Jump to content

Why Collision Volumes might be the better choice


Gaia Clary
 Share

You are about to reply to a thread that has been inactive for 2879 days.

Please take a moment to consider if this thread is worth bumping.

Recommended Posts

First to clarify: I have no strong opinion on this and i do not know why LL decided to prefer the Collision volumes over the Deformer. I only have some guesses and thoughts about this which i want to share with you. Actually i might be wrong with my assumptions. so please correct me if you know better. Lindens are invited for commenting :)

Collision Volume Bones are just Bones

Do you remember that LL initially was not even aware that collision volume rigging does work at all ? I believe that rigging the collision volume bones just worked because they have always been "simple bones"(joints actually). And since the SL skeleton obviously uses a standard method for animating there obviously was no extra fiddling needed (technically spoken).

Well... artistically spoken collision volume bones have a few ugly issues, but maybe these can be sorted out in the short future. just saying: its standard technology...

The Deformer's assumptions ...

The deformer seems to work on top of the existing mesh. Hence we must consider that this is probably an addition on top of the final mesh that has been calculated by the standard animation system. So the mesh deformer probably has to make a few assumptions about the mesh (and the skeleton) that it gets as input so that it can calculate the deformations correctly and thus deform the mesh as wanted...

Reasonable Conclusion... (standard vs. proprietary?)

From the above i conclude that the deformer might make future changes to the Avatar more complicated while collision volumes work in a more standard way using more standard techniques.

for example i have been able to integrate the new collision volume bones into our Blender Addon within less than 4 hours from the moment i saw the blog post until it was finally working. This includes the integration of the corresponding SL shape sliders as well.

I could do that only because the collision volumes are standard technology. And that technology is supported by Blender in the very same way as it is supported by Maya, 3DS, whatever... and at the end also by the SL animation engine.

Political choice or reasonable choice ?

I believe that "we dismiss the deformer because it was not our own idea" would be a very childish reasoning. I still believe that other arguments have much more value. I actually believe they really looked at the options and decided what makes sense for them as the slution provider.  And actually i believe (but i do not know for sure) that they are probably right at the end when they say:

"... it (collision volume bones) leverages more of the existing avatar shape system it is likely to be the more maintainable solution and to perform better for a wider range of users."

Conclusion

I understand that the Mesh deformer works nicely for many and makes it easy to create fit-for-all garments, while working with collision volumes seems to be much more complex. But i also can see how deciding to use Mesh deformer could possibly bind LL to something that can no longer move ahead toward the future. I mentioned already that collision volume weighting uses standards, while the deformer might do some propriatery (non standard) things.

Well this sounds like i dislike the mesh deformer. I do not! I just try to make some sense to the decision to use collision volumes. Actually now after its official maybe we will see some imrovements which may help us to create fitted mesh with less pain...

And again... i hope that i am not entirely wrong :)

  • Like 1
Link to comment
Share on other sites

Thanks, Gaia, for giving a fair assessment.

 

As much as I liked the deformer, the existence of collision volumes already being implemented within the viewer pipeline is a major bonus to it.  It also means that adding NEW collision bones is a relatively simple matter (assuming the code is even remotely well-designed) for LL to do.  This means extending the customizability of the fitted mesh system becomes simpler for them, even if it makes things a bit more complex for those of us trying to twist mesh capabilities into pretzels to get our meshes to do juuuussst what we want them to.

 

Now if we can just get them to expose the morph targets and such, we'll be in business!

 

Link to comment
Share on other sites

There are a few things that Collision Bones just can't do.  You can't make skin tight clothing that follows the curves of the actual avatar shape as that shape changes with different appearance settings.

There are a few things that the Collision Bones do better than the Mesh Deformer and a number of things the Mesh Deformer does better.

The Collision Bones are a lot and I mean a lot more work for the content creator.

The Mesh Deformer you just rig to the normal skeleton and click a couple of options in the upload window.

Hopefully with some work the Collision Bones can be modified so they are not so much work.

Personally I hope we can have both.  I posted a new topic where I describe how I think we can have both and everyone be happy.  While I prefer the Mesh Deformer over Collision Bones I can see the value in them just as I see under some circumstances creating using prims or sculpties can be a better choice than mesh.

http://community.secondlife.com/t5/Mesh/Solution-We-Can-ALL-Be-Happy-With/td-p/2341689

 

Link to comment
Share on other sites

As far i remember the collision bones thing isn´t a Linden baby too ? A few or only one content creator(s) used this "bug" as a feature.

