This was the first meeting in 3 weeks, Vir Linden, the meeting facilitator, has been on vacation. Next week there is no CC UG meeting because of the monthly Linden Staff meeting conflicts. So, savor this one.
Medhue streams the meeting live on YouTube. (With his new 21:9 ratio monitor.) The video remains on YouTube so, you can watch it at your convenience later or live as it happens. I usually record the meetings also, but this one I won’t post as I don’t have the time to process it and I see no point in having two raw footage videos. Plus, too few watch it anyway. About 3 times or more people catch Medhue’s video. So, I’ll probably skip publishing future processed versions.
I provide an index to the videos because I often want to refer back to something in the video. Text search works thus my indexes. AFAIK, there is no searchable way to find information in video format.
01:00 – Vir Linden is talking about the Animesh Project Viewer. This week three LSL commands are now appearing in the SL Wiki, being new.
- llGetObjectAnimationNames http://wiki.secondlife.com/wiki/LlGetObjectAnimationNames
- llStopObjectAnimation http://wiki.secondlife.com/wiki/LlStopObjectAnimation
- llStartObjectAnimation http://wiki.secondlife.com/wiki/LlStartObjectAnimation\
These show in the wiki as feature requests. Until the project viewer is released, they cannot be used.
There is no ETA for the project viewer. While it is way off, I think there is a good chance it will be this year.
03:15 – Alexa Linden updates us on Animesh and her testing during Vir’s vacation. It is also at this meeting that Alexa sort of hints at the Lab officially moving toward calling this new object animation feature Animesh, which until now has a primarily users’ designation for the feature.
One of the things the Lindens will do is provide test content for when Animesh rolls out. So, that has been accepted as a good idea by management and the LPWD (the Moles) has been given the project. I am not sure what they can do before a project viewer is out… planning?
Example Content – This shows animesh in action,
These were the zombies and dancing bears (bears?) Alexa used to test performance. These are the avatars we have previously seen in SL. But, these do not have people in them. They have been animated via the new animesh feature. So, they have no soul… (05:00)
06:00 – Performance information. The video is as Alexa recorded it in-world.
07:00 – Polishing positioning of animesh. As of now, it isn’t working well. Also, when used as attachments they are not working right. They do not move with the parent avatar.
Plus, the tools for controlling animesh attachments need to be devised.
Vir is not shooting for a finished viewer. He is going for one that works well enough we can get a feel for animesh, see problems in our use, and provide feedback. He wants that out soon.
Remember. Projects viewers are the beta versions prior to RC status. Some are closer to an alpha version than beta versions and that is probably going to be the case with the Animesh Project Viewer. We’ll have to wait and see.
09:15 – Asking about more controls for animation playback. Vir says more controls are in the To Do list. But, not part of this project. He expects the LSL functions they plan to add it should be possible to build HUD’s that provide the control.
11:45 – Question about whether or not the ability to use Animesh for attachments will extend the development time. Vir thinks it isn’t that difficult as it sort of works now, so… not significantly. This comes up again about 21:00.
13:15 – What will be needed to import animesh? Vir says basically the same as importing an avatar now. For the animations to work they will have to be in the animesh objects’ inventory.
15:10 – Transform matrix … this is about how well ATTACHED animesh objects’ animations will be able to follow the avatars movement. Vir says it should be fine. It is already sort of working.
15:30 – Increasing the size of the skeleton… If the current avatar has 150 bones and you add an animesh, it will have 300 bones. Is that going to be a problem? Vir says it will make a difference to the overall scene. But, whether the animesh is free standing or attached to an avatar makes almost no difference to performance.
15:50 – Will there be a size limit on animesh? Probably. Vir says they do not yet have one. Riva asked about being able to animate boat sails… these are large 50+m…
Vir explains he was thinking in terms of polygon count limits. Bur, size should be looked at too. Since there are already limits on what can be done with the avatars (size limits for distance between parts of avatar) we will probably see the same limits on animesh skeletons.
Once animesh moves into testing the Lindens will consider use-case feedback regarding size.
20:40 – Remote control of Animesh… well, we have LSL scripting that listens to chat commends… So, HUD’s and other control things should be easy enough.
21:00 – Animesh as Attachments adding development time… Vir does not see it as a big issue. The primary user case they are considering is the desire people have to attach animesh pets. So, they plan to have that ability in first release.
21:55 – LoD constant… came up… no response. Earlier it has been discussed. Animesh LoD will be like mesh body LoD…
22:10 – Load testing? Alexa had 320+ dancing zombies going at once. Walking in the region she saw no performance drop. She did see some interest list glitches. The same glitches we see now in a dance club. If you zoom way out, you see the avatars start to dance like jumping ants… speeded up jerky animation.
