Gaia Clary

Some remarks: The file name says: it works in blender 2.70 or newer so it should also work on Blender 2.76 The In world Shape does not help me to inspect the problem in the required details. I can only help in this case by taking a look at the blend file. We provide a ticket system for this kind of help (see the machinimatrix website) Please note that when i ask for blend files then i do this for the sole purpose of inspecting your issue. I have no intention to reuse your work in any way (just saying that your mesh work is treated as confidential :matte-motes-wink: ) AnaisCallipyge wrote: my version of Avastar is the 1.3-18, i use this on Blender v.2.76. I confess to not having hitherto imagined that the problem could come from there but now that you mention it ... I think that basically, this version of Avastar is dedicated to Blender 2.70 ? For the name of the entire file is this: Avastar-1.3-18_blender -2-70 For the shape, yes, it is exactly the same in Blender and SL ! Ok, I'll send you a copy of my body Mesh ! ;-) EDIT : Well, Gaia, I sent you my shape on the Beta grid ! Thanks again ! ;-)

When looking at the rig, i suspect you use Avastar? If so, then: Which exact version of Avastar do you use? Is the Shape that you use to make the screenshots in Blender the exact same shape that you use for the screenshots in SL? If the Shape in Blender and the Shape in SL are exactly the same, then i am interested in getting a copy of your blend file for inspection (Avastar has a ticket system for this on its Website).

Can't you put your particle emitter into a static object and then use one of the Avatar Attachment points instead of adding your emitter to the rigged mesh?

Drongle McMahon wrote: Thus in Blender the face numbers can be controlled by the material names. I must admit that i have not looked into this since a long time. So the SL Importer nowadays uses the Material names for ordering? And now (in blender) the order of appearance of the materials no longer matters and the new feature for re-ordering Materials has become obsolete ? [edit]: i just read further up in this thread that indeed now the texture face number is derived from the material order after materials have been sorted by name. Now i wonder since we (Blender) have the option to export objects in alphabetical order, should we also export materials in alphabetical order? Would this be beneficial in any way? Probably it does not matter...

I believe there are 2 situations when MAV_BLOCK errors appear where you can not immediately tell why they show up: When you create meshes with more than 21844 triangles per texture face The problem is that the SL Importer then silently splits the texture faces into smaller chunks. this creates extra materials. And as far as i understood, those extra materials are not necessarily recreated on lower LOD's. Consequently you get missiong material issues and MAV_BLOCK failures. Solution: Avoid texture faces with more thn 21843 triangles. When you create small objects with disconnected sub-meshes: In that case it looks like the problem is because the physics engine removes triangl;es with too small surface. this again can lead to removall of materials (although i did not exactly understand the reasoning). However, this seems to work in all cases where anything else failed: Solution: Add a custom physics mesh (a simple cube will do) and see if the MAV_BLOCK error goes away.

The following applies only if you used Avastar to import your XML Shape to Blender: Select the lower Body of the Avatar mesh Open the Toolshelf - Avastar Tab Locate the Appearance Panel in the TRoolshelf Select the "Shoes" Section Move the Heels sliders to 0

Hi, ChinRey Thanks for commenting and clarifying. One remark: when i mentioned "arbitrary primbuilds" i meant this as "unlimitted possibilities", so i am absolutely positive about prims. Concluding from your remark i fear that "arbitrary" can have a negative co notation. Sorry i was not aware of that. Please just read: "Prims are simple blocks for building almost unlimited creations" Now i am still very curious about the Maya avatar rig. Is that some official human avatar provided by Autodesk? Or is that a Rig created at Linden Lab by using Maya? Is there any detailed information about that rig available? And regarding the animation compatibility: As long as we talk about animations for human rigs, i am very confident that a motion transfer tools is available for that, like this one for example: http://blog.machinimatrix.org/avastar/help/motion_transfer/ While the example transfers any human rig to the SL Rig i see no problem to provide the opposite (transfer from SL Rig to Sansar Rig) as well.