However they didn´t have to put much new code inside the viewer.

I wonder if we are allowed to use the name - Liquid Mesh - because that´s what this creators used not LL. Trademark ?

I followed the tutorial on machinimatrix and only got  a mess so maybe this works better for LL, it doesn´t for me. But that could only be me ...

I´m only happy that now we have a way to make mesh clothes fit better (not perfect) that will boost the usage.

Monti

Link to comment
Share on other sites

Technically, I really can't have a serious opinion of which is better, as I have no experience with the collision bones and mesh. I haven't gone thru the workflow, nor can comment on it's end result, and I likely won't be doing any of that this year.

What I will say about the mesh deformer is that is works, works very well, and is no extra work for the clothing creator.

Where I will differ from what Gaia is saying, is that I do not see how the mesh deformer could be a problem down the line for LL, or any1 that used such a thing. The mesh deformer uses the positions of the meshes verticies. Mesh verticies are never going to change anytime soon. I don't know alot about collision bones, but I will say that it is always open to change, as technology moves forward. Mesh verticies are more "standard", than collision bones, in my honest opinion. LL could completely change the avatar and how every single thing works on it, and the Mesh deformer would still work.

In the end, is collision bones the better choice? I'd say that is impossible to determine at this point.

Link to comment
Share on other sites


Cathy Foil wrote:

There are a few things that Collision Bones just can't do.  You can't make skin tight clothing that follows the curves of the actual avatar shape as that shape changes with different appearance settings.

I see what you mean. Collision bones can not simulate exactly what the Morph shapes can do with the default Avatar, no doubt here.


Cathy Foil wrote:

The Collision Bones are a lot and I mean a lot more work for the content creator.

The Mesh Deformer you just rig to the normal skeleton and click a couple of options in the upload window.

Hopefully with some work the Collision Bones can be modified so they are not so much work.

I agree that making collision volumes is a nightmare compared to just click 2 buttons to get your mesh auto adjusted.

Actually as far as i understand  collision volume bones are nothing special at all. They are just bones like all others. However they have been introduced as a simple collision model of the character mesh. This collision model is used for more efficient calculation of collisions with other objects and avatars in the scene.

I believe that reusing collision volume bones as "shape bones" makes their usage painful. So i also agree that reworking their definitions might be a good step forward in making their usage easier.

Also i hope that we will see more bones in the future. Maybe the logical next step would be to keep collision volumes bound to their original task and introduce "shape bones" instead which actually could even replace morphs entirely at the end.


Cathy Foil wrote:

Personally I hope we can have both.  I posted a new topic where I describe how I think we can have both and everyone be happy.  While I prefer the Mesh Deformer over Collision Bones I can see the value in them just as I see under some circumstances creating using prims or sculpties can be a better choice than mesh.

As i said before i am not against the Mesh deformer. I just try to find a good reason for the decisions. Probably LL decided to use only one method because having both would constrain their future development. If they only would tell the world what they think, then i believe that a lot of questions could be answered instantly.

Link to comment
Share on other sites


Monti Messmer wrote:

As far i remember the collision bones thing isn´t a Linden baby too ? A few or only one content creator(s) used this "bug" as a feature.

The collision volume bones have been inside the Avatar since a long time, maybe even since day one. RedPoly found them in the Avatar definition files and concluded they might be usable for weighting (which turned out to be the case). So he uncovered a non documented feature actually :)


Monti Messmer wrote:

However they didn´t have to put much new code inside the viewer.

They didn't have to put any codeinsidethe viewer.


Monti Messmer wrote:

I wonder if we are allowed to use the name - Liquid Mesh - because that´s what this creators used not LL. Trademark ?

I can't tell if there is a trademark on "Liquid mesh". Ask RedPoly, maybe he knows. BTW "Fitted Mesh" is a bit of a weird name in my opinion. Actually i wonder why there must be a name for this at all. Having clothes that fit to an avatar should be the "default behaviour of clothes" ... no ?


Monti Messmer wrote:

I followed the tutorial on machinimatrix and only got  a mess so maybe this works better for LL, it doesn´t for me. But that could only be me ...

