New World Notes

How Offline Mesh Hurt Educational Projects in SL (Comment of the Week)

While many of us have talked about how the addition of offline mesh transformed Second Life into a consumer-focused world, educator Joe "Iggy" Essid offered this comment on how it hurt virtual world teaching projects like his:

I recall when Mesh arrived. Most of us in the .edu sector had already left, but I thought "there goes another magical SL experience": team-builds in-world. I think two students I knew could make things in Blender, and even then, it would be made outside SL and imported.

The magic of SL was always when folks interacted, in there. I didn't mind how much more realistic mesh avatars looked or how clunky the physics-engine proved; we were not building avatars or scripting HUDs, but we were making the spaces where avatars interact. I'm not a coder, so I imagine in-world building tools for mesh would be laggy and complex, but then building itself was complex already, the permissions system a mess, and the UI for SL clunky in the extreme.

Had in-world mesh tools existed, some of us in education would have soldiered on and given away Creative Commons content to colleagues, as a brave handful still do in OpenSim. Sure, we could keep using prims, but after a while, our builds would resemble the black-and-white people in Snowcrash's color Metaverse.

We dismantled our project, a virtual House of Usher peopled with actors, one of whom went on to win SF's Nebula Award. I also took down the version in OpenSim, after the host we had closed shop.

A former Linden told me "hey, buy some cool stuff from Turbosquid and import it to jazz up your build."

Sounded good until I priced it. Four items would blow the entire budget. We had already lost two hosts because of the .edu discount coming to an end. Game over.

Essid discusses his project to create a roleplay experience based on Poe's Fall of the House of Usher in this academic paper here and for HyperGrid Business here.

How Mesh Devalued Prims in Second Life & Made It a "Consumer World"

Works by the artist known as AM Radio in Second Life's pre-mesh age

Philip Rosedale's open challenge to name another metaverse with dynamic, collaborative content creation evoked this response from reader Vivienne Schell:

I think that Second Life's initial success was tied to in-world, collaborative content creation offering in-world tools. Leaving the Prim (do it by yourself) path and quitting further development on that field caused much of the decline SL went through after Sculpties and later Mesh became predominant.

The virtual sandbox has turned into a consumer world mostly, where many people pay money to a few people for creating some kind of "multiple social frameworks", but much of the original thrill which keeps Minecraft big (and made SL big) is gone. As a result Second Life has turned into a self-centered universe, where the latest mesh head or the latest body or the latest whatever shiny has become the most important issue for a core of die-hard users. That may be enough for a certain amount of people to log in, but it isn't -- by far -- enough to attract a broader audience.

Which sad to say, seems sadly true. Echoing that thought on Twitter is this from Jeff Berg, once best known as AM Radio, Second Life's most famed artist:

Watch & Share: How to Optimize Mesh Content Through Retopology

Here's a handy tutorial for Blender users who want to optimize their mesh content via retopology, i.e. remaking an existing mesh surface with more optimized geometry. Even if you're not a Blender user, it's definitely something you might want to share with your favorite mesh content creators who likely might (actually almost certainly) need some tutorial resources on making their mesh less lag-inducing.

This video via Black Dragon creator Niran, whose latest update to his SL viewer has a new feature for checking how well-optimized or not a mesh avatar accessory is:

SL Mesh Demos Demonstrating More Triangles Doesn't Mean More Detail

Quick: Before pressing Play on the video above, which set of SL mesh shoes and shirts are more detailed, and which have more triangles? The answer, as they say, May Just Surprise You. 

This video demo via reader Nadeja, who posted them as a response to last week's post, "Does The SL Community Care Enough To Clamor For Well-Optimized Mesh?" It's a perfect way of refuting the commonly held assumption that with mesh, more triangles means more detail. Not at all, not at all.

"Adding triangles helps only to some extent, but then it becomes wasted resources," as Nadeja writes in Comments. "Optimization means to find the right trade between visual and resources, to get the best look at the smallest cost."