Sami Buscaylet wrote: Q2. I don't really have a problme with the above if its a yes, however I do have a problem with it if my stuff wont be exported to my account on it if i choose to go to this new version of SL. So my question is Will i be able to take my stuff from (specially the newly bought) onto the new SL? If so how would this work? (I'm going to be really angry if i can't because I've just spent a lot of money on the new mesh body products.. if LL doesn't plan on allowing us to move our avatars to the new game and is planning to close SL I would expect some sort of compensation for my lost inventory). I have been thinking about Sansar for quite a while now. But unfortunately i am not amongst the lucky people who got more insight into Sansar. So all my thoughts are based on assumptions and rumours. However, lets guess a bit...: Mesh As far as i know today, Mesh creators will be able to import arbitrary Mesh models to Sansar. So all newer mesh based creations have a good chance to get into the new world as soon as the creators decide to do this. Then it is up to the creators if they arrange discount or free updates in the new world. Rigged mesh For Rigged items i am unsure where things will end up. We probably can expect to see a different Rig. But maybe creators will be able to create arbitrary custom rigs. Then it is possible to upload rigged mesh for the SL Avatar. But then we also expect that the new Avatars will come with more possibilities. And then who wants to run around with a SL Avatar when the Sansar avatars are so much more pretty ? (ok, this is all just thoughts) Sculpted Prims Sculpted prims are likely to be not available in Sansar. After all they have been a workaround to get "mesh" into SL long before SL supported true mesh. There is just no reason to add sculpties to Sansar. Regular Prims What will happen to Prims? I can not tell. They are simple building blocks for arbitrary creations. I myself have been learning a lot by making prim based creations. Hopefully we will see something similar in sansar. However in principle we could convert Prim builds into Mesh builds (and there are tools for this already around in SL) and thus migrate a lot of creations over to Sansar. But this is just in principle. I do not know if it makes sense. It depends on what building capabilities sansar will provide to users. Scripted Items It seems like migrating scripted items will be a challenge because it is likely that Sansar uses something different from LSL. Textures, images Probably no problem to get those into Sansar Audio, other media No idea Animations Well, this is very depending on how rigged Mesh items will be supported in Sansar. It only makes sense to migrate animations when the target rigs can handle them... After all human animations should be transferable... For sure this is a topic where we will get a bit of work when Sansar steps into the public :matte-motes-sunglasses-3:

maybe you forgot to set the Operator preset "SL Rigged Mesh" which is available at the top of the Collada export panel, see here: http://wiki.blender.org/index.php/Doc%3A2.6/Manual/Data_System/Files/Import/COLLADA

Either your armature's rest pose is not the T-Pose. Or your armature is not rotated according to what the SL Importer needs (avatar facing to positive X) Or your armature is rotated correctly but you did not apply rotation. things like that ... Some mesh developer kits are made for Maya and the creators have just exported the Maya rig as collada file. You need to adjust the Rig to blender. You could use the linden lab fitted mesh template blend file as guideline for how things have to be setup.

Hi; I can not give you immediate solutions, but maybe it can give you some ideas about how to proceed to find your way to your success... First of all: Fitted Mesh is not so easy to master because of a couple of things (whoewver wants to add more to the following list, please do :matte-motes-wink: : Fitted Mesh is a method that uses only Skeletal Animation to form the Avatar shape. While the SL Avatar uses a combination of Skeletal Animation and Morph Shapes (Shape keys in Blender) This results in serious problems when you try to use fitted mesh to adjust your meshes to the default SL Avatar Fitted Mesh uses a subset of the Avatar skeleton, the Collision Volume Bones. All of these bones are clamped to corresponding "classic" mBones (those 26 blue bones in Avastar which we had to use until fitted mesh became popular). So the Fitted Mesh bones never move independent from their classic counterparts (However the Pecks, the Butt and the Back bones need some additional thoughts...). This is actually good news, but you need to know exactly how to work with these bone pairs to get the best out of the system. The SL skeleton allows no more than 4 bone weights for any vertex of the mesh. If you do not know how to create a good weighting (that uses a minimum number of bones on each vertex) then you are likely to mess up your weightmaps when you export the mesh (because then only the 4 biggest bone weights survive on each vertex) The classic mBones can be moved around by rotating their joints (movement animations) and they can change their length by using the Appearance Sliders. Some bones (head, neck, ...) also can be used to scale up/down the nearby mesh parts. While fitted Mesh bones can only be used to scale the volume of the nearby mesh-parts up or down (by using appearance sliders) When you combine collision volumes and classic mBones then you can get all sorts of interdependencies with animation AND shaping your attachments. In other words: Welcome to the weighting hell :matte-motes-evil: We (the Avastar makers) have been looking for ways out of this and i believe that we are close to some point where the hell gets way cooler than it currently is, see here for the reference guide (currently work in progress): http://blog.machinimatrix.org/avastar/reference/creating-fitted-mesh/ But this is all still in its Alpha state and it will become available only with the next Avastar release. In the meantime you could check out the other reference guides on our Website where you can learn a lot about how to work with the SL Avatar in general, and a bit about how to get fitted mesh to do what you want.