I´m only happy that now we have a way to make mesh clothes fit better (not perfect) that will boost the usage.

We have been told a year ago that Collision Volume rigging was not officially supported by LL and we better not release tools for it to avoid its usage. Actually that is why i have stopped creating tutorials for how to use collision volumes. And actually i haven't even touched this feature for about a year until 2 days ago.

 

So now i am working again on the tools and try to make them easier to use. But i also have to wait a bit for the next decisions on the implementor's side. Will we get more/other bones, will collision volume bones be optimized for usage as shape bones, will known bugs be fixed or kept, etc. ...

 

Link to comment
Share on other sites


Gaia Clary wrote:


Also i hope that we will see more bones in the future. Maybe the logical next step would be to keep collision volumes bound to their original task and introduce "shape bones" instead which actually could even replace morphs entirely at the end.


Now, this would be a huge bonus if LL would consider it. Not the replace morphs part tho. A whole skeleton tho, with facial bones, breast bones, butt cheek bones and finger bones. Maybe also add in tail bones, wing bones, ear bones, and individual toe bones. That would be a ton of bones, but would give us almost every option we'd need to make almost any creature. Breast and butt physics wouldn't even be needed anymore, as the animators would add them in. With a full facial rig, we could have a fully animated face with any option. With fingers, we could make sign language gestures, and much more. Ulitmately, I'd love to see a bone attachment system, where we could attach whole new bone sets to the default. Imagine a guy with 6 arms, or a properly working centaur. There is no reason both a new default skeleton and attachment bones done later can't exist together, and it insures that every1 making facial expressions, or any other normal avatar movement, is using the exact same bone set, and every1 can use them.

Anyways, we can dream. lol

Link to comment
Share on other sites


Gaia Clary wrote:


BTW "Fitted Mesh" is a bit of a weird name in my opinion. Actually i wonder why there must be a name for this at all. Having clothes that fit to an avatar should be the "default behaviour of clothes" ... no ?


Yes, but it's not a default behaviour, LL didn't implement it in the beginning and existing mesh won't go away just because LL have finally done something.

Therefore, there does need to be a common name.

Link to comment
Share on other sites


Monti Messmer wrote:

As far i remember the collision bones thing isn´t a Linden baby too ? A few or only one content creator(s) used this "bug" as a feature.

 

RedPolly wasn't the first to use Collision Bones to deform mesh clothing.  About 7 months before RedPolly posted their video Prep Linden asked for mesh weighted to the Collision Bones as a test to see if this method was a viable option.  

I don't know if RedPolly got the idea from this post or rediscovered it on their own.  It is clear that LL tested it and rejected it seven month before RedPolly's video came out.

http://community.secondlife.com/t5/Mesh/Need-some-samples-of-a-rig-weighted-to-the-collision-volumes/td-p/1142973

 

Link to comment
Share on other sites

Well, that's interesting indeed. How sad that this thread completely slipped through my attention :matte-motes-sour:

However i remember how RedPoly annnounced to "uncover a secret" a few months later in one of the mesh meetings. And i also remember that one of the Lindens in the very same meeting created the impression that this was new for them. (ok, i might remember wrong, but that's what i remember.)

But well, ok... So what. Forget the past, step ahead :matte-motes-sunglasses-3:

Link to comment
Share on other sites


Cathy Foil wrote:

I don't know if RedPolly got the idea from this post or rediscovered it on their own.  It is clear that LL tested it and rejected it seven month before RedPolly's video came out.

 

I don't see in that thread where it is clear that they rejected it

Link to comment
Share on other sites


Medhue Simoni wrote:

Breast and butt physics wouldn't even be needed anymore, as the animators would add them in.


So what you're saying is remvove a feature that I can control and adapt myself and replace it with one where those choices  have been taken away from me in favour of what a designer of mesh clothing thinks I should do?

Link to comment
Share on other sites


Pandora Wrigglesworth wrote:

The #1 reason why collision bones should be supported is simply that a lot of creators are already using them and have been for over a year now.

If LL followed that logic, LL would have adopted

 

  • RLV code into the main viewer
  • Phoenix avatar extra attachment points
  • Wouldn't have killed invisprims used univerally in lots of shoes

I'm not saying that they should or shouldn't but LL did nobody any favours here having first made it abundantly clear that they weren't guaranteeing to support collision bones used this way and then doing a complete U turn and presenting it as "the solution".

