Avatar 2.0

	Warning!
	Linden Lab has no actual agenda nor plans to introduce Avatar 2.0 feature. This is a residents collaborating suggestion effort.

	Important! Page isn't finished.
	Sorry this page isn't finished, don't add anything yet... I'm working on it. -- ₪ 04:10, 18 March 2011 (PDT)

Introduction

We are often to ask ourselves who we are and why are you here, lacking a defined meaning to build our own persona as we came to existence in Second Life. We may choose limited set of default avatar's shape and skin, but are still bounded to the very same "Ruth" avatar mesh, which has plagued us so dearly since 2002. It may be hard to realize it has been 8 years since any improvement to avatar mesh. Only minor changes recently were new tattoo and alpha layers. In last 8 years, game technologies has changed numerous times, more than we can count. Model contents grew with such astounding details. Characters are becoming more realistic and packed with appeal outfits. To put it plainly, we're due for a new avatar mesh to keep up with today standards.

There are no question that we could provide much more tangibleness features. To impact us in a way we communicate, portrait, and express ourselves more than what we're have been missing out. In between relationship of another person online and yourself.

"Great, but we could do that with Rigged Mesh Import."

True, and this is just another reason why we must make a determination whither it's a good idea or not. How big is the file size of this mesh? Would it slow down data traffic network in Second Life? How would Avatar 2.0 help or make any difference?

Also, we must be aware how many skin/clothes designers are going to build their own rigged mesh avatar to be compatible with their fashion line-ups. How many different avatar could or would they make? How many are willing to build a personal custom-made just for you?

"What if I don't want rigged mesh import nor the new Avatar 2.0 mesh, will I be able to keep my old avatar?"

Yes and you won't even notice the difference by untrained eyes.

NOTE: One may argue that this wouldn't be 2.0 but 3.0 instead. While it's true, this would be the third set of default avatar for Second Life. Before we knew Second Life, there was a "LindenWorld". A concept stage, which had Primitars for their avatars. It was later replaced with a proper avatar mesh, Ruth, when LindenWorld evolved into Second Life.     Avatar 2.0 would embark as second fully mesh avatar. Unlike Primitars, composed in a way as one described a avatar with set of prim-attachments without the "Ruth" mesh.

Avatar 2.0 Versus Rigged Mesh Import

File Memories Size

This discussion would not be about how much glamors we can get out of, nor how much features we could possibility get. What goes on the servers, affects everyone to use it. So we must ask which would help us deal with traffic load between clients, sim servers, network system, and Second Life's database servers. Let's break down technical spec for current information.

• Rigged Mesh
  • Probability memories size of having 1 rigged mesh on a avatar (basic nude body), at maxim prim count score of 255 prims per mesh (roughly estimate of 300,000 polygons).
    • Exported with OpenCOLLADA are roughly about 43 MB
    • Exported with AutoDesk's FBX/COLLADA are roughly about 54 MB

• Avatar 1.0
  • Rough estimate of a normal Avatar shape file via XML (which currectly used to transfer between client and server) are about 0.2 MB

While there are no concept how much more data in Avatar 2.0's shape file we would like, which could easily double up twice the setting and still stay below half of a megabyte.

However, surely unlikely majority of residents wearing a mesh worth 300k polygons for casual social life. An educated guess of a typical average polygon count for a decent avatar mesh by today standard are roughly around 20,000 to 8,000 polygons, which brings file size down to 6-1 MB. Falcon & Nyx Linden have written their own compression packet which used for uploading DAE file to the server. Compression Ratio remains to be unseen for time being.

The total memories size for Mesh Import may not weight as much, as far it goes not taking separate Mesh for all 3 customized Level of Details and its physical boundary unaccounted. May raise memories size up to 9-2MB (compression unaccounted for).

Legacy & 2.0 Skin (UV Mapping) Support

Last thing Linden Lab would like to do is making all existing skin and clothes textures incompetent and making creation more complicated for beginners. A proposal suggestion are UV map switching for Avatar 2.0 mesh. Generally speaking, to have two version of Avatar 2.0 mesh with UV map based on legacy avatar 1.0 UV map and other in a proper and professional tight to reserve many texture area of a image. Switch UV map mode via Skin layer item with a checkbox set by the creator.

(A example shown on the right to compare UV map for SL's avatar chest piece and a professional-made game character.)

There are no actual development of Avatar 2.0 mesh yet, thus there are no plans how Avatar 2.0's UV map layout. However, it's very likely it could reduce down to two textures instead of having a third texture for each skin/clothes layers. Saving Second Life another load of data transfer.

((More to come.))