Land Impact Estimate in Blender via addon scripts?

[Jusden Jonstone]

I have this crazy idea, that many may think is impossible or totally not feasable, but essentially I have considered and done some reseach looking into the possiblity of developing a addon initially for Blender but possibly for other 3D Programs in the future that would enable users to calculate an estimate of the SL Prim Equivalence or Land Impact of a mesh model one is working on from within Blender itself. Bear in mind I know, getting an exact estimate could be very very challenging, but I do see some possibility to develop a feature like this. Basically could be implemented in a number of ways with calculations of estimated Land Impact and even rough L$ upload cost estimate (L$ upload cost being a bit more difficult). Ideally I would see this as a blender addon where one could periodically while modeling get the estimates of what any given mesh model's Land Impact would be. As someone who has developed blender and other 3D program addons and scripts, I do believe this is possible, but I am curious as to how much of a demand or interest amongst the SL Mesh creation community for such a tool understanding that it would be an ongoing effort to try to make any such addon as accurate as possible. Any people that would be interested in such a tool or would find it useful if it worked and was fairly accurate? Any input, suggestions, constructive critisism, or other developers who may be interested in helping implement such a tool? This is something i would work on developing for free and open source in hopes it would help benefit both content creators but also SL as well in aiding in creating efficient mesh models.

Please let me know your thoughts. 

Best,

Jusden Jonstone :matte-motes-big-grin-squint:

[Ashasekayi Ra]

Heya Jusden ^^

I'd be extremely interested in such a tool. I wanted to make one at some point. But, my ability to sift through LL's C++ mesh code was beyond me. I know Python at an ok level, but I've just started playing around with Blender Python. So, that idea didn't go far. I'd surely be willing to help as a beta tester.

Server weight seemed to be the one that would have been easiest to calculate. As far as physics, I wasn't sure if you'd be able to predict that one since part of that is closed off. But mainly, I was most interested in download weight. It would be great to test out different LOD triangle counts and how they affect the download weight without going back and forth between SL and Blender.

[Jusden Jonstone]

Great Ash, yeah I know some of the weights will be harder to potentially calclulate or estimate than others, although depending upon how the open source mesh upload code turns out for TPV's there may be some options to get more weights etc utilizing a variety of methods. But yes even a basic tool to get what you are describing would be an initial goal and blender being open source would be how i would implement it first with possibility of expanding to other 3d programs. I have and plan to chat with some other developers about ways to go about programming it and testing anything and will definately keep you informed with any plans to move forward, etc. I think it could serve as a helpful teaching addon for people learning to make mesh for SL in blender as well. So its great to get some feedback and ideas. TY 

[LindaB Helendale]

 

Definitely good idea. From the information in wiki, it is possible to calculate the land impact streaming cost / download cost component as a function of the radius of a sphere enclosing the mesh, for any given LOD scheme, assuming you know the size of the mesh in bytes for each LOD.

The byte size of the stored mesh depends on many things, number of faces, number of vertices, UV maps, etc, and it seems that there is some compression of data too. Anyways, a statistical model with some confidence intervals could be estimated.

I have reported some experiments here http://community.secondlife.com/t5/Mesh/Prims-Land-Impact-Whatever/m-p/1195631#M8403

 

 

[Jusden Jonstone]

LindaB thank you for that input and I agree with you on your thought process about how this could be done. Interesting experimentation you have done and yes I anticipated some statistical modeling would be a part of the computed estimates. Fortunately I have a background in computer engineering as well as in statistics which is much of the reason i had started exploring this idea, but getting other ppls feedback and input on the how, what and user experiences can help generate a roadmap to follow in terms of actually writing and implementing the code to start testing.

Best,

JJ

[Drongle McMahon]

I think the download weight should be fairly easy, although probably requiring estimation for compression ratios and auto-LOD generation.  Physics weight is harder because the details in the wiki are somewhat incomplete and we can't see the code because it's only in the server.  For hulls (decomposed shapes), the code is proprietary and the results extremely variable because of the number of settings. It would be very demanding to make something that would provide for all those parameters, and not very useful if they were omitted. So I would suggest aiming initially at server and download weights only. Finish that, and then think about physics weight as a separate issue.

[Kagehi Kohn]

This would be a very nice thing to have. Though, I would definitely like the next application that gets it to be Wings3D, simply because of two factors - 1) its interface is simple, and 2) while there are some quirks and bad assumptions it makes, like assuming you don't want to work with floating surfaces, and can't UV them (like you do to create a working door in blender), there is a version that is slowly improving that includes Carve3D functions, so you can produce CSG unions, differeces, and intersections with it (makes building a lot less of a hassle, since you just cut a hole in something, with something else).

I just got through spending several days creating three distillery stacks, which didn't *quite" come out as good as I would have liked, which, once resized properly (I do wish the exporter could set units..), ended up with an impact of 39. One of them I can probably drop, one of them ran into a glitch, where it refused to upload without physics set to maximum, and the third, I think the exporter fowled (supposed to have a hole, and hollowed out inside, but part of the hole is blocked with mesh that isn't supposed to be there).

I was expecting.. I don't know, maybe 20? lol The ability to select and object, and run it through something that could have said, "About 7 for that one, 15 for the next one, 16 for the one you went nuts on, when creating a handle on it...", would have told me, "You need to rethink this just a tad. lol