I'm still going with the "Not Invented Here" problem of including code from a source from which there appears to be huge internal political battles.

Whether collision bones are the right way or wrong way I am happy to watch the debate.

  • Like 1
Link to comment
Share on other sites


wesleytron wrote:

I don't see in that thread where it is clear that they rejected it

To me it seem clear that something has been rejected when LL asks for test mesh of a specific nature to test something out and then doesn't pursue or show any signs of interest in the method for 7 months not until RedPolly came out with the video.  Even then it wasn't until more and more content creators, who were tired of waiting for two years, decided to start selling Liquid Mesh that LL finally started moving on the project again.

Had Liquid Mesh not been starting to sell on the Marketplace and had others not kept bring up the Mesh Deformer in this forum, writing blogs and attending content creator meetings I highly doubt that LL would have ever implemented any mesh deformer at all.

Link to comment
Share on other sites


Pandora Wrigglesworth wrote:

The #1 reason why collision bones should be supported is simply that a lot of creators are already using them and have been for over a year now.

LL was never going to get rid of the original collision bones even if LL had decided to go with the Mesh Deformer.  The Mesh Deformer would not break any content using the collision bones.

It is going to take 2 to 4 times longer for content creators to adjust the weights of mesh weighted to the the collision bones.  I know I have had plenty of practice with the original collision bones.  Why does it take longer?  Because with the collision bone are one offset from the normal bones and two we now have to adjust the weights so that the vertices change shape smoothly when the scales of the collision bones are changed in the editing appearance menu.

If a content creator wants to use mesh clothing they have already uploaded to SL they will have to rig and adjust the weights all over again.  Only this time it will take 2 to 4 times longer.

Had the Mesh Deformer been chosen all the content creators would have had to have done, to make their already existing mesh compatible with the Mesh Deformer, would be to upload their old DAE files and in the mesh upload window select two options.

Link to comment
Share on other sites


Pandora Wrigglesworth wrote:

The #1 reason why collision bones should be supported is simply that a lot of creators are already using them and have been for over a year now.

What will happen to those already designed and sold "Liquid Mesh" clothes after the Fitted Mesh is released as ready to be used? Looks like that it's risky business to buy them at this stage as they might turn out to be unusable later on.

Quote from:

http://wiki.secondlife.com/wiki/Release_Notes/Second_Life_Project_FittedMesh/3.6.11.283899

 

"At this time, the new skeleton should be considered provisional and subject to change; we do not yet recommend selling or buying garments rigged to it. Since we may find reasons to improve it during this testing process, and any change to the collision bones will likely break garments rigged before the change, we want to make sure that we have a set of bones that we can all live with into the indefinite future before it is any more widely used."

(Underlining and bolding mine.)

 

Link to comment
Share on other sites

  • 2 weeks later...

Hello Cathy,

Just wanted to say that my first email to Prep was on 8/11/2011 asking him if I can use collison volumes without braking any ToS.When I was asked on the meeting about the date I said 8.11.2011 insted of 11.8.2011, because in my country this is how dates are formed but my thunderbird uses US format. So just to be clear, I was the one to inform Prep for this method and the forum topic came afterwards. My video was on my second product with collision volumes, the first one was jacket and my redRig was completly ignored by my customers :) I have timestamped object on the beta from August 2011 :)

Just wanted to clarify that.

Thanks

Link to comment
Share on other sites


Coby Foden wrote:


Pandora Wrigglesworth wrote:

The #1 reason why collision bones should be supported is simply that a lot of creators are already using them and have been for over a year now.

What will happen to those already designed and sold "Liquid Mesh" clothes after the Fitted Mesh is released as ready to be used? Looks like that it's risky business to buy them at this stage as they might turn out to be unusable later on.

Quote from:

 

"At this time, the new skeleton should be considered provisional and subject to change;
we do not yet recommend selling or buying garments rigged to it
.
Since we may find reasons to improve it during this testing process, and
any change to the collision bones will likely break garments rigged before the change
, we want to make sure that we have a set of bones that we can all live with into the indefinite future before it is any more widely used."

(Underlining and bolding mine.)