Look closer at this comparison of tops:

The top on the left, she notes, has "[l]ess than half of the triangles and it even has wrinkles. You optimize for the platform you have. You did that in the 90s and early '00s as well. Better shaders would allow more for sure, but that doesn't mean that you can't do a better job... The evidence with similar models shows otherwise."

But surely this doesn't apply to full mesh avatars, some skeptics may say.  Nadeja has you covered there too - watch below:

Does the SL Community Care Enough to Clamor for Well-Optimized Mesh?

Pictured: Slink Redux body bundle, a rare example of popular SL mesh avatar content that's also well-optimized

Last Friday I wrote about the new Black Dragon viewer update which checks whether a mesh avatar is well-optimized or not. Reader "Down With Underboob" (LOL) tried it out in SL over the weekend, to see how much mesh content was poorly optimized, and came away shocked:

So I took Black Dragon out for a spin at a fairly busy club. There's only one avatar in the place who is "decently optimized" per the report. This guy is wearing a system avi and system clothes... I like what BD offers, but please get real. Nobody wants to go back to looking like an avatar from 2008. Photographers like to use BD's capabilities, but again, they're not going to use system avatars to pass these complexity tests.

Show me some mesh bodies, heads, and clothes that pass these tests and don't look like they're from 2008, and I'll consider trying them. Does he think we're all going to run around bald? I don't want Warcraft hair in SL. Until then, I'll stick with what I have. I'm satisfied with the 15-20 fps I get at this club and I can get upwards of 100 FPS when I'm hanging out on my platform at home in Bellisseria.

I'd also point out that he's tilting at windmills if he thinks designers are going to start posting complexity/triangle counts in their vendor ads.

I'd say it's the responsibility of the many SL bloggers and YouTubers to report on the complexity/triangle counts of the content they feature, warning them that high poly items will likely cause lag. Failing to do so would be like a real life fashion reporter failing to mention the clothing items they're recommending are made of toxic materials.

As to her request for optimized mesh to try out, Slink is among the few mesh body creators well-known for optimization.

But is there a market for good optimization in general? Reader "Seph" thinks not so:

New Black Dragon Feature Shows Your Much Lag Your Second Life Mesh Avatar Is Causing (Likely a Lot)

Pictured: Niran in a Belleza 5.2 Bento mesh body -- which has over 1.5 MILLION triangles

There's a new update to The Black Dragon viewer for Second Life, and it comes with a great new feature: An option to check how well-optimized your mesh avatar is -- and shows you via a color code whether it is (green) or poorly optimized (yellow/orange/red):

So rather than confusing the user with an artificial "maximum" it now clearly states that it is simply recommended to stay below a certain point. Reaching this point also no longer immediately results in the "super bad" red rating but rather gives you the "eh" yellow/orange rating, see it as warning. As before you get a summary of all ratings together, right now its still very simple but i might actually add a 3-point answer that takes the 3 worst offenders and gives you hints and possibly even guides (or links to those) how to solve these.

As previously reported, most of the top-selling mesh bodies in SL are very poorly optimized, and a key culprit for Second Life's poor performance in general. Linden Lab refuses to do anything about this -- the CEO basically told me "So what?" when I asked him face to face -- but Black Dragon creator NiranV is a passionate advocate of optimization, and wants to improve SL's environment:

Black Dragon Creator Praises Optimization of Slink Mesh Bodies -- And Challenges Creators Like Maitreya to Follow Slink's Example

Black Dragon viewer creator NiranV, who has been ranting about the utter environmental disaster that most SL mesh bodies are, has rare words of praise (along with some advice) for the Slink mesh bodies created by Siddean Munro:

The body being 15k triangles is pretty good. That's really good optimized, although the hands and feets weighting in at 63k alone is quite heavy, there's still some work to do there.

I'd start by not bundling hands and feet together and rather making them a "pick-one" scenario where you attach/detach only the one you like. That alone would save a tremendous amount of triangles. Further she could make use of Bento to offer the "styles" of feet as animation override, given that the styles are things like normal, standing on tip toes etc. Anything else would fall under the pick-one scenario anyway.

