Gaia Clary

Phoenix Takashi wrote: Things I've tried/done: - Renamed the bones to the recommended bone names found on Gaia's website (I used a regular Blender rigify skeleton) I think this is your problem. There is nothing wrong with the Rigify Skeleton, except that it probably does not use the same joint positions than the SL Skeleton. I believe you would be doing better by using one of the ready made SL skeletons. Here are 2 locations from where you can get them: http://lecs.opensource.secondlife.com/fittedmesh/SecondLifeAvatarSkeleton.zip http://blog.machinimatrix.org/avatar-workbench/

Please take care that currently this only works reliably with daily builds of blender: https://builder.blender.org/download/ For the Sparkles addon i am now looking to find a good workaround to solve issues with Blender 2.73 and custom normals :matte-motes-impatient:

Drongle McMahon wrote: If it's Blender, can we have an intra-object version too? i think i can add that as well, yes.

here is our first implementation of custom normals support: You will need a daily build of Blender to get all of that it fully working (the Collada exporter will have custom normals export in the daily build of tomorrow)

Christhiana wrote: What would also seem to me like a nice feature to have... would be to be able to exclude the normals of a vertex from being automatically updated by blender. A 'lock vertex normals' option. this way you could delete geometry and keep certain vertex normals as they were before the adjoining geometry was deleted. In principle this can be done now. However some tools are yet missing. I am working on a few of these tools now and will add them to our Sparkles addon. Christhiana wrote: For example you if you wanted to split an object in two, you could make a copy and delete the opposite halves from each while you have the vertex normals along the seam locked. I have the welding operator almost ready now. this one will take care about sharp edges, edge split, smooth and flat faces, etc. It should be as easy as "select the 2 objects, press the weld button, done"

i can confirm that Avastar calculates the welding wrong in certain situations (however, it works for welding head/upper/lower body of the SL Avatar). Your solution perfectly matches to what i found in the last 2 days. And indeed it has to do with how the welding picks the right or wrong normals. BTW: Regarding the future of welding :matte-motes-evil-invert: : Actually we have started working on implementing export/import of custom normals. This will be added to the Blender Collada exporter and to the Avastar exporter as well (plans are to provide this wiuth blender 2.75) But this is only half of the solution. Because you also need to get the custom normals from somewhere. So i plan to move the "welded normals" feature from Avastar's collada exporter into an Operator, such that you will be able to: Use the welding function to generate the normals (similar to how Avastar works today) Use the Collada exporters to export the Normals to SL You can expect to see this in the near future.

I have been playing a bit more with Avastar's welding feature and after testing a bit more i ended up with something that works nicely: The ring at the bottom is made of 4 mesh elements. i have placed one element on the top of the ring so you can see where the seams are (inside the green circle) The horizontal faces are all flat shaded. While the vertical faces are all smooth shaded. I exported the quarter ring element with the options: weld edge normals (to enable welding) weld all visible (to take all adjacent meshes into account) I could not see the issue that you report. So imho your problem should be solvable with normals welding. However i will now go and look into support of custom normals :smileyhappy:

I have added a sticky page to our Blog where we can collect examples for when this issue appears and how it can be fixed: http://blog.machinimatrix.org/blender/dae-parsing-issue/ I hope that we eventually can add some extra checks to the Collada exporter in order to flag problematic meshes. But first i would like to collect examples of failure.

That IS exactly what i meant, the welding still seems to take place, but there is an additional issue related to the particular light setup. And that's what i need to better understand. i think that i know meanwhile why this (what you see) happens and if my assumption is true then actually the only good way to fix it is to support custom normals (and some additional user work to get the normals to do what you want). But this will take me some time to get it figured out. Possibly i even need to think about an additional normals toolset... Well, i always have more ideas than i can realise :matte-motes-bashful-cute-2:

but it should work exactly like that. you used the edge split to make sharp edges. then you make the sides smooth shaded. and finally the adjacent edges of your parts would be welded as wanted. But maybe i did not understand your issue well enough. However i found some other issues which have to do with the lighting setup in SL. I need to go deeper into this normal stuff again, there is definitively some room for improvement. But i meanwhile believe that i better do the support for custom normal exporting to Collada first, then use that experience to improve the Avastar exporter as well...