for avatar attachments it seems very common to use a super simple LowestLOD like a cube (12 tris) or a thetraeder (4 tris). This alone reduces the streaming costs significantly. Maybe this solves your issue already. The reason why this works is because for some reason the LOD switch happens later for Avatars compared to rezzed items. Thus you must be extremely far away from the avatar before its attachments get switched to the lowest LOD.

Blender does neither support child faces nor child materials. I remember that i have seen some weird behavior when the mesh contained many duplicate verts AND more than 21844 faces (triangles) on one material . Then the material assignment sometimes becomes completely weird and it may look like material inheritance does happen.

We try to describe the full process of making an attachment here: http://blog.machinimatrix.org/avastar/reference/create-an-attachment/ However this document has become a bit long and probably it also has become a bit too complex over the time. However i always invite users to talk to us. Any feedback (especially from users who have dificulties with the documents!) is very much appreciated so we can learn where we need to change the documentation for better understanding.

Please create a simple example blend file which demonstrates the problem and create a report at https://developer.blender.org/maniphest/task/create/?project=2&type=Bug If this is possible for you then you can add the project "Collada" to the report and if you like you also can assign me to the report. It ends up on my desk anyways :matte-motes-sunglasses-1:

The support for fitted mesh in Avastar has been very "basic" up to Avastar 1.5. Technically it works and you could do fitted mesh all the time with Avastar and i know a few who tell me its as easy as it can be. However for me the detailed process for creating "good" fitted mesh has so many oddities that i could not create a good fitted mesh documentation yet. We first had to create some reasonably stable fitted mesh tools. And this is what we will pack into the Avastar 1.6 release. We are currently working on the documentation for the new release which will include a chapter about how to work with fitted mesh. And we are preparing a beta version which hopefully can be made available soon (probably sometime in october 2015) If you are very familiar with Blender and Avastar (classic rigging) then you could give it a try and test Avastar 1.6 alpha: http://blog.machinimatrix.org/2015/06/20/27118/ and watch the playlist at: But this should ONLY be used for testing purposes!

Devriv wrote: Delete the two files “startup” and “userpref”. Then copy these two same files from the backup folder and paste them there as replacement. Then close everything, restart the computer (not necessary but a good idea), and open Blender to check if the changes have taken effect. Please take care that the blender versions are the same when you do the copy. It might or might not work otherwise.Well, normally newer blender versions can read older configuration files, but sometimes this can get messy. And it is really not necessary to restart the computer. But restarting blender is indeed necessary to make it fully work.

I don't think that you can export your character in the "current pose". What you need is to find the bvh file of the current pose and import that to blender as a new action. Then you can aply this action to your armature and scrub over the timeline to find the pose that you want to use for your model. Here is some documentation to achieve this with blender/avastar: http://blog.machinimatrix.org/avastar/help/motion_transfer/

Here is a quick answer for blender: Pure blender: Apply armature modifier, apply shape keys, then export with Avastar: Freeze Shape as static object, then export

possible causes are: The origins of the Rig and the Mesh(es) differ You can move (object ->transform) the origin of your meshes to the origin of the Rig before you bind The Scaling and Rotation of The rig or the Meshes is not normalized You can try to apply scale/rotation of the meshes before you rig them You scaled the Avastar armature in object mode and not in edit mode This is actually not a mistake However you should then Export with "Apply scale to Armature" in the advanced export attributes section of the collada pannel. Importing with other viewers than the SL Default viewer appear to have issues at times. I always recommend to use the SL default viewer for importing rigged mesh. If that fails then something is definitively wrong with your item. Depending on your special kind of avatar you also can try to add some hover height into your specific animations for your quads, although this sounds like a workaround to me. On a second thaught, maybe you did just that and that could possibly even cause the issue. All of the above are just wild guesses and the true cause is yet something else. Maybe your meshes are just too big or the Object center (pelvis) is placed at some weird place, many things can go wrong...

