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:

“There’s one thing I don’t understand. The thing I don’t understand is every motherfucking word you’re saying.”

Anyway, Patch's basic takeaway -- be patient:

In short, it's going to take a while, and you need to support your body maker of choice in the transition by encouraging them and providing the feed back they need as they release beta tests of your current body, because if they see it as a thankless task, they may just give up and let go. And that's no good for any one of us.

And don't forget to watch Cassie's mesh round-up above!

Update, 10/22: I didn't accurately convey what Cassie said, as she explains in Comments:

I saw no difference in lag with any of the bodies I had, since I was also in an area that had no lag at all. Since this video, I've been constantly using BoM, and while not everyone can see me yet, it has been a big improvement. I can wear 20 layers at a time, including stuff that creators no longer in SL, had made. The numbers are a huge difference for the bodies that were specifically made for BoM, not the ones that are currently only relays. If you look at both Signature and Slink, their numbers are much lower, and they took the time and effort to do so, with a lot of love and care in their system.

Apologies to Cassie, and thanks for her explanation!

Comments

[Olivia]

I think mesh clothing is so popular and is the cause of a lot of complexity and that won't change much.

But people using mesh body "onion skin layers" for skin, tattoos, freckles and so on.... That can now be baked onto the skin so it should mean fewer textures have to be downloaded by viewers. Also BOM can be used for applier clothes too of course but they are less popular than they used to be. I still use appliers for underwear but not for much else.

Personally I would like to see an optional version of my Maitreya Lara mesh body without the tattoo, underwear and clothes layers (or perhaps just underwear). As currently the body is basically 4 bodies one inside the other (skin plus the three layers outside). All of this geometry presumably has to be loaded regardless of whether there are any textures on each one.

[anon]

This really always comes down to linden not doing its job, just trying to push its work on others.

new engine
new engine
new engine

If their current employees cant do it, then fired them and get people who can or lease a engine that works the best with your files. its 2019 and your trying to constantly stitch together a 2003 engine time and time again. Im pretty sure when the engine was made in 03 they couldnt even think of ray tracing,GPU Power, virtual reality, peer to peer, or even blockchain . All we can keep doing is fighting cause the company is lazy, and keeps saying downgrade downgrade back to 03.

There is no incentive for mesh body creators and clothes makers to update in second life. They dont get monthly payments from their customers , so its really just you dont like it "thanks for the fish" mentality.Its said in the slink group chat everyday.

BOM wont be truly a wtf moment till the next bento update comes up and they have to update their bodies to "Bom versions that are downgrades" and they say oh yea we stopped supporting your body that you paid for and liked for system avatar 2.0

[Cassie Middles]

I said it doesn't improve performance? I don't remember saying that. I said I saw no difference in lag with any of the bodies I had, since I was also in an area that had no lag at all. Since this video, I've been constantly using BoM, and while not everyone can see me yet, it has been a big improvement. I can wear 20 layers at a time, including stuff that creators no longer in sl, had made. The numbers are a huge difference for the bodies that were specifically made for BoM, not the ones that are currently only relays. If you look at both signature and slink, their numbers are much lower, and they took the time and effort to do so, with a lot of love and care in their system. :<

[Pulsar]

Slink did exactly that by removing alpha cuts and the onion layer The other popular female bodies are confusing the SL folks: someone believes they need an HUD etc for BoM, when in fact it's just like the classic avatar to use - including lots of layers per time (makeups, tattoos...). With SLink it's a much better BoM experience. I gladly purchased the body as they deserved it. SLink even has a couple of options to avoid the body-paint effect.
Kudos to Ruth too, the open source body, mod and lightweight.

[camilia fid3lis nee Patchouli Woollahra]

anon:

BoM as it currently stands is basically the same baking system as applied to System Avatars currently, plus a few nudges forward to recognize the value of asymmetry in skins (FINALLY, having arms with different textures and actual words on tattoos that aren't reflected into gibberish), and then ported over to be painted on mesh avatars rather than the underlying System Avatar.

It's early days, but for effects that don't require anything but diffuse, we should try to use BoM as an option, if only to encourage those responsible for this push in efficiency to continue working and add on other desirable options such as support for baked alpha cuts, normals and specular/reflectiveness, perhaps even for the classic System Avatar as well as a backport to keep Ruth 3.0 competitive for the low-load fashionista.

If you want something bad enough, you need to show that you'll support the use case for it, and this is what should encourage the attempt to use it more often where feasible. Granted the mesh body designers will need to meet halfway in the middle with the right changes to make the full value of the shift felt, but nobody said it was a one-paddle boat.

And frankly, I don't know who you really are, but I am getting really sick of this kind of s--t, where people don't appreciate the difficulty of upgrading a system that has been in almost non-stop operation for sixteen years and where upgrades need to get it right on first or second bounce as rollbacks may cause bigger problems than they solve. Please call me back as soon as you've managed to push visual upgrades over the same timescale on the same kinds of audience size as Second Life because I really don't see you being capable of offering anything but bellyache. THANKS.

[anon]

That just seems like a straw man argument and fanboy talk. shutting down criticism because im not a computer programmer doesnt mean im not aware of tech and shouldnt say. Wow shouldnt i be able to cross sim boarders. oh wow my avatar did this yesterday , but cant today, Wait why do i have to buy the same skin again when the uuid didnt change, wait why do i have to make my own alphas in paint when i didnt before, wait i dont get the big update because i didnt buy the whole store of feet and hands.

Games go dead from bad patches and ban hammers all the time. I want bom to work , but it wont work if its a consumer downgrade. I never say its not a technique update, i always say consumer for a reason.

Do you really think a new engine at the end of the day on a program 17 years old wouldnt improve 99 percent of the problems in sl? I only say new engine now , because it doesnt seem avatars will every be able to go past the system 03 avatar. If the new technical update needs to limit avatars back to system, then the writing it on the wall. I get it your a purist , but consumers are not.they dont fight on forums they just leave and competition has never been higher for second life, what keeps a lot in secondlife, is that they invested in items with real cash. If bom saids they no longer matter , your giving them more of a reason to venture out.

[camilia fid3lis nee Patchouli Woollahra]

anon:

There are things I used to wear or use a decade ago that I no longer use. There may technically be no expiry date on anything in Second Life generally, but in practice things change and improve all the time, enough that sometimes what we were once fond of becomes laughable or even technically unusable.

And the BoM Update is not limiting users back to system. They can continue using applier-based systems side by side while switching over to BoM or system avatars when they need to cut the load they place on the grid. As for upgrading the graphics engine, there were already two major shifts earlier (Windlight and ALM) and SL no longer has the laughable mockery in quality that it once had back 2003. In fact, work continues apace with such efforts as EEP as well as tolerance of thirdparty efforts in juryrigged VR and pushing the envelope by burning compatibility with legacy shaders (take a bow, NiranV).

One other thing you'll need to understand when looking at the slow pace of development in SL on anything is that changes once pushed to the grid may not be possible to roll back without causing damage, and that these changes may cause a marked change in the load placed upon the grid as a result of them being spread to a very large number of people. Where possible, expect LL to make changes that take advantage of pre-existing systems and formats rather than make extra work to support more stuff.

When you look back at what has been added since 2003 to cope with advances made to the system, you'll notice a marked lack of new asset types and formats: besides adding the mesh uploader with later patching for classic/fitted/Bento and animesh rigs, disabling lossy compression for sculptmap textures, and adding the ability to reference extra textures for use in ALM on surfaces, there is basically very little that is new in terms of the visual asset types LL has opted to support in SL between 2003 and 2019 - everything is either a reuse of previous definitions, or subtle additions. This is something you get a lot in any software development that has to support significant amounts of legacy content, VR/VW or otherwise - balancing consistency with change.

And I'll put my name against this claim any day.