To summarize this, there's still a bit to go but this is definitely the right way to go and other creators like Maitreya should definitely have a look at this example how to do it.

Alicia responds by adding more background on Slink feet -- and problems Slink faces when trying to compete against market leader (and leader in poor optimization) Maitreya:

Slink Mesh Bodies Are Well-Optimized for Second Life Avatars -- Here's How

Slink, a popular virtual world brand for avatar enhancements, produces what's basically the Tesla for mesh bodies in Second Life -- not just sleek and stylish, but energy efficient and good for the environment. This is because their creator, Siddean Munro, painstakingly optimizes them to have the least number of triangles without hurting visual quality.

This is an important update to a much-discussed New World Notes post from 2018, where I included Slink among the list of mesh bodies found to be poorly optimized, a key source to virtual world pollution -- i.e., lag and non-displaying avatars. Ms. Munro disputes Slink's inclusion in that post, so I will update it accordingly. And since then, in any case, she has updated her mesh bodies to be more efficient.

"I do work hard to make sure all my meshes are as optimal as they can be, use as few textures and scripts. as possible and that they are also laid out and written efficiently," she tells me. 

As evidence, she offers this image below:

Linden Lab Needs to Punish Poorly Optimized Mesh -- Or Reward Well-Made Mesh (Comments of the Week)

Good reader thread responding to my somewhat tossed off remark that, "While you should definitely encourage users to customize their avatars, you also want to impose optimization restrictions so avatars don't become a bloated lag-inducing disaster like they have in SL."

The problem, as Alicia observes, is the company behind Second Life has enabled (encouraged?) poor optimization:

Linden Lab has set the maximum for textures at 1024x1024, and most creators end up making their textures too damn big (any texture size over 1024x1024 max), which then will be forced to the max of 1024x1024; but, remember that LL never set limits on how many verts/faces a mesh item should have. The textures issues, on the other hand, can easily be fixed if people simply went with the allotted sizes (64x64, 128x128, 256x256, 512x512 and the 1024x1024 max) SL can manage.

Also, knowing how to keep the MB down on the textures also helps if the person knows what their doing in GIMP, PhotoShop, or any other photo manipulation program. When it comes to mesh, you still have mesh hair that shoots an avi's complexity near or past 100k, and the same thing can happen with jewelry, clothing, etc. Verts/faces counts also increase/decrease when you resize any mesh object (trees, shrubs, etc.) All of these individual issues ultimately becomes a huge problem that never gets handled as LL only wants to continue working as is...

Everyone knows that LL will never impose restrictions on how maximum number of verts a mesh item can have.

Pulsar argues that Linden Lab should set out rewards and/or restrictions around mesh:

Why It Will Take Awhile for Bakes on Mesh to Make a Difference in SL

Commenting on Cassie's video comparing SL mesh bodies, and her note that Bakes on Mesh doesn't seem to improve performance [Ed. Not exactly, see update below], Patchouli W provides this analysis:

BoM's true value contribution will only surface once onion-skin geometry is stripped away from BoM-optimized versions of the same body and people stop wearing more than one copy of the body geometry because of the removal of said onion-skins. The other optimization that will need to happen - and this is trickier because of limitations on how BoM works currently as well as so many legacy mesh clothing items being reliant on them - is the removal of legacy alpha cutting in favor of a BoM model that can handle 'cuts' made by using system layer alphas - this would reduce the amount of duplicated geometry required to fake smooth transitions between disparate pieces of geometry as well.

Only when these two changes are pushed through in full will the full value of BoM in cutting the impact of mesh bodies on the SL grid be fully realized. Then there's the issue of trying to port this work over to cope with normal and specular/env channels baking, which is an entirely new kettle of fish since the way normals are blended is a little different compared to diffuse and specular texture layering.

This may be true or not true, but what's striking to me is all the 3D graphics technical jargon one must know to understand the state of Second Life mesh. Or as it was expressed in the immortal line from The Limey: