listed problems were enough for the Lindens to start considering them and wondering if releasing the Deformer as made was a good or bad thing.

Then there are all the requests for better mesh layout in the avatar and the problems the poor layout ireates.

Then there are the weighting problems of the avatar. Weighting could be better. In fact one can copy the improved weighting into their viewer and take advantage of it now. Only you can see the imp[rovement in your modeified viewer, which has to be modified each time you update.

There were enough concerns from users to make the Lab want to reconsider and come up with a comprehensive fix rather than allow even more content to be made from a new process that might turn into broken items if they did change the avatar. Result: stalled Deformer.

 

One of the biggest problems from the Lab's point of view, is marketing items made from different base avatar shapes. How would people know which shape was used for which dress, pants, shirt, etc. The Lab saw this as one confusing mess and was simply not going to go there. So, until they have time to figure this out... we are stuck.

 

Thanks for pointing to my articles on the Deformer.

[Sassy Romano]

I'm not convinced that these are the issues at all.  Here's how most companies handle that:-

More features asked for than originally conceived - not in scope

Avatar with better weighting - not in scope

Which source shape was used - not in scope (use default avatar)

You will already know though that Darien Caldwell did a change to permit uploading the source shape used so that effectively means that there's data to deform off the default avatar anyway.

As for LL being concerned about the marketing of items, when has LL ever worried about that issue?  If that were truly their biggest concern, one would have thought that they might have realised that in the implementation of mesh that not all avatars were the same size and shape since that was THE FIRST instance of not being able to wear mesh that fitted and yet they said "here's mesh that won't fit you".  No, that's not their concern at all, never was, never will be.

I don't think anyone ever felt that this was going to be perfect but we've tolerated sculptie parts in clothing that clipped, nobody screamed that this was utterly unacceptable or that a few vertices looked wrong, i'm not convinced that people would be screaming that their deformed mesh was slightly imperfect either as long as the overall presentation was of a good fit on the majority of anatomically possible avatar shapes.

Since LL have also gone completely silent on the JIRA, one can only assume that it's simply a case of no appetite to continue otherwise, they'd be seeking input or if they didn't want a community involvement, would be stating that they were either considering their position or otherwise.

 

[Perrie Juran]

---

Sassy Romano wrote:

 

Since LL have also gone completely silent on the JIRA, one can only assume that it's simply a case of no appetite to continue otherwise, they'd be seeking input or if they didn't want a community involvement, would be stating that they were either considering their position or otherwise.

 

---

What motivation is there actually for LL to invest any resources (translated, "MONEY") into this.

Though it does seem odd that they don't. 

To improve the In World experience they have invested in giving us things like materials, lighting and shadows, etc, etc, etc.  But we get the impression that they don't want to invest a single penny on actual Avatar appearance.

The other side of this is that as long as people continue to buy and wear half-arsed mesh out fits and as long as there are people dumping mesh clothes made from copy cat templates on the Market, there is no Motivation.  And even if we all, Merchants and Consumers went on strike and refused to supply or buy this mesh junk, they'd still have no motivation.

Right now any effort on mesh clothes by LL appears to be lip service and publicity motivated.

I do know that any company has to consider their R&D budget.  LL is no difference.  Here it is a question of priorities and with no motivation to invest in this we are stuck.

I personally do not understand their logic.  CHUI and Materials are minor needs and concerns compared to Avatar Appearance. 

Do they not understand how ridiculous our cartoon characters begin to look when sitting on a couch textured with "materials?"

[Madeliefste Oh]

---

Sassy Romano wrote:

There is no incentive for LL, there is no driver for it and in my view, implementing materials without this will lead to not just two but (if it ever gets implemented), three classes of mesh.

There will be "original mesh", undeformed, diffuse map only.

There will be "new mesh", undeformed, diffuse, specular and normal map

There will be "newer mesh", deformer, diffuse, spec and normal map

---

Materials is not a new form of mesh. It is more then all a new from of texturing, that can be applied to mesh, to sculpties and to prims. It changes nothing about the shape of the mesh or the prim... only you have more texture layers now for suggesting height differences and shininess on your object. 

[Sassy Romano]

I know it's not a new form of mesh but the point is that there will be (ideally) lower poly mesh with materials.  

How do I know if i'm buying a diffuse only item, or a diffuse map + normal + specular mapped item?  Is it important?  Those that have documented this on their MP listings think so!  Oh that's you

When I see that "mesh jacket", i'm now going to have to take into account the methods used to texture it and that's what I mean by a simplification of old mesh, new mesh and newer mesh.

If we continue to see very high poly models with every fine detail modelled in the mesh instead of taking advantage of materials then the point of it has been missed so there *should* be differences in old mesh vs materials mesh in some cases.

[Medhue Simoni]

Exactly Sassy!

These are just excuses so that LL can do nothing.

The Deformer works AS IS!! No more work is needed, except maybe some final touches.

Another excuse is there is no interest, which is BS. That is the longest jira ever created, probably in the history of jiras. That ALONE, tells you there is more interest than any other jira they are working on.

[Myra Wildmist]

---

Medhue Simoni wrote:

Exactly Sassy!

These are just excuses so that LL can do nothing.

The Deformer works AS IS!! No more work is needed, except maybe some final touches.

Another excuse is there is no interest, which is BS. That is the longest jira ever created, probably in the history of jiras. That ALONE, tells you there is more interest than any other jira they are working on.