Vir thinks because the dancing bears all share the same UUID’s there is a big performance enhancement form that. More testing is needed. But, this suggests to me this is not going to be the heavy performance hit I was expecting.
25:30 – Compare 300 dancing bears to 40 avatars. No clear comparison. But, there is way more work to keep avatar info updated than for updating animesh. Think about the key and mouse information feeding into the system that updates the server on what the avatar is doing and how all that info has to be relayed to all the other users in the area. That is NOT the case with animesh objects. There is no keyboard, mouse, chat, or other info coming from the object. The little bit of messaging is nothing compared to an avatar. So, even 320+ animesh characters are way less scene impact than a couple of avatars.
26:45 – Mesh baking, still working on updates to the baking service. But, the programmer was pulled off to work on something else. Current works sounds to be more on getting the viewer working with the larger textures. Also see video at 48:00. More details about what is required and what is happening.
28:00 – Will animesh characters be able to cross regions? Yes…
30:40 – Herds of elephants and cats…
31:15 – How are Animesh LoD’s handled? The same way as avatar mesh objects. But, may be not the exact same equations.
Some think animesh will do wonders for doors and windows. I don’t see it. But…
33:35 – Physics – Avatars when animated do NOT animate their collision bubble. So, waving arms do NOT bang into things. Animesh will not be physical. You might add a cube for a physical layer. But, the cube is not going to animate.
You could even have multiple physical things in the animesh linkset. But, they are not going to animate… this is confusing. The animation may move those physical ‘things’ in the linkset, but as I understand it, that movement will not be seen by the physics engine. So, while those 50 foot tentacles may wave all over the place they won’t bang into things.
35:00 – Riding the elephant… nothing new. Just reiterating what we already know.
36:15 – Animesh and avatar movement speeds. If an avatar rides an animesh or attaches an animesh could we change the walk or run speed of movement?
38:00 – Animesh and Pathfinding? Vir hasn’t tested nor has Alexa. She’ll check with Dan. But, should work as animesh id just an object.
38:50 – Adding AI for animesh? Follow work for animesh is to support more NPC features like wearing clothes. AI is being left to the scripter for now.
Pathfinding problems exist and are not animesh specific. People don’t build for and seldom update their regions for Pathfinding. So, Pathfinding objects often get stuck. However, smart programming gets around that.
Vir considers the Pathfinding fixes outside the scope of Animesh.
41:30 – May be make an Animesh island for new peeps coming into SL… Alexa says they will do a big announcement and put these things in the Linden Viewer Library. Plus, LPWD will be making Animesh visible.
Neat would be to have starter avatars with animesh pets…
45:45 – Grid Wide Experience comes up… I wants from users… no info from Lindens other than Vir doubts Animesh and Grid Wide will release at the same time. Vir does not expect Grid Wide by Halloween. So, you are safe. No apocalyptic zombie herds running across the grid.
48:00 Mush more information on the Mesh Bake engine.
In this discussion, it should be clear the reason for this upgrade to something as complex as the Server-Side Avatar Baking process. The current bake service is limited to 512×512 resultant textures. The mesh body avatars and clothes are often using 1024×1024 textures. The upgrade will bring the service to the same level.
Then, the hope is the mesh bodies now made from multiple layers will be reduced to a single mesh skin layer similar to the Classic avatar.
In the Classic avatar the skin, tat, underclothes, clothes, and jacket layers are baked into single texture mapped to the avatar skin. The similar process is planned for mesh bodies. This would create a HUGE reduction in the number and size of textures used by mesh body avatars.
53:40 – Tradable outfits… probably not.
55:00 – Speculation on what may be created…
56:40 – Animation speed…
59:00 – Standard JIRA disclaimer… Just because the Lindens ACCEPT a JIRA does not mean they will do it. It means it is something they want to keep in their considerations.
1:00:00 – Animate ocean? Not currently in the plans.
No meeting next week.
1:03:00 – Particle Rain – Rider Linden
We could make proper mounts without messing with spare bones or having bones affected by other bones. Exciting 😀
“If you zoom way out, you see the avatars start to dance like jumping ants… speeded up jerky animation.”
Impostors animations speed/rate has been fixed in the Cool VL Viewer v184.108.40.206, released 3 weeks ago (change log here: http://sldev.free.fr/forum/viewtopic.php?f=3&p=8575#p8548 ). It was a twofold bug (i.e. two combined bugs causing one bogus effect), and probably explains why it has been plaguing all viewers since impostors have been implemented, years ago…