 

Any decent merchants would keep records, and update the product as needed. This isn't a new thing. I sell many products that I love so much that I change them and update them on a regular bases. Every1 gets the updated version. Even when I sell things through my website for Poser or Daz products, my delivery system automatically sends out emails with new download links to the updated products. LL could definitely make things easier on us merchants to send out updates. LL also hurts the process cause of their limit on how long you can view transactions, which means we all have to collect our transactions ourselves, which I don't have to do anywhere else on the net.

I just always hear people bring up broken products, and I just don't see what the problem is, unless the creator is not making money and long gone.

Link to comment
Share on other sites


RedPoly Inventor wrote:

Hello Cathy,

Just wanted to say that my first email to Prep was on 8/11/2011 asking him if I can use collison volumes without braking any ToS.When I was asked on the meeting about the date I said 8.11.2011 insted of 11.8.2011, because in my country this is how dates are formed but my thunderbird uses US format. So just to be clear, I was the one to inform Prep for this method and the forum topic came afterwards. My video was on my second product with collision volumes, the first one was jacket and my redRig was completly ignored by my customers
:)
I have timestamped object on the beta from August 2011
:)

Just wanted to clarify that.

Thanks

Thanks for clarifying that RedPoly. :)

It is always nice to give credit where credit is due.  Using the collision bones in such a way was a brilliant idea you had.

I know what it is like to have something been credited to someone else.

I guess the question is now why after asking for test rigs so long ago did we hear nothing from LL about this?

Despite what some may think I don't dislike the Collision Bones method.  It has its drawbacks just like the Mesh Deformer has its own.  If we had to have just one deformer I would prefer the Mesh Deformer over the Collision Bones.  Why?

1. The Mesh Deformer would have been more useful to more content creators.  Really tight clothes, custom mesh avatars and replacement mesh body parts are really not well served by the Collision Bones method of deformation.

2. The Mesh Deformer is way easier to use.  Just rig your mesh to the normal skeleton bones and adjust weights like we all been doing for quite some time now.  The Collision Bones at best will always be two to three times harder to rig and adjust the weights to because now the content creators not just have to adjust the vertices weights so the mesh looks good when the joints bend but also how when the joints scales change.

3. Existing mesh can be made to work with the Mesh Deformer in a matter or a few minutes.  Just locate your original DAE file, change your Debug setting in your viewer to use the SLM file and click a few options on the upload window that's it.  To do the same thing with the Collision Bones method you have to rig your mesh all over again and adjust the weights from scratch.

4. The Mesh Deformer has way more potential usefulness as LL adds more avatar features such as when we will be allowed to add bones of our own or our own morphs.  Since the mesh deformer uses the vertex positions of the avatar mesh it could have been modified to use the vertex positions of avatar mesh we created.

5. The Mesh Deformer since it used the avatar's mesh vertices to deform the content creators mesh the deformations follow the actual avatar's shape.  The Collision Bones method are if a series of balloons are attached to the skeleton that inflate causing the clothing mesh to be stretch.  This will never follow the actual avatar shape that closely.

6. Any new and improved default or additional mesh avatars with better vertex distribution, weighting and morphs LL makes the Mesh Deformer will perform even better but the Collision Bones method will not improve.  It will have the same flaws and drawbacks as it has now.

I am glad we have at least one deformer now.  I really do think it be best to have both the Mesh Deformer and Collision Bones method either working separately or in tandem.

 

Link to comment
Share on other sites

I could not agree with you more Cathy.

Last friday, in the 3rd Party meeting, which I watched on Youtube, Oz gave a reason why it took so long. Basically, he said the people that had the knowledge to tackle the issue were all working on the rendering update. This shows just how bad LL management is. The clothing deformer just needed some1 to spend a couple of weeks looking at it, and making a decision. That's it! Maybe if they decided to go with collision bones, then a few days to add the bones. They then could have gone back to working on the new rendering system, and updated the avatar when they had more information. We would have gotten workable clothing 2 years ago, but because LL can't manage well, we sat for 2 years with a completely unexceptable issue. Imagine the amount of users they pissed off. Imagine the amount of time and money they wasted. Imagine the amount of OUR time and money they wasted. It all just astounds me.

  • Like 1
Link to comment
Share on other sites

You are about to reply to a thread that has been inactive for 2879 days.

Please take a moment to consider if this thread is worth bumping.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...