I am using Medhue’s video for my index. He has one of the new 21:9 ratio screens. So, you’ll likely see a big letterbox display like I see on my 16:9 screen. Medhue streams the meeting live. The resulting video gets saved as 720p video, which is fuzzy if you take it full screen.
00:00 – Medhue setting up and getting the stream going.
03:30 – Vir Linden says they are submitting a project viewer to QA soon. Provided they do not find something horribly wrong, we will soon have a project viewer. I think there is a good possibility we will see a viewer in October… if things go well.
04:00 – Vir is working on transform matrices for attached animesh. He feels he has it to the place users can start helping with the testing. So, a project viewer will soon release.
Transform matrices are about how the animesh’s vertices get moved as the avatar moves and the animesh animates.
04:50 – Alexa Linden has spoken with the moles about creating animesh content. This will give users items they can test and play with as soon as animesh is released to the main grid. The moles have started work. We’ll hear more over time.
06:00 – Attached animesh transformations was the big blocker holding up the release of the project viewer. Now Vir is working on limitations for animesh. So, the number of animesh attachments to the avatar (currently 1), land impact surcharge, avatar rendering cost surcharge, and something else Vir couldn’t remember at the moment. Basically, check what the performance impacts are and working out what the render costs should be.
These are limits for initial testing on the ADITI grid. Vir and other Lindens are clear that they initially start with the most restrictive limits. Then if it makes sense, they will loosen the restrictions if possible. Meaning of the render impact is less than they imagined, the limits will loosen.
07:40 – Mesh baking service changes are in QA. Then a load test is planned to see how the change from 512×512 to 1024×1024 textures affect the servers.
08:20 – Wiki Page on animesh has not been released. We likely won’t see any wiki stuff until the project viewer is out.
09:16 – How animesh attachments will work? Same as any other attachment. It will be the root of the animesh that attaches to the avatars attachment point. Then one can edit the animesh’s position and rotation on the avatar.
Vir was also asked if animesh could be attached to the HUD points. For now, yes. But, doing so does not make sense to Vir and he thinks they will turn that ability off.
So… if you have an idea for a use, let Vir know.
The problem is it would be complicated to make work. We are trying to shoehorn a 3D thing into 2D screen space.
12:30 – How does the 20k poly limit for animesh work? If you have a 15k poly mesh and add a 6k poly (total 21k) as a linked object the system would reject it as over limit.
The system will check animesh at upload for the 20k limit. Also, when a link-operation is performed the system will check for the 20k limit.
Limits are checked on the server. So, no third-party viewer workarounds.
14:15 – Is there going to be a comprehensive API for placing textures into the bake-stack? This is a question about the new bake engine that will handle attached mesh items similar to how the classic avatar clothes are baked.
With mesh bodies, we will need a way to control how layers stack and which textures are on which layers.
Vir says an API is not part of the current bake engine change. This stage is about increasing the resolution.
14:40 – Since the server knows about poly counts to perform the link-test for animesh, would it be possible for a script to get that information from the server? Vir thinks it a reasonable request. However, it is not part of any of the current projects.
15:00 – 20k of polygons for animesh is not enough!?! Vir says that may change. The problem is the high poly count content being used now that is impacting performance. Vir won’t say 20k is the final limit. BUT… whatever the limit, it will be low enough people will have to learn how to optimize their animesh content.
16:00 – The animesh Land Impact Surcharge is 200. Vir feels this is high enough to limit the use of animesh but, still allow a reasonable number of animeshes in a region. As with other limits, they may change as the Lindens can collect more information from actual use.
Vir hopes 200 is too high. This is a starting place. It is a starting place limit for the ADITI testing.
18:40 – How is the bake server handling decompression of mesh details? Vir tells us the server does not decompress the mesh details. It uses a size based on an estimate of mesh size made on the size of the objects storing the triangles… These are the same values used for land impact and streaming costs.
I’m not sure I understood what is being said. I take it that rather than having a size as in height, width, and length or even the number of polygons we are talking something more like file size. A file size is sensitive to the number polygons in a mesh. So, a 10k polygon earring would be ‘bigger’ than a 64m cube house made with 12 polygons.
20:00 – Is animesh going to be a bigger load? Yes. Previously the system only had to look at meshes being rendered in a scene. With animesh, they will have to look for attached animeshes too. They will be testing the performance impact.
20:40 – Can users get a high poly count for testing? Vir doesn’t think so. He is concerned about overloading the test servers. Current limits are not based on data. But, estimates.
Medhue points out that a lot of the existing content will not be usable as animesh because of the polygon limit. A 50k poly limit would allow more existing content to be used for testing.
Discussion runs on what the limit should be and why testing limits should be one way or the other.
29:00 – Vir clarifies the Bake Server project. Phase I is to increase the resolution. Phase II will be to build in the features to allow the attached mesh to use the baking service.
Then back to the poly limit discussion.
Vir points out that the limit is set on the server and tells the viewer. So, there could be regions with different settings, which would allow more varied testing.
Also, most of the load from animesh will be on the viewer side.
38:00 – Baked Mesh is planned to work on animesh. Phase II of Bake will add the ability to use the service for animesh.
More limit discussion. Generally saying a limit of 50k is more reasonable.
51:30 – More discussion on how limit may be applied.
Lots of discussion in chat…
54:50 – Animesh Project viewer is just going to QA.
More discussion on limits.
58:00 – Vir summarized limits discussion. There will be limits. We do not know what the final limits will be.
No CC UG meeting next week.
Vir hopes to have a project viewer by the next meeting.