there is one nasty trick that you could use, however it also has a downside (see below): Basically the relative locations of the sculpty vertices is encoded into an RGB color map (the sculptmap), where each axis can have a color value between 0 and 255: R (0-255) G (0-255) B (0-255) Normally the entire color range is used for sculptmaps. And because of this a sculpty always rezzes with a bounding box of <1,1,1> (a 1 meter cube) But nothing keeps you from reducing the color range in your maps. So, when you reduce the range of used numbers on an axis, then you can manipulate the object scaling in that range. For example: When your sculptmap for example is restricted to color values in this range: R (0-255) G (0-255) B (64-192) Then your object will get the visual dimension of (1.0, 1.0, 0,5) However by reducing the color range you also reduce the already bad spatial resolution for the possible vertex locations. Remind that you have only 255 possible vertex location on each axis when you use the full color range... However you have to calculate the range by taking the relative scaling into account. And because reducing the color range can have negative impact on the look of your sculpted prim (it gets more and more bumpy when you reduce the color range), the better option is to rescale the sculpties after they are imported into SL. You could do this by using an LSL script where you only need to specify the bounding box size. In blender we have a tool for this, which creates LSL scripts for generating and scaling sculpted prims in world. I do not know if such a tool also exists for Autodesk products. Well, you need to think twice or even more often about all of this. But once you got it, all becomes rather easy and understandable, even the 'weird" features of sculpted prims begin to become logical :matte-motes-sunglasses-3: after you stared at this for a while. Here is an introduction into sculpted prims (sorry for the very very bad voice in that video, i aught to change that sometime in the near future)

Hi. I never heard of a "link mirror tool". So maybe you can explain briefly what a link mirror tool does? Is it about mirroring meshes? Is it supposed to work inside a 3D tool (Blender,Maya,...)? Or is it supposed to work in SL (as LSL script)?

The bind process does never change the topology of a mesh. Actually the binding normally does not change the mesh at all unless you have enabled the option "Alter to Rest Pose" which does not make any sense in your particular case because your mesh IS already in Rest Pose. It is possible that Avastar unintentionally changes some display settings (then you actually have found a bug). If you can not figure out what happens, then you can always send us a report to our ticket system. We will take a look at your mesh and find the reason why this happens.

Alwin Alcott wrote: you have to be at the creation section of these forums, there you have a bigger chance to find one who wants to help you. I always thought this Forum was the Creation section... Am i wrong? Here is a small collection of videos that might help to get your models textured: http://blog.machinimatrix.org/blender/start-here/ You probably can step right into the "Materials & Textures" Video: And just to keep you from falling into the Upload pitfall: Here is an article about uploading your mesh and textures to SL: http://blog.machinimatrix.org/avastar/reference/first-steps/ Scroll down to the bottom of the page and read the chapter Import your Character to SL

Ezkabach Sorciere wrote: I had to select "avatar" in the modifier menu and then the armature was applied. I then used the collada export option in the Avastar tab, making sure to select the needed selection boxes. Can you be a bit more specific about what you mean by "applying the armature" and "Selecting avatar" ? Actually when you apply the armature then the rig gets applied to the model, thus you get the mesh frozen in the current pose and the armature is deleted. Which would make your mesh a static model (no longer rigged). So you must have done something else... Maybe you previously disabled the modifier somehow and your activities enabled it again... After following the uploading tutorial, the quad avatar was uploaded successfully and an AO fixed the front leg crossing. The front leg crossing is a "normal" behavior for quadrupeds. When you just wear the mesh after upload then you get the crossing of the legs. (which is also indicated in the SL Importer's preview window). The reason is because you have bended the front limbs by 90 degrees. while the standard standing pose "assumes" the arms are in T-Pose... The only way to get rid of the crossing is to create your own animation (pose or animation sequence) and use an Animation Overrider in SL. I wished the SL Mesh Importer would allow to use a user defined test pose instead of the default standing pose during import or even allow import of Animations for the rigged mesh together with the mesh. But i am afraid this wish will never come true.