---

Medhue,

If the deformer works as is, what's the issue, then?

Someone mentioned LL might be thinking about reworking the base skeleton. If they are thinking about that, would that affect the deformer project?

I would think it would, based on what I've been reading, but I'm quite new to this topic.

 

 

[Medhue Simoni]

---

Myra Wildmist wrote:

---

Medhue Simoni wrote:

Exactly Sassy!

These are just excuses so that LL can do nothing.

The Deformer works AS IS!! No more work is needed, except maybe some final touches.

Another excuse is there is no interest, which is BS. That is the longest jira ever created, probably in the history of jiras. That ALONE, tells you there is more interest than any other jira they are working on.

---

Medhue,

If the deformer works as is, what's the issue, then?

Someone mentioned LL might be thinking about reworking the base skeleton. If they are thinking about that, would that affect the deformer project?

I would think it would, based on what I've been reading, but I'm quite new to this topic.

 

 

---

The issue is that LL won't progress the jira and testing. We kind of have that situation where 2 teenager sit next to each other, both wanting to kiss, but both fear the consequences of the kiss. Another Schrodinger's cat situation where the fear of the consequences stalls progress. Maybe not so much fear of the consequences but fear of choosing wrong.

I have no idea what LL has in the works. LL doesn't tell us, or even hint. If there are issues with the feature that need to be addressed, LL has yet to tell us what those issues are. Everything others have mentioned are just speculations. Those things are not really issues. When designing anything, it can always become an endless project by adding more and more and more, with each feature being less and less needed for release. The deformer does exactly what it was designed for, and actually, it does much more than was originally conceptualized.

The only thing holding it back is LL.

[Medhue Simoni]

---

Nalates Urriah wrote:

The Deformer project is stalled. STORM-1716 is full of posts of people wanting it to do more than it was ever intended. Some of the listed problems were enough for the Lindens to start considering them and wondering if releasing the Deformer as made was a good or bad thing.

Then there are all the requests for better mesh layout in the avatar and the problems the poor layout ireates.

Then there are the weighting problems of the avatar. Weighting could be better. In fact one can copy the improved weighting into their viewer and take advantage of it now. Only you can see the imp[rovement in your modeified viewer, which has to be modified each time you update.

There were enough concerns from users to make the Lab want to reconsider and come up with a comprehensive fix rather than allow even more content to be made from a new process that might turn into broken items if they did change the avatar. Result: stalled Deformer.

 

One of the biggest problems from the Lab's point of view, is marketing items made from different base avatar shapes. How would people know which shape was used for which dress, pants, shirt, etc. The Lab saw this as one confusing mess and was simply not going to go there. So, until they have time to figure this out... we are stuck.

 

Thanks for pointing to my articles on the Deformer.

---

Almost every single word in this comment is complete and total speculation or opinions.

The jira is not full of posts wanting more features. Just read it. My interpretation is that many of the people posting anything negative have limited knowledge about all the factors, and even less knowledge of how to make, rig, and weight a mesh properly. Making good mesh clothing is a skill, which means not every1 will be as skillful at it. So, we have people that can't weight a mesh commenting on how the deformer handles their mesh. They just copied the weights from the avatar and think it is going to work perfect.

A better mesh layout for the SL avatar? Who would think this is in the scope of the deformer? If LL wants to make a better mesh, guess what, we can add it to the meshes the deformer works on. So, this doesn't have any affect on the deformer at all.

"Then there are the weighting problems of the avatar. Weighting could be better." Give 2 artists an avatar and have them both weight it. Neither will even remotely resemble the other. This is again, an opinion. That said, the avatar could get a once over in weighting. Some1 already did it. It already has it's own jira. LL already looked at it. What happened? Nothing! If it was an issue, which it is not, then it's seriously a 5 min fix that LL just needs to implement.

Marketing? Really? That is not, and should not, be of LL concerns. When innovation hits the market, the market sorts these things out, not the government. Governing bodies are incapable of sorting these things out or predicting the best ways to sort them out. In every case, it's the merchants and creators/inventors that have the knowledge to sort these things out. IMHO, for the most part, there is no need to use any other shape to design for. At least for most creators. There might be some special markets, like full figured clothing, or large breasted clothing, which the deformer already allows for.

What I do see very clearly in the comments of the deformer jira, is many creators that don't understand what needs to happen to get their meshes to fit every1. They don't or did not understand that their mesh is going to change. It has to change to fit every1. These creators literally expected that the shape of their mesh would not change. Basically, they expected that the deformer would perform magic. As some1 who has dealt with mesh clothing that deforms to almost any character on other platforms and programs, our deformer works as well or better than any I've seen or used, especially considering that the SL avatar is so low poly.

[Myra Wildmist]

Thank you everyone for responding to this post. You were all very helpful, I learned a lot from all your posts, and I think I actually understand what's going on. Just thought I'd recap what I'd consider the two main points to take away from this thread:

• In world deformer items, such as Liquid Mesh, aren't real deformer mesh, because the real deformer mesh as identified in the JIRA would require a update to the viewer to accommodate the deformer implementation.
  
  
• Currently, the status of the Mesh Deformer project is unknown and appears to have stalled. LL isn't saying whether they plan to advance it and any speculation on why is just that - speculation.

Again, thank you, everyone for your help.