ok, so i looked closer at what avastar does. The caveat is with flat/smooth shading. Below you see 4 setups (from top to bottom). The 2 lower pieces are made of 2 adjacent sheared cubes. The 2 upper pieces are joined sheared cubes with duplicate vertices removed. All pieces have been imported to SL with welded normals enabled: What you can see is: The Welding normals option does not work on single objects. Because of that 1 Object flat and 1 Object smooth both are not affected by the welding normals option. However they import to SL as they look in Blender. The Welding normals option only works on smooth faces! So 2 Objects flat is als not affected by welding normals. When you look at 2 Objects smooth, then you see a difference between what is displayed in Blender and what is shown in SL: In Blender you see how the objects would look without any normals trickery, so you see a hard edge Due to the welded vertex normals, the edge between the 2 adjacent sheared cubes now disappears. hence in SL the 2 Objects smooth looks exactly the same as the 1 object smooth. So Avastar indeed welded the normals as expected. And here is how you would solve your issue in your case: The 2 faces at the front side (marked with "s") have assigned smooth shading, while all other faces have assigned flat shading: Now Avastar welds the smoth faces of the 2 sheared cubes et voila, your seam is gone.

ehm... well... what shall i say... it looks like a broken feature :matte-motes-sour: ok, i count this as a bug report and try to fix it.

ah, i see. Avastar only takes boundary edges (open edges) into account. Actually it first determines for all involved objects which of their edges are boundary edges, then for each vertex that belongs to a boundary edge avastar tries to find a matching boundary vertex on other objects. When it has found a match, the vertex normals of both vertices are averaged. I believe that as soon as the custom normal export is supported we can drop the Avastar welding feature anyways. So i rather not try to support welding of solid meshes at this time :smileysad:

can you give me an example what you mean by "solid meshes" ?

Please check this page to see how to weld normals with Avastar: http://blog.machinimatrix.org/avastar/help/export_mesh/ its described in the advanced options section.

We have plans to add custom normal export to the Blender Collada exporter. But i can not yet tell when exactly this will become available. I hope for Blender 2.75 but there are soo many other things to be done as well :matte-motes-dont-cry:

Yes, Avastar has a "weld normals" function which fixes this issue upon export. But we also plan to add "export of custom normals" directly into the Blender Collada exporter. However i can not yet tell if this will be ready for Blender 2.75 or later.

Absolutely sure! We have taken care that Blender can export for Second Life out of the box. So of course you also can make avatars without Avastar. Avastar is really just a tool that tries to make some tasks in the process easier.

I believe the main problem with all of this is that users get the impression that creating mesh content for Second Life is as easy as making prim based content. However there is a lot of knowledge that you need to learn first before you get things running. And making rigged mesh (avatars or avatar attachments) puts one more level of complexity into the task. And finally, there is your 3D tool that also needs to be understood basically, before you can get reasonable results exported. Well, for Blender things are rather easy, because we have actually taken care that it can export for Secondlife out of the box. However, you need to get the right input (demo blend file) for this to work, and you need to know how to export this to be compatible for SL. The right import for fitted Mesh is here: Official Fitted Mesh downloads You may want to know the details: The Fitted Mesh Kit If you want to do classic Mesh, this helps: Avatar workbench (you then want the 2.65 version) Now you need to know how to export your rigged mesh correctly: Check in the Blender Wiki: The Collada Exporter (You want to select the Preset "Second Life rigged") Check how to import to SL: First Steps (Scroll down to "Import your Character to SL", that part works independent from any tool) That should be helpful i hope. If you are interested in more details about blender and Second Life, you might be interested in this page as well: Avacademy That page contains an unsorted set of videos about aspects of 3D topics which are more or less relevant either for making models in general and/or for making content especially for Second Life. I hope this helps for the beginning.

OpenCollada is in first place a programming framework. This project actually has added some reference implementations for Maya and 3DS Max. Those reference implementations are plugins and they should be also available with the Autodesk distributions, or at least these plugins should be readymade available from Autodesk. I only can speak for Blender here. Blender also makes heavy usage of this Collada framework, and its exporter comes out of the box along with Blender (and it works reasonably well so far) BTW, Blender developers have been actively involved in helping to maintain the OpenCollada framework. So... long explanation, short result: you should not try to make OpenCollada work for you unless you are a developer and you want to help improving this framework :matte-motes-evil-grin: Try to find the Collada export plugins for your tool instead.

What format do you use for transfering your model from ZBrush to Blender?

Could you give a short information ion which way the maya Dae exporter is by far better than the Blender exporter? Where does the Blender Collada Exporter need improvement? [Edit]: ah, its because of the bind pose matrix, i see now. Well, there is no support for bind pose matrix in Blender (yet). So the Blender exporter can not do much here. Using the custom properties is indeed a workaround. However Avastar actually DOES support a bind pose. So no hack needed there, just the plugin... But well...

can you provide an example?

What you describe sounds like some scaling issues in the mesh itself. If you like i can take a look at your file. In that case you can open a report at http://blog.machinimatrix.org/avastar/tickets/ and add the blend file for inspection. (the blend file needs to be < 32 MBytes though)

According to Linden Lab they plan that your Avatar and even your Linden$ will be recognized in both worlds.